SURVEY OF DYNAMIC COMPU1ATIONAt. TAX POLICY … · life-eyc!e behavior, intergenerational issues....
Transcript of SURVEY OF DYNAMIC COMPU1ATIONAt. TAX POLICY … · life-eyc!e behavior, intergenerational issues....
UNIVERSIDADE NOVA DE LlSRO4 F aculdade de poundwfwmia
SURVEY OF DYNAMIC COMPU1ATIONAt
GENERA EQUILIBRIUM MODELS FOA
TAX POLICY BVAUAlIOI
Alfredo M Pereira o Working paper N81
i i
UNIVERSIDADE NOVA DE LISBOA
Faculdadc de Economia
Trav EamptGvio Pillto
CampoliJ~
1000 LSBOA Outlbro I 1987
Ocober1987
SURVEY OF DYNAMIC COMPUTATIONALGENERAL EQUILIBRIUM MODELS
FOR TAX POLICY EVALUATION()
Alfredo M Pereira
University of California San Diego and
Faculdade de Economia Universidade Nova de Lisboa
and
John B Shoven
Stanford University and NBER t
r) We would like to thank the participants of the NBER Applied General Equilibrium Workshop
Sianford May 1987 the International Symposium on The Social Accounting Matrix Methods and
Applications Naples Italy June 1987 and of the II World Basque Congress Donolla - San
Sebastian SpaIn September 1987 for helpful comments and suggestions The usual disclaimers
apply
bull
1
Survey cl Dynamic Ccmputational General Equilibrium Models
for Tax Pelley Evaluation
1 Introduction
There is a well established and fast growing body of literature focusing on tax policy evaluation
using disaggregated computational general equilibrium (CGE) models That such models have become
so popular Is hardly uPrling In fact the CGE approacih has several advantages over more macro
orienled aggregated models or analytical parllal equilibrium analysis First ttre CGE methodology
allows the study 01 differential impacts acnoss sectors of production andocnossconsumer groups
Second U allows one to consider the int~ractions among different sectors and agents so that the
policy evaluation is not biased by ceteris paribus assumptions Third and in a more technical vein
it makes use of flexible computational numerical techniques Analytical tools often become
intractable for disaggregated models Fur1hermore the CGE modeller does not have to be confined to
small changes in parameters This i an Important feature because large changes In policy
parameters ere often nremplaled in most tax reform proposals
Historically the field was fostered by Ihe early pioneering work of Harberger (1959 1962
9SG) Herberger uses a highly aggregated analytical model to focus on the taxation of capital
Income The first fully dlsaggregated computational general equilibrium model was introduced by
ShavenWhalley (1972) to evaluate the effects of differentiallaxalio of Inceme from capital in the
us
Until the early eighties most empir[cal general equiiibrium work for tax policy evaluation
i I I
I I I
I I r
bull
2
static tnvestment passively actjusts to saving The typical instruments in the US tax system are
modelled and the government budget Is balanced This state of affairs Is discussed and surveyed In
Shovn (1983) Fullerton-Henderson-Shoven (1984 and Shoven-WhaJ)ey (19S4) t I
Integration etc - has Induced Important developments There have been attempts to Incorporate
life-eyce behavior intergenerational issues allocation of savings investment decisions and
adjustment costs corporate financial decisions and government deficits In to the CGE models Underlying all of these features is the need to endow the models with a dynamic structure w~houl I which most of the above issues cannot be adequately addressed
IThe first efforts to incorporate dynamics into the CGe tax models are associated with bull
Fullerton-Shoven-Whalley (1978 middot1981 1993) Consumption decisions incorporale some t I
int~rtemporal aspects and an Intertemporal equilibrium path for the economy Is obtained by Ihe 1
Isequencing of static short-nm equilibria The equHibna are connected through the evolution of the I
capital endowments which_ tn turn depend on saving middotbehavior A compr~1e discussion of a recent I i Iversion of Ihis model is provided in BallardmiddotFulierton-Shoven-Whaney (19SS) I
More recently a whole new generation of models has been developed by Andersson (19S7) Auerbach-Kotlikoff (1983 1984 1987) Ballard (1983) Ballard-Goulder (laeS) Bovenberg I (1984 1985 19B5) ErlichmiddotGinsburgh-Heyden (laB7) Fellonstoin (19B4 ISBS) Goulder I (19BS) Jorgenson-Yun (1984) Pereira (leaSh 19B7e ta87d) and Goulder-Summers
(1987)
tThis article focuse~ on slJch efforts Our scope is narrow By concentrating on decentralized i
I i
Mo~e recently j the ~hrust towards better descrIptive power together with the nature of the tax
policy issues under analysis effects of a ponsumption tax capital taxation corporate tax
3
numerical W~ieral equilibrium modelling designed 10 addresS tax polic issues we abstract from
several ether related areas and approaches First we abstract from analytle models along the lines
of Abel-Blanchard (1983) and Judd (1985) Second we abslract from centralized growth-type
megels of laxation like Charnley (1981 1982) Third we abstract from eentralielti models
I
associated with World Bank researchers See for example Adelman-Robinson (1978) and
Dervis-de Mele-Robinson (1981) Finally we abstract from other CIl fields See for example
ShovenVlcalley (1984) for a sUlVey ollotematlonal trade applications DecaluwmiddotMartens
(1985) ane Robinson (1986) for applications In the area of development James (1985) for a bull I I
survey of economic history applications and A Manne (1986) and BOrges (1988) far all I encompassing surveys
The models included in our comparfsion are I
I 1 8allardFuliertan-ShavenmiddotWhalley (198S) l 2 Andersson (1987)
3 AuerbachmiddotKotlikoff (1963 1987)
4 Bailarcl (1983) and BallardmiddotGould (19aS)
5 Bovenbeg (1985 1985)
6 ErlichmiddotGinsburgh-Heyden (1987)
7 FeltenSleln (1984 1985)
S Goulde (1985)
9 GouldermiddotSummers (1987)
10 Jorgenson Yun (19B4)
11 Pereira (1986b 19870)
bull 4 I
The kei features cf these models afe summarized in Table 1 which provides a structure for the
discussion Most of toe models Ihal we examine could be said to be In the Yale tradition That Is
they wer developped oy sudnts oi Herbert Scarf or students of the students of Herber1 Scar This
shyincludes BalladmiddotFuerM-Shoven-Whalley Andersson Ballard-Goulder ~)Venberg FellenSlein I
ibull Go-cider GouldermiddotSJmmers ard Pereira The exceptions from this heritage are Auerbachshybull
Kotlikoff Erllch-GinsburghHeyden and JorgensonmiddotYun I This survey is organized ltis follows rne second section provides a brief overview of I
Inon-dynamic CGE models The third seolion offers a discussion of modelling economic behavior and
equilibrium The fourth section focuses on model implemenlalion and policy ev~uatlon The fifth
Isection summarrzes empirical evidence en selected policy issues The sixth section illustrates the
power and wea~nesses of dynamic modelling with the issue of corporate laxintegration Finafly the last section summarizes the state~of-the-art in terms of dynamic modelling for lax policy
evalutlon and sugg-ess areas of future research
_
--
--- -~ ----- __-shy
_ _ - _- - Con~um9rs_ 8amphavlor
Bajla rd -Fulflonmiddot -Shaven-Whalley
1985 Anderaaon
1087
Auerba ch~ Ko1Uko 1983 1987
BaHElfdmiddotGouldor 1983 1905
Sovanborg 1985 198e
ErHchmiddotGlnaburghmiddot middotHeydeR
1987 Feltensteln 1984 1966
Gouldar 1985
GouldorSummor 1987
Jotgeneon-Yun 1984
Porelr 1966 1991
Count dynamlc dl9agg~ objecth lastes constraint ondog endog OvorlapBnqvMI of datil
year _seltlng labor no savings gonorst mollves
inlertsmporal 12 Income otility eel IntertempOfll1 yes yo no noLS 1973 Ymiddotmiddot
_ groups max ---_
0 Intertnmporal rllcufsivs
Swaden 1984 yes aggrogate utility CES eq of molion 00 yos na no consumer max for weaHh
tntortorrporal Isoeiastlc~ 00-1983
ill yo 55 poundtUS u~illy middot19831 In tcemporal yesmiddot 11El7 yes yo 00
cohorts rn~x CES-19C7 plauslblo
bYPothlC1 inlsrtamporal
19n yes 55 ago uliJity Isoelastlc Intertemporal ybullbull ye ye ysLS BFSWexpanOOd _ cohorts max
0118 inlertemporal recursivo LS plausible ye aggregate ollllly lsoolastic eq ot motion no yes no no
hYPOlhotic consumer max for weallh jnterlemporlll Boigium 1960 1983 ro aggrogate utility St()ne~Gearyjner1emporal no yo no no
consumer max Inlerl(lmporal two periodmiddot two yes
Austrs1i 1991middot1982 yes 2 income ulllly CD yeluly ybullbull tn the flrsl no no 2-perlod) groups max constraint porlod
Intertamporal LS 1973 ybullbull 55 aga utility eel intarhtmporal ye ybullbull ybullbull y
BFSW bullbullponded cohorts max
intert9mpor~1 W 1973 ves aggregate ullily 180085110 intertemporal no ybullbull no no
BFSW bullbulllMded consumer max (I~gt IntertempQ(a~
W 1955-80 ybullbull aggregate ulilily trenslog intertemporal ybullbull yes no no consumer max
Inlerlemporal recursive 1973LS yes a income ulllily CmiddotD eq of motion yos ybullbull 00 no
BFSWxpanded groups max for wealth
middot5middot
bull-- _---- -- _-----shy~ -~
I AlJlt 1 - uynafllc (lIt MOOero fur all runey tVlIIVllllQlI
dynamic dluggI utllng
I Ba~ll-a rd-fullerton
-Shoven-Vlhalley I no 19 sectors 19S5
Andarsson 1987 ye 2 sectors
~ Auerbach~KotHkof no~1gB3
1983 1967 yesmiddot1g87 1 sector
Ballard~Go uld ar 1983 1985 no 19 sectors
Bovenborg 1965 1986 yes 2 sectors
Erllch~GIn5burg hshy-Heyden no 24 seclors
1957 1__ Feltenataln 1964 196e no 30 sectors
GoUldsr 1985 no 3 sectors
~ GoutdarmiddotSumrnamprs
1987 yes 5 sectors
_ Jorgen9ClIO-Yun ona sector
1984 no produces two goods
Poralr 19S~ 1987 yos 4 sectors
~bJecllve
Produc~rs
technology behavior InpIJ1 1-0
matrix adl costs opt Imal
capital stock
sialic cost min
CES in valu$ added
110
capital labor yes 00
np Invest adjusts
to sFvnns
max 01
interlempora flot rash flow
cnsl min-lOS) Imax of valul)
of the firm~1987
ill
CD-ISBa CES-1987
capital labor
captel labor
no
no
YA quadratic in
I nO-191)3J
qutldratic in Ilk-1gB
ye
YAS
static cost min
CES in vafl1oadded
10
capita labor yes no
no invest edjuss
10 savifQs max ot capita yas ~
value ot the CD labor 00 quadratic In yo Ilrm Ilk
static CD in several no global ltlost mln vaiue added
110 ltsp1al
Jabor Ves no Inves adjusts
to MVrHIS
statle oost min
____ slatio
cost min
max of value of the
firm equivalent 10 two~stage
~ost min me) of
CO in value ad~ed
110 CES In
value added 110
CESin shyvalue added
110 dual
trBRslog price function Cobb-Douglas
capUal labor
cepHal labor
capital labor
capital labor
capital
yes
yes
yes
no
no invost adjusts no to S8vinos net of
El-xog 9011 deflcitsl no
no invest adjusts to savings r
yes quadratic In Y
Ilk
no yo I
yes Intartamporal net cash flow
In V Added 110
labof Y quadratic In I
y bullbull
- 6 -
~ --- ~ ~- -- -~------ --- --- - --~--~ ---~ shy
_
- ---
--
--------
- - - --------
H n v I taxes
olllng allowed debt Ballardfullamprton~
dynamic objective constrfnt tupendlttire (IlleUs endogenous
aU major balanced trade yoarly ybullbull rly endogenous -f ~Shovon WtlitUy US taxes with C(in~Hlt00 pubHc uility balancod cQmposilon no no debt
eiasticllv junel l 1985 rnagt eCP-REV~d1 aU fn3jO[AndtH3Son vearly balanco small opan Gcon J Swodish inturn fa to cr raturn1987 no - inaia dlJn axogonolJs 00 Ini ITnininod poundgtPfUV
-I L)~ in~h~ --~ ~~ ~pound~~---1- ----shy
runrac h~KQllfkofi ai gulD
1963 1937 US m1l3syo - intertnmponl tixogancuamp yo ys cos~d ampConoroy bandmiddotfinanced flon-oppoundIla
B iii lIa fe-Gould 9r nil llltl1gtr balanced 1rndayearly yearly crcoge)Jus
US t)X6S wiih conslanl I
1963 1965 no public utility balanced cOlYlposWon~ no nD debt max budnt ExP~RCV alastlcHy lunct
Bovenberg simp~iriad 1985middotclosod econybullbull rl~ no - balanced exogenous no no debt US laxes 19a6~two counlry1985 1986 bodgel model
Ert leh~G tn aburghmiddot yeally simplified Inlert walt funet Ishorl~run ttads Heyden no balanced exogenous 00 no Belgium
taxes _ desloullbtium flltUensleln
1967 budaet EXP=REV 6-1mprifia-dmlm costs of yoarly yo rast of tho world
1984 1985 producing a ExTax Rev exogenous financed with yo Australian ~s a Ihlrd conSUmer public good money and bonds non-optimal taxes group
Goulder yaarly annual and endogenous aU maJor
laquo1()Sed economy1985 publlc utilily In1ertemporal composillon bull yes yes U~ taxesY rna versions IlEFElQgtfpoundV bond-financed (non-opllmal balanced trademiddotGoulder-Summer yeary yoarly endogonous bull all major
1987 00 public utility balMc composition no no US taxes with constant max budGel EXPREV alasticilY fund Jl)rgenaon~Yun yoarly aU major
no balanced exogenous no1964 US taxes closed economy budGet inteftemporal recurslve endogenousPereira all major
yo public utility eq of molion lovel and yo yos1995 1987 US laxes closed economy rna for debt composition bondmiddot financed oplloll
- 7 shy
bull - -__----- _ - ------ -- ------- _-----
---- ---
------
------
ll-ltiLt 1 uynamlc Lut MOOSS fOf I ax YOtlcy tv8luallon
Inanell ~ets and maf~9ts ojumLrtm- ----_ ----c_~
privata do btl GlvldtHlI pL~lc aliocalQflo rls shy
1rlc ypa expGct assets equity retention dobl 01 savlnq$ promlums path
Sal ra rd fu Barton phys cap1 used 10 buy soquence lt
Shavon-Wha1ley no financial exogenous oxogenous no physicBl 00 bull 01 sIalic 1E myopic 1985 assets capital flqrlil -
Andersson phys cap used 10 buy 1987 no financial no no yo physical no dynamic ~ perfect
assets capital Auerbach~KolUko If phys cap usod to buy
1983 1987 neo financial no no physical cap no dyoamlc ~ perfectYmiddotmiddotnSSf)ts and Qovt debt SailBrd-Goutder phys cap
1983 1985 no financ[al assets-shy -------
Bovenberg phys cap 1985 1986 no iinenclaf
asso1s Ertlch-Glnsburgh~ phys cap1
-Heyden no financlaJ 1987 assefs
FeltoRsteln phys cap 1984 1986 Invostment is
bondmiddotfinanced
used to buy
exogenous exogenous nObull physl~al no dynamic ~ perfllcl capital
used to buy
no no no physical capital
no dynamic ~onlinuos lima
PI perfect
used 10 buy no no no physical no dynamIc PF perfect
capital jllNt) porlod) bull used 10 buy
no no ye physical cap no dynamic PI p4)rfect and govt debt (Iwo Iods) bull
Goulder phys cap used to buy 1995 no financial exogenous exogenous yenbullbull physical cap no PFlfE bull IdynamIc perfoct~assels and am dabt slaUc
- -~ - shyGoulder-Summers phy$ cap
1987 bonds and no e)(oganOU$ no used 10 buy ye dynamic PF perfect _ bullbullultv bonds MullY (exoaenous
Jorgensonyun phys cap 1984 no financial
essels -~ - ------
Pereira phys cap 1986 1967 bonds end
equHy
________
no
8xogenoos
no
exogenous
no
ye
used to buy two type$ of physical
copltal used 10 buy
go dobl bonds equity
no
no
dynamIc
-
dynamlc
~
lE
perfect
fIbullbulllble (nonperlecB
middot6middot
~-- _middot~ ___~w_ - - ___ bullbullbull___ -------- shy~
-----
------
- ---
-
IlflpnUItll tit lun pVII-y PIlUAlIV
q6H1brla oqual fax polley ovnluaU-on efUclency derlvatton Ioorllhm
calibration param~hr computation COmelfed llold tndfcaHH eHecls
Bnl jard~FuU6rtQ n base year back soll1lion steady-sta1e yo inlsrgroup inerseel or Iii
VS transilion lump-sum with Interfemporai 1965
~Shovn Whallmiddotv replication wllh exog Morrills nmamo1ers alQorilhm _~ond~middotsta1$ ears Inc CVor EV
i~noerucn sleady-sfala no agg(ogaled lnlalsectora 1967
base year back sobHon SiMNON package replication wHn oxog valiant of VS Iransilion no Indicamptor lnlertamporal
paramelsrs multiple shoo ling 6~~~l~slat8
AU8rt)ach~Kotllkot steady-slato yo Inlerguna-r a lions
1893 1llH
quatitaUve by replication a$Sium~lkml Gauss-Soidel liS Uln1WOn any rule -lilh 1111ermpOrB
sonsililJ1y I _ ~poundI~Jjnnelill $t(lfdvmiddot~luiA J3djlt15S ~~ nnrilW ~--------
a ail r nimiddotGould or bas9 9tH bac~ solution Ballard~ staadymiddotslalo )05 kltHSclo(a~
1983 1985 VS Iransllinn lump~sum dynalTlk CV inlerlemporal steadY slafe parameters alaorlthm replicaUon+ with oxog Gourder
+stead~stato (ers loc inlareel10rat BCvonborg qunmaiive by dynamic version stoadY-$Igtto yes 1985 1985 YS transition some govt eN Inlortempora
S9rHltivilv OnearizaJian method ropllcaUon assumptionl of Johansons
+steady-sfata paoo palh Erllch-Glnsbu rgh~ actual vorsus no aootagalod inlersectoral2 bftse years back solu1ion variant of Negishis
Heyden replication wIth alltogbull optlmizaUon countorfaclUal no ~odjcator inlottempora 1987 paramlJters aocroach
faltonstaln baso yeat back solution actual versus no aggrampgalfld intersElclornl replication wilh exog Merrills counlerfaclutll no indlclltor middotlntamprlempOull1984 19B6
parameters al90rithm 1981middot1982
Gouldar staady~slala Inlar5ecioralbase year back souUon Variant or 1S lronsilion no dynamic CV inlortempora
stoady state parameters Newlon 1985 feplicalion+ with iUog Fair~Taylorl
+slaildy-sla1e Intargeoorat Gtlulder~Summers base year back solution varIant of stlJadymiddotstate fixed no aggregated fntersectoral
vs trarlsition government Indicator Intertempo1al steady state oatameters Newton
repticaUon+ with exog Fair-T aylQf 1967 +ste-adv-stata bullbullbullndlno financial
Jorgenon-Yun econometric steady-stale Money melfic In1ersecloral
- - 8slimalion vs traosUian no social weUare functton intertampotal1984 +~teadlmiddotstatl)
~ shy
qtJalitaUv8 by NPSCt optimal path ye generalized inlersectorai 1986 1987 Perelr
replication 8S$umptlonl optimizatIon V$ lumpmiddotsum dynamic lntertemporat se~~ttilitf ____Jllgorilhm opllmal pah pen inc CVoclV financial
- 9 shy
bull ~ ___ - _______ - ____ bull___ OM bullbull __ bullbull __~____ ~_
--
- - - --------
--------- - - - --------
---
1nltJ1 bull I ~ 1J11ltI1 v I ~
-
Sa Ita rd-Futlertonmiddot intargration consumption lax ~StHlven~Whall~y VAT capila~ or labor tax
Ch~mqflS1985 Andersson
consumpion tax neutral corporals lax system 1987
it ue rb a c h~ i(at II Sltotf COIlSUft1ptian lax (poundpird ItS hibor Mx~on ufocto of taxation on capital fOfrnJilti1
dvnamio fiscal policv effects af OQvernmanl defioit (3tlard-Goulder
1883 1987
allocts 01 the adoption of a 1ge3 1985 pure consumpHon lax
im~rtmce of cons fCreSClht -Bovenborg analytically orianiad approach
1985 1986 effects ot pure conumption tax capllal income taxation in Opo economies
Erllch-Glnburghshy-Heyden reat wage policios in Belgium
1987 Felter1steln 1984 - theorotical
1985 ~ linnncial crcwdlng out1984 1986
Goulder effects of tax- vs bond-financed 1985 chanQes In pubtlc expondllufe
~~--
Gouidr~SUmmampfI torpora1$ lax cvts reduced 1987 ITO increase 10 gasolioe taxes
Jorgenson-Yun
effects ot several programs of tax reform on the allocation of ~~ptai
1984
-- Purelra
corporate la)( Integration effects cf romiddotinlroducinQ imreslamnl tax credits
1986 1967
10 bull
bull -~- ~- --- _ - ---____--_ - -----_-_ --~ ---- ~
1 1
2 Brief overview of non-dynamic CGE models
The origin of most or the models we survey goes back to the work of Scarf (1967 1973) whlgt
first developed a reliablo algorithm to compute equilibrium price for a ArrowmiddotDebreu economy
HiS algorithm used simpicial Bubdivision techniques and can -be shown to be the computational analog
of the fixed point theorems previously used to prove the existence of equilibrium His technique
could soW a model with an arbitrary number of consumers and commodities as long as all agents
were price takers consumers were subject to budget constraints demands were continuous andmiddot
production did not display increasing returns to scale The alQorithm while guaranteed to converge I
I I I
was relatively slow for prcb1ems involving more than S1If 20 dimensions A maior improvement in
computational speed Woo offered by Merrills algorithm (1972) which used tpe same fundamental
ideagt as Scarfs procedure
Scarfs model (as is true for the standard Arrow-Debreu model) does not include a government
sector ~ neither taxes nor public goocs At one of the mosi promising appHciUions of the new
computational techniqu w~s in Iha area of tax policy evaluation Shoven and Whalley (1973)
exlended the general equilibrium model and computational approach to include a wide rry of taxes
and a governl11ampnt spending plan The original ShavenmiddotWhalley model was static (although the
different commodi~ies CQuid be considered simitar goods available at different datest as In
Arrow-Debrau) and government was aumad to run a balanced budget The data are arranged as a
social accounling matrix The models spacification and calibration is checked by solving it in the
presence of the base set of laxes The result should ba exactly tha initial social accounting matrix
After having pasd this replication check the model is solved for a counterlactua equilibrium In
the presence of a new tux design The result is once again a social accounting matrix The two
equilibria are compared in order to tSS($S the impact of the new tax plan The first uses ot this
bull
1 Z
model for lax policy evauajon were ShovenWhalley (1972) Whalley (1975) and Shoven
(1916)
Completely static models as Shaven-Whalleys are unsatisfactory for many tax reform ISsues
These include corporate lax Integration effects of investment tax credits effects of accelerated
depreciations consumer Or e)penditure tzxes~ importance of saving subsidies like IRAs etc These
are essentialy dynamiC ksues They involve not only the allocation of capital across sectQfsbut
perhaps more Important the capital intensity of the economy But the capital accumulation and
capital reallocation take time and may involve adjustment costs Because of these issues Ballardmiddot
Fulierton~Shoven~Whaloy took the first steps towards developing a dynamic model
In this conlexl the ShovenWhalley model has been extended and implemented for Ihe US
-economy In BuIIBrdmiddotFulienonmiddotShovenmiddotWhalley (1985) While Ihis book completely documents the
mo~el it was used in several publicalions beginning in 1978 Their mode) consists of nineteem
production sectors twelvo households and fifteen consumer goods It includes a very detailed set of
taxes including the federal and state personallncome taxes federal and state corporate income taxes
SOCial Security taxes pro1J0rty taxes unemployment Insurance excise taxes sales taxes etc It
has been calibl att-d to reproduce the 1973 US economy Recentty a version corresponding to the
1983 JS economy has been developed
wThe 8alard-Fullenon-Shoven Whalley model is dynamic In the sense that consumers face a
choice between current consumption and leisure versus future consumption (which can be
purchased via s wlngs) The consumer classes act as if they were maximizing a nested CES utility
function over the domain of a CES aggregate of contempo-raneous consumer goods and leisure and
future consumption subject to their income constraint The parameters of those funcHons
determine the shares of income devo1ed to each commodity to saving and to the purchase of leisure
They also determine the two key elasticities in the models ~ the elasticity of labor supply with
bull
13
respecl to tgte rbullbull1 aoer lax wage rae and the elasticity of saving with respecl 10 the real after lax- rate of retum to capital
In the model consumerS have myopic expectations regarding lulure prices and in particular
regarding the future rale of return 10 capilal Ballard (1983) and BallardmiddotGoulder (1985)
inco~ora1e both perfect foresight and limited foresight into this model Future consumption is
acquired by buying a fixed composition portfolio of real investments that offer an infinite annuity
of returns
The production side of Ihe model is compielely static The mod1 Incorporates a constantmiddot
olasicily 01 substitution betweeen prlmary inputs in production (capital and labor) and fixed I coefficients for intermediate inputs The model distinguIshes between Industrial outputs and
eonsumer goods for the simple reason Ihat tile data are classified differenUy Industrial sectol$ I I involve such categories as forestry and fisherle metal mino9 and publishing and printing while
consumers purchas-3 fUfHi1ure 2ulomoblles and books This fact is recognized in the model by Ibull
incorporatirg a secord stage Of production which converts industrial outputs intO consumer goods
This technology is uStally mod9-led as a flxed~coefficient conversion matrix
The BallardmiddotFINton-ShovenWhalley model assumeS that the private sector finance
marginai ioveslment with the same composition of debt and equity as currently exists in each sector
This is the same arsumrtion that Herberger originally used Investors all hold debt and equity in tile
same proportion Thampreiore this ownership can be aggregated into simply capital ownership For
tax purposes however the separate treatment of debt and equity is taken into account at both the
corporate and persofa Io=vel SimllarlYJ the dividend poIicies on corporale equitY are established
exogenously There are no government bonds in the model since there are no government deficits
The modol is solved for a sequence (as many as 100) of temporary equilibria wltll consumers
allocating income between present and future conumption at each point in time The palh for 1he
14
economy ~ a sel of connected equilibria The connectIon is pra~ided by capital accumulation
Capital accumulation is endogenous and determined by saving The model stans with a sociagt
accOunting mi1rix In the bzse case the economy is assumed to be in a steaqy state growth -path (along
which all relative prices ~fe constant) The madel solves for both the naw steady state growth path
and the transltion 10 it zf~er a policy intervention rhe authors have frequently addressed the
question-of howong rt tKes to effectively settle un10 a new steady S1ate growth path
The dynamics of the model are limited however in that future consumption is collapsed into a
composite commodity Aso the absence of government deficits and the lack ofproductlon dynamics
limits the realism of the BallardmiddotFuliertonmiddotShoven-Whatley model for the analysis of dynamic
policy issues (such as the adoption 01 a consumption tax or Ihe elimination of the investment tax
credit) The models we survey try to Improve upon Ihe Ballard-FlIiertonmiddotShaven-Whalley
dynamic modelling The models Jn~orporate more forward lookTng behavior They improve the
dynamics on the conampumrtkm side of the economy and some of them introduce true 1ynamlc behavior
on the part oI the producers We turn now to the issues inY9lv_ed in developing such dynamic models
15 I I ~3 ~ssues in the modelling Of dynamic behavior
We will 100k firt at the speclficalion of economic behavior of consume(s~ producers
goyemment and Ihe r~sl of Ihe world The modemog of corporale and household financial decisions
I is then addr5sed Finally Ihe eencept of equilibrium is discussed
I
i
I 1
31 Consumers behavior
Early efforts 10 build dynamic features into the economic behavior 01 consumers are due to
Ballard (1963) and Auerbach-Kotllkoff (1983 1984) Their work has been closely followed by
several subsequent authors In faet dynamic household behavior Is the single most pervasive aspect
amohg the models surveyed
All the models Inc-O1orale some loro of lila-cycle behavior Household behavior is dewmloed
by the maximization of an addWVely separable time invariant intertemporal utillty function The
utility lune1ion is defined oyer the domain of the consumption goods In the economy In most of the
models leisure is also an argument in utility so that labor supply is optimally determined
Utility maximi4ation is subject to a lifetime intertemporal budget constraint which equalizes
the presen1 value of consumers income and expenditure More reeentiy in Andersson (1987) bull
Bovenberg (1985) and Pereira (1986b 1987d) the constraint i$ defined as a sequence 01
recursive equations of motion on wealth This has the potential advantage 01 accomodating liquidity
constraints However It should be recognized that in the absence 01 liquidity constraints (ie when
consumers are free to completely borrow against future income) the two specificallons of the
househofd ccnstraint are essentiay equivatent Furthermore tn both versions saving is optimally
determined as a way of translaring wealth intertemporally
1 6
Some lodels now have a sophisticated dynamic specification of hoosehokl In Ballard (1983) -
BallardmiddotGoulder (1985) and Auerbach-Kotlikof (1983 1984 1987) consumers behavio is
embedded in an iterQEneraional setting Each of them has S5 age cohorts simulataneously alive
Som-e measure of jntergener1ional altruism is considered in the form of bequest motives In Ballard
(1983) and Ballard-Gouldr (1985) consumers derive utility directly from bequthing parl of
their wealth Therefore bequests assume 1he form of a scrap value of terminal wealth of a
generation In AuerbachmiddotKotlikof (1983 1984 1987) each generation empathyz bullbull with Ihe
utility of future generations ovar bequethd wealth In addition Ballard (1ge3) includes detailed
demographic poiections wfill the intent of incorporating in the policy analysis the effects of the
bull Post World War tI baby-boom
In the real world household decisions also include the optimat allo~atiOn of saving among
shyalternative physical or financial assets Theories of household portfolio behavior are relatively
Iwen~establshed However they involve uncertainty as a crucial element Accordingly they have
middotnot been incorpora~ed in the context of the determinIstic dynamic models we sutyey here We will 1 Come backt-O this Issue below
32 Producer behavior
The efforts to build dynamic features into the economic behavior of producers are more recellt
and less widely adopted The first attempts are due to Bovnberg (1984 1985) and Summers
(1965) Dynamic behavior has been more fully incorporated In the recent models of Andersson
(1987) Auerbach-Kotlikoff (19B7) GouldermiddotSummers (1987) and Pereira (1986b 1987d)
Part of the reason why production side dinamlcs has been more slowty adopted is the weak
supply of accepted theories referring to the dynamic behavior of the firms In the models referred
I 17
to above gynamlc producion and investment b~havior are middotinduced by the existence of capital
adjustment costs and linked to Tobins q theory Adjuslment costs are designed to capture both the
incomplete mobility of capital aoross industries and installation costs ~e the costs of adjusting
capital towards its cpHmai level)
Adjllstment costs can be conceived as internal to the firm and measured in terms of foregone
outpul along the lines of Lucas (1967) Alternatively adjustment costs can be viewed as actual
external cests incurred together with Ihe purchase costs along the Unes of Gould (1969) With the
exception of Pereira (198Gb 1937d) all of the CGE authors follow the former approach
The firms in the economy maximize their market value as the present discountad value of the
future stream of dividends In Andersson (1981) and Pereira (198Gb 1987d) firms are seen as
maximizing the present discounted value of net cash flow Maximization is constrained by Ihe
adjustmenl cosl lechnology and ~n equation of motion describing Ihe evolution of the capital stock
It should be noted that undereogenous divldendrelention rules the two problems are equivalent II
should also be nOled Ihat a salisfectory economic rational for th bullbullistoo of dMdonds with the
present tax code is missing in the profession (Shoven (1986))
The models with static formulation of producers behavior are characterized by passive
investment behavior Investment merely accomodates to saving in the eccnomy With a dynamic
formulation induced b adjustment costs real investmen decisions are forward looking Investmenl
Is endogenously and optimally determined by the firms A fundamental difference between the
shonmiddotrun in which capital stock is given and longmiddotrun in which Ihe level of capital is allowed to be
optimally determined is emphasized
This extra richness of produc1ion dynamics is not without costs A careful look at the summary
tables will clearly show an inverse relation between the adoptlon of dynamic features in production
and the level of disagreggation of the production side of the economy In fact with production
bull
j
1 6
dynamics $e dimension of the problems is immensly increased Let IS be more specific
In a static framework with constant returns to scale production technology the output level ismiddot
indeterminate The optimal aliocation of inputs can be obtained by cost minimization with any
feasible output level generatillg zero profits Such zero profit conditions are used to solve for the
output prices in terms of the factor prices and hencemiddot to reduce the dlmensionality of the problem
from the -number of commodities and factors to the number of factors Thus even if the model deals
with 30 production sectors the computation of an economic equilibrium can take place using only
the dimensionality of the primary inputs in the economy
Now under certain regularity conditions on the production and adjustment costs technologies
leading to enough concavity of the optimality objective the intertemporal output path for the firm is
endogenously optimally and uniquely determined even with constant relums to scale technologies
(see Pereira (1985 1987a) Accordin~y with adjustment costs in the model optimal profits
will in general be non-zero This result has crucial implications for the computatiQn of equilibrium
in the model~ with production dynamics With adjustment costs no reduction of dimensionality is
possible The curse of dimensionality returns as a binding constraint
The introduction of adjustment costs adds another significant complication to the model With
capital being less that perfectly mobile in the economy different rates of return on capital will
exist in different sectors This is a difficult problem to tackle conceptually in the absence of
uncertainty Also in the real world producer maximization choices include also its choice of
financial ratios (debVequity) and payout rates (dividendretained earnings) These financial
subjects are difficult and much Clf this behavior is still taken as exogenous by the modellers We
will come back to these issues below
middot i
19
bull 33 government behavior
middot
Two issues dominate the modelling of government behavIor First govemment behavior has
Ibeen typically seen in Ihe CGE liIerature for lax poliCY evaluation as constrained by yearly balanced
I budsets (see for example 8allard-Fullrton-S~ovn-Whafley 098Sraquo The analysis of
govemment defidlS and public debl in a CelE context requires a dynamic selling Second the level of
government expenditures is either exegenouty given as in Auerbach-KoUikofi (1984 1987)
6ovenbrg (1984 1985) Feltenstein (1984 1986) and Jorgenson-Yun (1984) or
endogenously (but no optimally) determined by the balanced budget conditions as in
6allard-Fullerton-Shoven-Whalley (1985) Andersson (1987) Erllch-Ginshurgh-Heyden
~
(1987) Gouder (19aS) and Gouder-Summers (1987) In the second case the composition of
bull public expenditures is often optimally determined However the leve~ of government expendi1ures
can only 00 endogenously and optimally determined if the government is seen as an optimizing agent
and is allowed to run deficits
The first attempts to deal wiU the government defiCits in CGE tax models are due to
Auerbach-Kollilltoff (1984) and Feensain (1984) In Auerbach-Kotlikof (1984) government
expenditures afe exogenous They grow at the rate of growth of population However given 1he tax
structure and tax revenues yearly deficits and surpluses are atTowed subject to an in1ertempond
constraint that the present value of future tax revenues equals the present value of future
expenditures
In Fellenstein (1984 1986) government expenditures are also exogenously given He also
allows the government to run deficits to finance expenditures in excess to 1ax revenues However
surpluses are returned to CQnsumars in the form cf transfers Accordingly government ts not
subject to any constraint regarding the future repayment of public debt
In GouJder (19851 the 9Vernmenl maximizes a static social welfare ~nctio subject to a-
bull
20 I i
I
I I I
I bull
balanced budget conSirtlL This follows the optimal alocatkm of government expenditures alpng
the lines of Baliard-FHeiton-Shoven-Whalley (1985) Ttle model Is generallzea to allow for
exogenous changes in HIe time path of government expenditures Two financing alternatives are
consjdeft~d and contr8s1ed EldditionaJ tax revenues and bond issuance
Pereira (198Gb 1137d) auempts to address both the incorporation of deficits and the
determination of govement expenditures The path of government expenditures and the path of
delicilSlsurpluses (ano therefore the path lor debt) are endogenously and optimally determined The
government is seen as mltJximizing n intertemporal social welfare function given the 13x structure
Optimization is subject to a sequence of recursive equations of motion reflecting the evolution of the
ptlbllc debt allowing fDr government budget imbalances It should be stressathat this specification
01 the government cOMtcalnt is equivalent to the specificaUon in Auerbaeh-Kotlikoff (1983 1987)
when no liquidity constraints affect -gove~nment behavior
The extra richness in the treatment of government behavior allows tfte examination of
government debt poices in a truly dynamic setting Also financial crowding out effects induced by
government deficits can be analyzed (see Auerbach-Kottlikoff (1987) Feltenstein (1986) and
Perer (198Gb 1987e))
34 Foreign sector
Most of the open economy models follow the assumptions of balanced trade with import and
export net demands characterized by constant elasticities along the line of Ballardmiddot
-Fullerton-Shaven-Whalley (1965) Such is the case of Salfard (1963) Ballard-Goulder
(19S5) and Goulder-Sllmmers (1987)_ In Feltenstin (1985) the rest of the world is treated as
21
an addjlion~ consumer group
Bovcoberg (1986) develops a model In which two economies lIr considered each following
intertempoml perlect foresight paths These economies meet in the international forum Their
trade relaroMhips are characterized by yearly balanced trade accounts
None of the new generation of dynamic cae tax models has yet IncorPorated the lnternational
capital flows as done in aoulder-Shoven-Whalley (1983) for the earlier Ballarrshy
-Fullertonmiddot Shaven-Whalley (1985) model This is unfortunate as there is an important
Intertmporal aspect to international lending The first al1empts along Ihese lines are due to
Andersson (1987) and Erlich-Glnsburgh-Heyden (1987) Andersson (1987) in his model of the
Swedish economy adopts a closed economy appraoch in which rates of return in the domesJic
economy are largely determined by the international capital markets Fidinterest rate induce
Intlnatlonal capital flows which determine and finance the International lrade imbalance In
lurn In ErlichGinsburghHeydenmiddot (1987) foreign trade Is generaled according to an
intertemporal Irade welfare lunction with constant import and export elasticities In the shortmiddotrun
they allow inlernational trade inbalances which generate capilal ftows to the domestic households In
the long run however trade balance is assumed
35 The need for financial markets
A dynamic economic Wucture not only provides the ideal environment 10 model many features of
economic behavior it also permits one to Incorporate the financial side of the economy If the
government is allowed to run deficits the question 01 deficll financing automatically follows If
investment is optimally dewmined and returns to capital are different aClOSS sectors problems of
investment financing arise If there are seeral final1cfal assets In the economymiddot government
4-
I i
ibull
I I I I
I
j
22
bonds privpte bonds and equity (or simply physical capital installed inmiddot different sectors) the
problem of allocation of saving among assets with potentially different returns arises
The papers surveyed var greatly with respect to l~e extent of their attention to the financial
side of the economy At Clne extreme are the models in ~anardmiddotFuliertonmiddotShoven-Whalley (1985)
Andersson (1987) Ballltld (1983) Ballard-Goulder (1985) Bovenberg (1985 1986) and
Erlich-Gjnsburgh-Heyden C19B which are devoted exclusively to the real side of the economy
In turn Auerbach-KotliKoff (1984 1987) Feltenstein (1984 1986) and Goulde (1985)
allow for government debt In these models saving finances changes in government debt and physical
capital Private and publlc 2ssets are perceived by the households as perfect substitutes The
allocation of saving merely adjusts to the relative demands for funds
Feltenstein (1984 1986) is the only model surveyed which introduces money Government deficits are financed by issuing money and bonds according to an exogenously given rule Money is
demanded by consumers for transaction motives and an exogenously given fraction as a store of value
On the other hand government bonds and physical capital are the vehicles for the intertemporal
transfer of wealth
Summers (1985) and GOlldermiddotSummers (1987) introduce a whole menu of financial assets shy
firm specific equity capital Different assets earn different rates of retum However such rates
are equal up to constant and exogenous sector speCific risk premia Therefore the introduction of
constant exogenous risk premia as helpful as it may be in the context of calibration does not solve
the main issue of the non-optimality of the allocation of saving Also talking about risk premia in a
deterministic context is somewhat unsatisfactory
Pereira (1986b 1987d) also introduces a whole menu of assets private and public bonds and
firm specific equity However all the assets are expected to yield the same rate of return and
therefore perc~jved as perfect substitutes The different asset types allow consideration of
23
exogenOU$~8tUequjty triO dividendlrc~enlion rufes and therefore seve-ral sources of investment
financing bond equity and retained earnings
The nonmiddotcptimalHy of the allocation of saving and the absence or exogenelty of corporate
financial rules is _a reftectlon of the limitations of the determInistic approach Either in a context of
perfect anticipation oi the future prices Of in general within the t$atm of point price
expectations all the papers follow a deterministic approach Under such circumstances consumers
either expect diHerent rales of return (inclusive of risk premium) across assets in which case
they wilt buy only one asset (that with highest rate) Of they expect equal rat~s of relurn in which
case Ihey are indifferent about the asset composition of Iheir portfoliO There is no way of
Iradingmiddotoff rales of return and risks to obtain an optimal interiof solution to the problem of the
allocation of saving
The most advanced contribution in the modelling of saving allocation in a CGE setting is due to
Slemrod (1980 1983) in the context of a static one-period model In his model consumers act
according to a two stage separable decision process They Hrsl decide on how much to save Then
they decide on the allocation of saving according to an Indirect uttlity function dependent on the rates
of ratum and variances offered by the different assets in the eoonomy The sourCe of riskiness in the
ampConory comes from an uncertain marginal product of capital On the other hand aside from
portfolio decisions the res of the economy is insulated from uncertainty
The most complete contribution in terms of the treatment of the corporate financial rules is
FulienonmiddotGordon (1983) In a variant 01 Ihe BailardmiddotFulienon-ShovenmiddotWhelley (1985) model
Fullerton and Gordon have capital intensity and optimal financial decisions Jointly deter~ined
through a two siage process The cosl of financing capital is minimized by trading off the tax
advantages of debt against the epected real bankrupcy coSls inherenl with high debVequity ratios
Given the optimal debtequity ratio the level of inveSlment Is chosen such that at Ihe margin the
24 l
bull
rerurn pn ~uity equas the return on bonds plus an exogenous risk premium
36 Market assumptions equilibrium and expectations
Virtually all of the papers arc characterized by Walrasian market clearing assumptions All
markets are perlect1y competi1ive Atomistic competition among agents 1s assumed even thoughmiddot only
a flnile number of agents is considered Virtually no mark~t disequilibria or price stickiness afe
considered The only exception is ErlichmiddotGinsburghmiddotHeyden (1987) bull In thei model of the Belgium
economy I the _wage rate is fixed in the short-run Therefore in the shOrlwrun desequilibrium in
the labor market will generate endogenous unemployment However in lh$ long-run all prices
Including the wage rata are Ilexible and accordingly ali markets ciear
Given the dynamic nature of behavlor in the economy markelmiddotcfearing prices in each period
depend on expectations of fture prices and on tax variables In the economy There are essentially
two ways at interpreting the economic equilibrium in such a dynamic conte~bull If future prices are
perfectly anticipated (io expectations are self-fulfilling) a perfect foresight equilibrium
prevaHs Then future actions are merely the implementation of current decisions for future
periods However jf ptica expecla110ns are not perfect (Le agents make mistakes w1th re$pect 10
future prices) then a temporary or shortrun equilibrium prevails Markets ciear and clearing
prices depend on fu1ure price expectations Current plans about the future are typically not
precisely implemented They will be revised as more or better information becomes available to the
economic agents See Grandmant (1982) for a comprehensive survey 01 the temporary equilibrium
literature
Wilh the exception of Gould (1985) and Pereira (19S6b 19S7d) all the models surveyed
adopt th~ concept of a perfect foresight equilibrium In turn BaliardmiddotGouldr (1965) considers a
J
2S
flexible am9unt of foresight In terms of the number of years over which price movements are ~
foreseen Pereiras model (1986b 1987b)) is flexible in a somewhat different way in that it can
indude any range of foresight from myopia to perfect foresight The choice between the perfct foresight and lemporary equilibrium is ultimately to be made on
philosophic grounds It can be argued that less than perfect expectations Imply that agents are
irralional in some way (see Auerbach-Kctiikoff (1987 p 10) However the reverse argument can
be made One can question wiether agents are reaHy rationar and perfectly knowledgable about
future prices
Recent evidene of Balird (1987) Ballard-Goulder (19851 Goulder (19851 and Pereira
(1986b 1987d) suggests that the choice in modemng expectations is an important one They show
that the degree of foresight inl0 thc future (ranging from perfect foresight to-myopic expectations)
may have dramatic impacts on the pollcy conclusions of Ihe model Accordingly the best research
strategy may b to design models which are flexible enough 10 allow for different rules regarding the
formation of expectations With the exception of the articles just mentioned sensitivity analysis
bullhave previously not been performed along this dimension
In terms of implementation the two concepts of equilibrium - perfect foresight and temporary
equilibrium have different implications The dimensionality of the equ1ibrium soll1ion algorIthm
is involved Suppose we have a model with ten markets to be run for a period of 50 years Aside
from normalization a perfect foresight model implies computing prices In 500 dimensions while a
-t~mporary equilibrium model requires sorving SO equilibria each in ten dimensions Given that
computational speed often varies with the cube of the number of dimensions the lemporary
equilibrium formulation is potentially strikingly more feasible However Ballard (1987) and
Goulder-Ballard (1985) have devoloped techniques to greatly speed the computation of a perfect
foresight equilibrium
26
j
I
4 Jrnplemenlalion and Issues on policy evaluallon
41 Calibration and equilibrium comparlslon
CGE models at typically parameterized by the use of a calibration procedure Some parameters
are exogenously given However some crucial parameters are determined in such a way that the
model replicates the data for a given base year See MansurmiddotWhalleY 11984) for an eXlensive
discussion of 1~i$ issue
CalibratIon in a dynamic context is generaliy interpreted as requiring two properties First
replication of base year data is required Second the model is parameterized to simulate an
intertemporal balanced growth path when Ihe base policy is maintaineq This Is the approach
followed by Ballard (1983) BallardmiddotGoulder (1985) Goulder (1985) and SummersmiddotGoulde
(1986) It follows the practice of BallardmiddotFuliertonmiddotShovenmiddotWhalley (1985) with their
sequential cquilibrium dynamics
Other authors Auerbach-Kotlikof (1983 1984 1987) Boyenberg (1984 1985 1986)
and Pereira (198Gb 1987d) follow whal we cali a qualitalive calibration The slructural
palametars are exogenously chosen so that lh economy follows a reasonable path into Ihe fulure
ThaI has 10 do wllh the fact thaI given the recursive nature 01 the dynamic economy and aside tom
such structural parameters only iniUa stock values are needed to run these models Given initial
conditions on the stocks of say private wealth capltal and government debt agents will optimize
and Ihereby generate a lirst round of net demands and equilibrium conditions In lurn the
equilibrium prices will determine the evolution of the stock variables into the next perIod
There are several potential problems with calibration in the conlext of dynamic models The
assumption of a steadymiddotstate growth ~ath in the base case can be questioned First while
27 I Ibullbull
steadymiddotstaU is a possibity it certainly is not the only meaningfull solution to dynamic models
Even in the case of a perfect foresight equilibrium the model implies an equilibrium path which i may or may not involve colanced growth The model not the modeller should dictate the nature 01 I themiddotbase case path Second in the context of a temporarY equilibrium path a steady state solutlcn is Inol a likely model QutcOn1tl In fact unlessmiddot expectations are stattc short run behavior consistent
with a steady-state evolution will in general not be generated On the other hand if static I expectations are self~fufjili[1g we have in fact a perfect foresight model Third even the base year
replication requirement may cause problems In Ihe cOntexl of temporary equilibrium Any
calibration parameter would be conditional on expectation rules which is probably an undesirable
feature
The nalur of the two-requirement calibration strategy is very muchin the spirit of the bull
traditional design of the comparision of alternative equilibria comparis_ion between a steady-state
base case on one hand and alternative paths including a transition period and a finat sleadyyenstate on
the other hand Pereira (198Gb 1987d) compare different (not necessarily steadystate)
equilibrium paths The arguments against such a procedure are based on the idea that the impact of
policy changes can be observed most easily since all departures from the steadymiddotstate can be
attributed to the altemalive polioy On the other hand the base case so dEfined as a steadymiddotstate is
consist~nl with previous work in a less dynamtc setting and Iherefore allows a common standard
for comparing model results
42 Computation algorithms
We are still al a stage in which basically each author uses a different computational technique
Th development of dynamic models - in parlicular wilh adjustment costs andior perfect foresight
bull
28
has corresponded with the decline in the use of fixed point algorithms In -fact given the relative
arge dimensions inevitably involved such algorithms lend to be very inefficient at the best and
chen prohibitively stow See Stone (1985) and Preckel (1965) for a comparative assessment of
different computation techniques Among the models surveyed only Ball1rdFulle(ton~
-ShavenmiddotWhalley (1985) and Feltensteln 1985 use Merrills variant of the fixed point
algorithm 1echnique
AuerbachmiddotKoHikoff (1963 1984 1967) follow a three stage procedure They first compute a
base case steady-state then a revised case steady state and finally transition path for the economy
berveen these two steady-states tn all stages a Gauss-Seidel iterative procedure is used
Ballard (1982) Ballard-Goulder (1965) Goulder (HiSS) and Summers-Goulder (1986)
~
use a method developed by Ballard and Goutder which is similar in many awects to the FairmiddotTaylQr bull
(1983) algorl1hm Short-run equilibria are calculated (using Merrills algorithm) parametric on
nrice expeC~lions The model is then iterated to generate self-fulfilling intertemporaf expectations
and the corresponding perfect foresight equilibrium In a relatively similar approach Andersson
1987 uses a simulation program SIMNON developed In the University of Lund This program
can handle two-paint boundary problems In a fashion consistent with the multiple shooting
algcrlthm (see Lipton-Poterba-Sachs-Eummers (1982))
Boyenbergs computational approach (1985 1985) differs from the other models surveyed in
that he relies heavily on analytical techniques Computations are done by using a dynamic version of
Johansons linearization method Being essentially determined by the continuous time nature of the
I I
f
model thiS linearization model has the disadvantage of confining the analysls to infinltesinal changes
around the base case equilibrium (see Bovenberg (1985) p 53) Pereira (1986b 1987d) uses an optimization algorithm NPSOL - developped by
rGillMurray-Saunders-Wrlght (1986)) This algorithm is used 10 compute the sequence of
I
29
Shortmiddotrun temporary equilibrium which makes up the intertemporal equilibrium path The
equilibrium conditions are seen as nonlinear equality constraints in the minimization of an artlcial
objective function The prices are normalized to the unit simplex by an additional linear equality I I 1 bull constraint
II ErlichmiddotGinsburgh-Heyden (1987) follow a unique approach in that they use a variant of the
optimization technique introduced by Negishl (1960) The economic equilibrium can be generated i
as a solution of a mathemallcal program Ihe objective function of which is a weighted sum of the
utility functions of the various agents while tho constraints set consists of the market dearing 1
conditions Ginsburgh-Heyden (1985) have extended Negishis resulllO tho case of downward price
lrigidities I
Finally the paper by Jorgonson-Yun (1984) is also unique in tIlat it is the only I
econometrically estimated model Different blocks for the consumption and production Side of the
-
model are separately estimated to provide the necessary structural parameters i
I The diversity of computation techniques is yet another indicator of the exploratory nature of the
body of literature surveyed in this article
43 Equal yield comparisions
The link between a base case and counteriactual simulations Is usually provided by the concept of
equal yield Shaven-Whaley (1977) discuss the meaning 01 equal yield in a general equilibrium
context when government is not allowed to run defICits Equal yield is interpreted to moan constant
public utility Government base case utility is maintained in the counferfactual experiments With
balanced budgets the concepl of equal yield is unambiguous The new equUibrium prices and the
balanced budge condition will del ermine the minimal expendilure and taxes needed to maintain base
30
case publi~ utility Accordingly in general equal yield is inconsistent with equal nominal tax
revenue Some change in tax revenue Is necessary Different tax replacement schemes ate
considered to assure that enough tax revenue is collected This is the approach essentially followed
by most 01 the papers surveyed Only Feltenstein (1985) Goulder (198) and Jorgenson-Yun
(1984) chose not to follow an equal yield strategy
The question is of how to Interprete lila concept of equal yield when the government is alowed to
fUn deflcLts The optimal leve of expenditure for base ease public trliUty can now be taAinanced
bond financed or financed by a mix of bonds and taxation We have several versions of equal yield In
particular equal yield may now be consistent with equal tax revenUE Furthermore some meaSure
ltgtf financial crowding out effects of government deficns can be inferred from the comparislon of the
several equal yield alternatives These issues are extensively discussed ~Pereira (1987b)
44 Price expectations and the dynamic generalization of compensation
Indicators
The sum of equivalent and compensation variations over households is the most widely used
aggregated measure of efficiency gains or losses In the context of $teady~state eomparisions andor
when complete future markets exist andlor perfect foresighl is assumed there are no difficulties
associated with the use of the standard Hicksian indicators In fact correct future prices are known
in these cases
If expectations are not selfmiddot fulfilling andlor future markels are not available a dynamic
generalization is necessary 8allardmiddotGoulder (1985) provide some steps in that direction by
defining an indicator that accomodates periods far into the future when households do not have
perfect foresight but a steady~state prevails On he other hand this issue is more fundamental in
i
[
I
i I I I I [
I I
bullI
31
the contex~ of a Gner1 lemporary equilibrium ftamawor~ In such Circumstances Pereira
(1986a) dev~lops a cnamlc generalization which is obtained as the present discounted value of a i i
sequence of short-fUn optimal expenditure functions consistent with a base case expected future I stream of IlWities I
32
5 Empirical evidenee from selected policy Issues
Dynamic tax models have generated several important results which escaped Sialic modellers
The following is a seleCpoundd set of issuos whIch stress the marginal benefits of dynam~ modelHng of
economic behavior in innovltlttive areas of analysis The discussion of a consumption tax emphasizes
the benefits of dynamic tusehold behavior and intergenerational aspects The study of the
flliminatlon or the re~intToduction of the investsnemt tax credits is made meaningful by the dynamic
modelling of production betlavor Modelling governmenl in a dynamic conlexts allows the modeller
to study the irrrpact of government deflcits Finally a dynamic framework lets us appreCiate the I
relevance of consumer expectations in terms of the evaluation of policy alternatives
i
i
51 Consumption Tax
t
By Its very nature a consumption tax can not be adequately investigated with a static model bull
Fullerion-Shoven-Whaicy (1983) use Ihe model of the US economy s described in
Ballard-Fullerton-Shoven-Whalley (1935) to evaluate the movement from the currenl US tax Isys1em to a progressive consumption tax Since their modeJ Incorporates a Jaborleisure choice
where leisure is an untaxed commodlty their results reflect the fact that both the consumption tax
and Ihe present system are dslortionary
Concentrating on the rntertemporaf distortions they show that sheltering more saving from lhe
current US income tax could improve econornic efficiency even if marginal tax rate increases are
necessary in order to maintain government fevenue~ At first you have a revenue shortfall
However the economy moves to a higher sustatned growth path and uitimatefy lower tax rates
cangenerate the same revenue path Also wages increase in the long run with this policy The
bull
bullbull ~
33 shyt
i
I
I
present value of welfare gi[lS for a potiey of complete savin9 deduction wilh marginal tate
adjustments (consumption lax) is simulated to be around $500 billion to $600 billion 01 1973
dollars
They also investiga1e th 1ngth of ttme it takes th~ economy to adjust to these policy changes
Roughly they estimate the long run to b thirty years although this figure Is very sensitive to the
specification of the savings elasticity - a crucial parameler in this model
52 Investment T~x Credit
Goulder-$ummers (1 S87) address the impact of eliminating the lnvestmenl Tax Cradit (ITC) on
intersectoral capital formation and on economic growth Their model isYa1ticularly adequate to
addies$ this issu~ in that is postulates a forward looking investment m9we with adjustment costs
They show that the eliminating the ITC causes a reduction in the rale of investment Inveslment is
estimated to fall by about 7 in the short-run and by about 12 in Ihe new long run steady state
In lurn a previous policy anncuncement lowers Ihe overall al1ractiveness of Investment and leads 10
a downward shift of the Investrrent profile On the other hand the combined effect of a revenue
neutral simultaneous efimiraron of the ITC and reduction of the corporate tax rates is a long run
reduction ef the capital Sieck by 35 This pattern suggests that a revenue neutral increament in
both Ihe corporale tax rates and the inveSlment lax credits would be preferable 10 the formula
adopted in Ihe US Tax Reform Act of 1986
Pereira (1987d) addresses the impact of the ITe from Ihe oland point of Ihe curTenllax syslem
under the Tax Reform Act of 1986 Given Ihe concern over the potential depressiVe effects upon
savings and investment of the new tax system in conjunction with deficit reduC1ion mechanisms of
the GrammmiddotRudman type a pertinent question is should Ih lTC be e-introduced In the context
bull
34
-of his mod~1 of the US economy Pereira focuses on the trade-off between the distortion in the
intersectoral allocation of capital induced by the lTC the lower tax revenues it generates and
potential financial crowding-out effects on one hand and the positive effects of lowering the relative
price of new capital goods on the other hand Preliminary results confirm the qualitative results in
Goulder-Summers (1987) suggestting a welfare gain of about 2 of the present value of GNP in
the best scenario of absence of replacement taxes Furthermore preliminary results show an
important time pattern to welfare gains with the average benefits increasing the further you look
into the future This pattern is due to the presence of constraints to intersectoral mobility of capital
and inertia in the adjustment towards the optimal capital levels as reflected by the presence of
adjustment costs
I
I
bullAuerbach-Katlikaff (1gB) in the context of their intergenerational model of the US economy
consider the impact on savings capital formation and interest rates of deficit policies and balanced
budget increases in government consumption They conclude that deficit finance and government
consumption can significa-ntly crowd out capital formation and lower the welfare of future
generations However crowding out from deficit finance is a very slow process because it results
from increased government spending over potentially long horizons Also deficit policies may
substantially influence the longmiddotrun interest rates while leaving the short-run rates essentially
unaffected Finally and in opposition to the central role adjustment costs seem to play in both
53 Financial crowding outeff~cts of government deficits
Goulder-Summers and Pereiras models Auerbach-Kotlikoff suggest that the time path of interest
rates induced by a policy of deficit financing seems to be insensitive to the presence of adjustment
costs
bull
35
54 The role 01 consumer expectatlons
The expectations analysis has uncovered one of the benefits of dynamic modelling
Ballard-Goulder (1985) find that the appeal of adopting a consumption tax depends on the level of
foresight possessed by the consumers Furthermore additional foresight may be welfare worsening
This is a second best type of result This tends 10 occur under policies that lead to capital deepening
and declining rate of return to capital over time To the extent that consumers have more foresight
they will be beller equippod to anticipate the fall in the rental price of capital Consequently if a
saving incentive is enacted people save more with myopia than with perfect foresight Given the
exi~tence of taxes on capita and the discrepancy between the private an~--sccial returns to capital
the greater savings with myopia is beller socially Ballard-Goulder find that the welfare gains
from a consumption tax are reduced by about 10 when we move from myopia to farleet foresight
Therefore the attractivenness of adopting a consumption tax seems fairly [obust across these
foresight specifications
I
bull
36
6 Tbe power and weaknesses 01 dynamic modelling the example 01 corporale
tax Integrallon
The Issue of corporate tax integration allows us to sh~w the stre-ngths and weaknesses or the
dynamic approach Empirical evidence using CGE techniques indicates that depending on the precise
scheme of integration and the 1ax replacement methods integratiOn may have substantial effects In
Ihe work of FuliertonmiddotKingmiddotShoven-Whalley (1980 1981) total integration was found to yield an bull
annual static eHielency gain of $4 to $8 billion in 1973 dollars Simulated dynamic gains may be as
large as $695 billion or abcu 14 of the present value of tulUre consumption and leisure in the
US economy These gains result primarily from interindustry reallocalions of Investment and an
improved inter~emporal allocation of consumption
Mora recent work emphasizes hOw consumers asset portfolio decisions and firms finanCial
decisions affect the efficiency gains from integration Slemrod (1980) focusing on consumers
asset ponfoio deCisions finds static efficiency gains which are about twice -as Iarge as those
repOrted iiy Fulierton-King-ShovenWhalley (1961) FuUrlon-Gurdon (1983) focus on the
firms financial decisions They roporl efficiency gains of 6 of GNP from the elimination of the
tax distor1ions favoring debt However when they eliminate the corporate tax and repiace it with
increaSed persona income taxes addlHonal dls10t1ions are created in the optimal labot~leisure
decisions These distortions tend to dominate the analysis slJ(h that the overall effects of complete
integration are very modest Galpr-LuckmiddotTodr (1986) address simultaneously consumers
asset pOrlfollo decisions and the firms financial decisions They also find very modest efficiency
gains from integralion
The above results are important but may be severely biased There are severa aspects of
economic behavior and modelling crucial for the study of Income tax Integration whiCh have no been
i 37
captured innny of the above papers One aspect is the absence of government deficitsmiddot a balanced
budget is assumed It is true Ihal resource crowding OUI is caplured In these mOdels but financial
crowding out induced by government spending is nOI Second investment is not derived from
optimtztlcn behavior Investment behavior passively accomodates endogenous saving decisions As
a consequence the differential impacts of policies in the incentives to save and to invest are not
captured Aso full capital mobility across secors is assumed with instantaneous capital
adjustmenlS towards optimal levels This assumption rules qut different costs of capital across
seclors and therefore differentiated reactions to tax policies changes Finallyhe modelling of both
gov9rnmerlt deficits and of endogenous real and financial investment decisions necessitates the
~nsideraiion of a dynamilt framework and the introduction of financiaJ assets govemment bonds and
iprivate financial assets A dynamic framework also highlights the efficiency effects of Integration on r
ithe ptimal intertemporal decisions The above models are either sIalic (Slemrods and
GalpermiddotltJckemiddotTode(sj or a dynamic sequ~nce of otherwise Slatic model bullbull I Pereir a (1986b) develops a dynamic applied general equilibrium model bull to study the tbull
efficiency effects of integration on the growth and allocation of investmenl across sectors Special
Iattention is paid to the structural affeelS of resource and financial crowding oul induced by
gove mect spending and deficits and to the real and financial Investment reactions across industries
to both tax integration and flnancial crowding out His mode departs from previous work on income
lax Integraticn and for that matler from most 01 the CGIE literalure for tax policy evaluation in
several direcllons It encompasses an endogenous sequential equilibrium struelure founded on
dynamic behavior with flexible expectations Government deficits are optlmaUy determined
Investment decisions are tward looking and the resull of optimizing behavior Several financial
assetsmiddot public and privatemiddot are considered
Simulation resulls Indicale that the welfare gains from Integration under large deficits are at
32
best lJlodest when measured in terms of the GNP The elimination of the corporate tax and its I replacemenl by increased income lax rates yields long run benefits which are never larger than
2 of the presenl yalue ct telr consumption and leisure onder the previous tax regime and 1
under the current tax regime However the short run w~lare effects of full integration tend 10 be
very low and in in some scenarios even negative This is a flew intertemporar pattern of efficiency Ieffects which reflects an adjJsiment lag in the interindustry investment decisions due to the
existence of costs of adjustment On the of her hand partial integration achieved by excluding I dividends from the corporale lax base syslematically yields negalive efficlencyeffeclS In Ihe long Inm these can be as high as 3 of present value of the future value of consumption and leisure This
is a new second best effect suggesllng Ihat less than complete integration may have perverse I efficiency effects
The dynamic slructure of PerclrBs model snowing for an improved freatment of real investment
declsions and government deficts provides a potential setting for a more accurate measure of the
costs and benefits of inlegration The simulation resullS suggest lower benefits and an intertempora Ibull
pattern of growing benefits
Simulation results also clearly suggest robust negative effects from partial integration How
reliable is this result If dJvidends were deductible from the corporate lax base (as are interest
payments) the preferenlial Ireatmant of debl over equily would be ellminaled Corporallons should
be expecled 10 react by decreasing the optimal deWeqully rallo However corporale financial
gecisions are exogenously set In Pereiras model as in aft of the other models surveyed that go as far
as dealing with the issue
Also withdrawing dividends from Ihe corporate tax base is a way of eliminating Ihe double
taxation of dividends at belh Ihe corporate and personnal Income levels How will the optimal
aliocellon of household saving accomodate that change in Ihe relalive return to corporale equity
39 I
t However hOusehold portfolio decisions are exogenouly set in Pereiras model as in all of the other
models surveyeltj go as far as dealing with the issue
IIt is sate to say that the evaluation 01 the benefits 01 partial Integration as defined above are
sevrey underestima1eO Meanwhile the distortionSoenerated by the increase in the marginal
personal tax rates designed to ralse the revenue foregone by the dividend exclusion are fully
aCCltlunted for A good measure of the costs of integration together With a poor evaluation of the
benefils may very well be the reason why partial integration is Simulated 10 yield neg alive
eHikiency gains
On a different vein as most of Ihe model surveyed here Pereiras is a closed economy model It
is legitimate to wonder how the resullS would change If international capital flows were anowed
IThis almost certainly would affect financial crowding out since a good d~al of the capital inflows
could be used to finance public debt
For these various reasons the results should be treated as preliminary but they strongly I suggosl that the dynamic structure provides valuable new insights into the corporate tax integration SSU2
40
7 Concluding remarks I I I i
What has been acccmplished Ihis far The success of Ihe CGE research in laelding the challenge
of dynamic modelling can not be denied
I iGreat progress in the modelling of household behavior has been achieved Promising
developm~nts in the modeUing of production and government behavior have also been accomplished
Interesting innovations in the concept of equilibrium and computation techniques were discussed
The policy analysis was greatly enriched by considering transitional effects together with longer
run steadymiddotstate equilibrium paths generalized equal yield strategies accomodating different
financial crowding out impacts and generalized dynamic policy evaluation indicators
An important set of issues have been addresed by the models surveyed Traditional issues
focusing on the effects of capital taxation and consumption taxes have been pursued In terms of
capital taxation for example the intertemporal nature of the issue was enhanced by allowing an
optimal evolution of capital stock in the economy and forward looking optimal investment decisions
In Andersson (1987) GouldermiddotSummers (1987) and Pereira (1986b 1987d) this is coupled
with a improved treatment of several tax provisions like investment tax credits and depreciation
allowances In terms of the consumption taxes developments in the specification of Intertemporal
household behavior the introduction of overlapping generations and the modelling of
intergenerational links via bequest motives as well as a closer attention to demographic evolution
provided a much improved economic setting for the understanding of the several aspects of the
problem
Furthermore dynamic modelling has permitted the addressing of a set of new issues Policy
Issues ranging from dynamic tax policy analysis to the evaluation of exogenous changes in
government expenditures to the impact of different methods of financing government expenditure to
4 1
the importance of tinanciai croHrii1g out to 1he relevance of consumorS expectations in tne policy
results have been addressed
Despile all of the progress and in part due to such progress several avenues are wide open for
much neded additional research The following seem to be the most Interesting and promising
areasmiddot First the specincation of liquidity constraints to household behavior may help to obtain a
better understanding of policy chenges that affect directly the interest rates in the economy Second
major attention needs to be paid to the incorporafion of a roraign sector (with open capital markets)
Third some efforts are needed in terms of finding adequate computation techniques now that
dynamic modelling eally makes the curse of dimensionality worse than eVOrbull
The modelling of financial markets is the single most unsatisfactory speet of the dynamic
models surveyed here~ The problems associated with provlding a mo~ adequate treatment of
financial markets and in general endogenous and optimal financial deci$ions bull both at the corporate
and at the personal levelsmiddot have 10 be addressed That necessarily requires Introducing uncertainty
into the models To b adequate this must go well beyond the mere assumption of point expeetalions
with the whole array of modelling implementalion and evaluation problems thai generates
bull I i
42
REFERENCES
Abel A and O Blanchard 1983 An Intertemporal Model of Saving and Interest Econometrica
51 pp 675-92
~ Andersson K 1987 Tar-ation of Capital in Sweden - A General Equilibrium Model Ph D
Dissenation UnIversity of Lund Sweder
Auerbach A L KotlikoH 1983 National Savings Economic Welfare and the Structure of
Taxation in M Feldstein ed Behavioral Simulation Methods in Tax Policy- Analysis Chicago
University of Chicago Press
Auerbach A l Kotlikoff 1984 Social Security and the Economics of Demographic-
Transition in H Aaron and G Burtless eds Retirement and Economic Behavior Washington DC
Brookings Institution
Auerbach A L Kotlikoff 1987 Dynamic Fiscal Policy Cambridge Cambridge University
Press
Auerbach A L Kotlikof and J Skinner 1983 The Efficiency Gains from Dynamic Tax
Reform Internatiooaj Economic Reyiew 24 pp 81middot100
Ballard C 1983 Evaluation of the Consumption Tax with Dynamic General Equilibrium
Models Ph D Dissertation Stanford University
Ballard C 1987 Tax Policy and Consumer Foresight A General Equilibrium Simulation
Study Economic Inquiry 25 pp 267-284
Ballard C and L Goulder 1985 Consumption Taxes ForeSight and Welfare a Computatonal
General Equilibrium Analysis in J Piggott and J Whalley eds New Deyelooments in Apolied
General Equilibrium Analysis Cambridge Cambridge University Press
Ballard C D Fullerton J Shoven and J Whalley 1985 A General Equilibrium Model for
t 43
Tax Eoliey Eyalua~iQO NBER University or Chicago Press
Borges A (1987) bull A Survey of Compula1ional General Equilibnum Models CECD Economic
Slud~ 8
Bovenbig L 19B3 imperfect Capital Mobility in a Perfect Foresight Herberger Model
Mimea UC Br~eley
Bovenberg L 1984 Capitel Accumulation and Capital Immobility Qmiddottheory In a Dynamic General Equilibrium Framework PhD dissertation University of California at Berkeley t
Bovenberg L 19850 Dynamic General Equilibrium Tax Models with Adjustment Costs in A iManne ed bull 1985 EtQnomrc ECUjilbrium Model Formulation and Solution Mathematical
lProgramming Siudy 23 Arnsterdam NorthmiddotHoliand
Bovenberg L t 1986 Capital Income Taxation in Growing Open Eeondmies Journa of Pubtic
Eamp(omlcs 3t pp 347middot376 IBovenberg L ald W Keller 1981 Dynamics and Nonlinearlties in Applied General
Equilibrium Models Internal Report Mlmiddot81middot208 Department for Statistical Methods CBS bull
Vooiburg The Netherlands
Charnley C 1981 The Welfare Cost of C~pltal income Taxation In a Growing Economy
IJournal of Political ECQnQm~ 89 3 pp 468middot496 I
Charnley C 19810 Efficient Tax Reform in a Dynamic Model of General Equilibrium I Mimeo Yale University I Charnley C 1982 On the Optimal Taxation In Stylized Models 01 Inlartemporal General
Equilibrium Mimeo Yale University I Decaluw B and A Manens 1985 Pays en Oeveloppement at Modeles Calculables OEquilibre
General - Une Revue de al litterature empirique Monographie Sene 9 Volume 3 Centre de
Recherche et Developpemont Economlque Unrslt de Montreal
44
bull
I I
I
I
1
Erlich $ V Ginsburgh and L van der Heyden 1987 Wher~ do Real Wage Policies Lead
Belgium bull A General Equilibrium Analysis European Economic Reyiew (forthcoming)
air R and J Taylor 1983 Solution and Maximum Likelihood Estimation of Dynamic
Nonlinear Ratioant Expeclations Models Ecooometrica 51 pp 1169middot1186
Feltenstein A 1984 Money and Bonds In a Disaggregated Ec~nonYin Searl 1-1 J Shoven
Eds Aoolied ~eoeral EqJiliDCiulD tioalYsis Cambridge University Press Cambridge
Feltenstein A 19B6 A Dynamic General Equllibrlu Analysis of Financial Crowding Out
Theory with an Application to Australia Journal of Public Economics 31 No1
Fullerlon D R Gorden 1983 A Reexamination of Tax Distorllons in General Equilibrium
Models in M Feldstein Ed Behayora Simulation Methods in Tax pOlicy Analysis Chicago
University of Chipgo Press
Fvllerlon D J Shaven and J Whalley 1983 Replacing the US Income Tax with a
Progressive Consumption Tax JoulOal of Public ECQnomics 20 pp3middot23
Fullerton D Y Henderson J Shaven 1984 A COmparision of Methodologies in Empirical
General Equilibrium Models of Taxation in Searl H J Shoven Eds ~pQlied Geoaral Jguilibrium
lIoalyss Cambridge University Press Cambridge
Fullerton 0 A King J Shaven and J Whaney 1981 Corporate Tax Integration in the United
Statesmiddot a General Equiiibdufi1 Approach American Economic Review 71 pp~ 677~691
Galper H R Lucke E Tader 1985 Taxation portlor Choice and the Allocation of Capital A
General Equilibrium Approach Mimeo
Gill p W Murray M Saunders M Wright 1986 Users Guide for NPSOL (Version 40) A
Fortran Package for Nonlinear Programming T echnicsl Report SOL 86middot2 Deparlment of
Operations Research Stanford University
Glnsbur~h V and L van der Heyden 1985 General Equilibrium with Wage Rigidities An
45
AppJication to Belgium in A Manne ed t Economic EQumprit1m~ Made Formulation and $oJioo
Mathematical Programming Study 23 Amsterdam NorthmiddotHolland ~
Goulet J 1968 1 Adjustrent CO$1S in the Theory o(the Firm Reyiew of EconomiC Studies 35 bull
pp47middot55
Goulder L 1985 Tax-Financed versus BondmiddotFinancedGhanges in Governent Spending
bull Mimeo Harva(d UniversilYmiddot
Goulder l J Shaven and J Whalley 1983 Domestic Tax Policy and the Foreign Sector The
importance of altemative foreign policy formulations to results from a general equilibrium tax
analysis model In ~n methods in Ibe antico of income from caoia1 ed Martin
Feldstein Chicago University of Chicago riSS
Goulder l L Summers 1987 Tax Policy Asset Prices and 3rowtly ~
A General Equilibrium Analysis NBER Working Pper No 212B
Grandmont J 1982 Temporary General Equilibrium Theory Ch 20 in Handbook 01
Mathemalical Eccnornics Vol II NorthmiddotHoliand
Harberger A 1959 The Corporate Income Tax An empirical Appraisal In lax Revision
Comoendjum Voll House Cornmitee on Ways and Means Washington DC Govemment Printing
Office
Harberger A 1962 The Incidence of the Corporate Income Tax JOULn1 of Political
Economy 70 pp 215middot240
Harberger A 1964 Taxation Resource Aliocalrlon and Welfare fn lb Role of Dir bullbulll and
Indjresl Taxes io tbe Fede[al Reserve s~I~Ill Princeton Princeton Unlveresity Press
James J 1985 New Developments in the Applications of General Equilibrium models to
Economic History in Piggott aand Whalley Eds
Jorgenson D and K Yun 1984 Tax Policy and Capital Allocation Harvard Inslitute of
bull
I
I f
Economic Research WP 1107
Judd K 1965 ShonmiddotRun Analysis of Fiscal Policy in a Simple Perfect Foresight Model
Journal of poJiHcal 5CCJiW1l pp 298middot319 r
Upton D J Potbc J Sachsa nd L Summers 1982 Multiple Shooting in Rational
Expectations Models fuoometrlc 50 1329middot1333
Manne A ed 1985 fconQrnjc Eoullibrium Model FOrIDujaljQo and Solution Mathematical
Pogrammlng Study 23 Amsterdam NorthmiddotHolland
Manne A 1985 Or the Formulation and Solution 01 Economic Equilibrium Models In A
Manne ed ECQOQlli-Jr~ibmiddotum ~tQdel EQlJulailon and Solution Malhematical Programming
Study 23 Amsterdam NorthmiddotHolland
Mansur A and J Whalley 1984 Numerical Specification of Applied General Equilibrium
Models in Scarf H J Shoven Eds Aoplied General Eouillbrium AnaiysectIsect Cambridge Universily
Press Cambridge
Merrill 0 1972 Applications and Extensions of an Algorithm thaI Computes FixedmiddotPoints of
Cenain Upper Semimiddotcominuou Point to Set Mappings PhD Disserlation Unlvel$ity 01 Michigan
Negishi T 1960 Welfar Economics and Exislence I an Equnlbrlum for a Competitive
Economy MelroeCOOQ11ka 12 92middot97
Pereira A 1985 On the Existence and Uniqueness of Optimal Ou1put Path lor CRTS Firms in
Adjustment Costs Technologies Mimeo Stanford University
Pereira A 1985a Consumer Expectations and Ihe COSI 01 UllUty In an Intertemporat
Framework Mlmeo Stanford University
Pereira A 1985b Corporale Tax Integrallon and Government Deficits A Dynamic General
Equilibrium Anaiysls Mlmeo Stanford University
Pereir A 1987 On Ihe Computation of the Users Costs of Stocks Ecooomi (forthcoming)
47
Pereira A bull 1987b Equal Yield Alternatives and Government Oeficlts Mimeof Stanford
University
Pereira A 1987c A Dynamic Applied General ~uHlbrium Model for Tax Policy Evaluation
Comlletamp Docurnent2tion Mimeo University of California San Diego (in progress)
Pereira A 1987d Should the investment Tax Credit be Rei(ltroduced Mlmeo University
of California San Diego (ir progreSs)
Piggott J and J Whalley eds 1985 tJew QeyeQpments in Applied General EQuilibrium
bullenalysjs Cambridge Cambridge University Press
Preckel Pbull 1985 Alternative algorithms for computing economic equilibria In A Manne Ed
Radnor R 19amp3 Equllibrium under Uncertainty In K Arrow M Intrilligaor ods
I ibull
Robinson S 1 g86 Mullisectorai Models of Oeveloping CountrieS a Survey Oivision of
Agricullure and Resources UC Borkeley WP 401
Scarf H 1967 The approximation of fixed points of a continuous mapping ~MM Jouwal of
epplid Mathornallcs 15 pp 1328middot43
Scarf H with T Hansen 1973 00 the Computation of EconomiC Eguilibria New Haven Yale
University Press
Scarf H J Shoven Eds 1984 61l1llied Genera Elujlibrium 6nalyssect Cambridge University
~ Press Cambridge
Shoven J 1976 The Incidence and Efficiency Effects of Taxe on Income from Capital
Jurnal of ramie1 Economy 84 pp 1261-83
Shoven J 1983 Applied General Equilibrium Tax Modemng IMF Slaff PaoerS 30
Shoven J and LB Simon 1986 Share Repurchases and Acquisitions An Analysis of which
Firms Participate NBER Working Paper No 2243
bull
48
IShaven J and J Whalley 1972 ft General Equmbrium Calculation of lhe Effects of
Differential Taxation of Incoma from Capital in the US Jauroe of public Econgmics 1
pp281middot321 (
Shoven J bull and J Whalley 1973 General Equilibrium with Tax$s a computation procedure
and an existence pr00f ReviEhi of Economic Studies 40 pp 475middot490
Shoven J J Whalley 1977 Equal Yield Tax Alternatives A General Equilibrium
Computa~ui0nal Tfchnique J~HJrnal of PUblic EconomIcs 8 pp 211~224 bull
Shaven J J Whalley 1984 Applied GeneralmiddotEquilibrlum MOdels of Taxalion and
International Tr8de An Introduction and Survey Journal of Economic literature 23t pp
1007middot1051
Slemrod J 1980 A General Equilibrium Model of Capital Income Taxation Ph D bullDissertation Harvard Univnrsity
Slemrod J 1983 A General Equilibrium Model of Taxation with Endogenous Financial
Behavior in M Feldstein edbull Behavioral Simulation Methods In Tax Policy Analysis Chicago
UniVerSity of Chicago Press
Stone J 1985 Sequential Optimization and Complementarily Techniques for computing
Economic equilibrium in A Manne Ed
Summers L 1981 Taxation and Corporate Investment a q-Theory Approach BroQkinas
Summers L 1985 Taxation and the Size and ComposHion 01 the Capital Stock An Asset Price
Approach Mlmeo Harvard University
Whalley J 1975 A Geneal Equiloibrium Assessment of the 1973 United Kingdom tax
reform Economica 42 pp 139middot61
1JLTIllOS JIQll~)ltG PAPERS PUBLICADOS bull
I
I I
j
112 70 - LUCENA iJiCIIJo ~ IIEnvironrilent Honitoring and Orqanization Structure T (Setellibro 1987)
nI 71 - L~ENA Diogo A Note on the Representation of Information Stxueture (SeteJObro 1987)
nQ 72 - LUCENA D1CgO Envuorment 111)nitoring and Organization
Structure II (Setelilhro 1987)
ril 73 - AlmmE3 Antonio Pais e CAsPAR Yitor bull Tributacao Inoentivos e InvesUlllentos Analise Qualitativa noYellibro 19S7)
n2 74 - liAlltOSA rose Pedro Opt1lllal lIage R1gidity A SUggestedtlthodology to Teet tb Theory and an Application (Outubro 1987)
nI 75 - CABRAL L 11 B Three lIotes on SYJIIlletric Gemes witt itsYJIIlletric Equilibri (NoveJObro 1987)
n2 76 - LUCENlt D10gOTo Search or Not to Search (Fevereiro1967)
nQ 17 - SA Jorge asconcelos euroI A Theory of SynergyB (Fevereiro1987)
n2 76 - VILARES tanuei Jose Os Be IntenaMios IlIIportados ColIIO Factor de Produao (Julho 1987)
nQ 79 - SA J~rge VasCQncel03 e HoY to Copete and Co~unicate in tature Industrial Products_ (aio 1988)
n2 80 - ROIl Rafael Learning and Capacity Expansion in a New larket Under Uncertainty (Fevereiro 1988)
n2 81 PEREIRA Alfredo llarvilo Survey of Dynamic Computational General EquilibriUlll Hodel for Tax Policy Evaluation (Outubro 1987)
Qualquer in ormacao sabre o 1IIorlltlng Papers ja publlcados sera prestada pelo Secretariado de Apoio aos Docentes podendo os ~e3OSf
ser adquirido5 ne Secyo de Vlndas da Faculdade de EconollLia UNL na Travessa Esteviio Pinto Cltpolide - 1000 LISllOA
Ocober1987
SURVEY OF DYNAMIC COMPUTATIONALGENERAL EQUILIBRIUM MODELS
FOR TAX POLICY EVALUATION()
Alfredo M Pereira
University of California San Diego and
Faculdade de Economia Universidade Nova de Lisboa
and
John B Shoven
Stanford University and NBER t
r) We would like to thank the participants of the NBER Applied General Equilibrium Workshop
Sianford May 1987 the International Symposium on The Social Accounting Matrix Methods and
Applications Naples Italy June 1987 and of the II World Basque Congress Donolla - San
Sebastian SpaIn September 1987 for helpful comments and suggestions The usual disclaimers
apply
bull
1
Survey cl Dynamic Ccmputational General Equilibrium Models
for Tax Pelley Evaluation
1 Introduction
There is a well established and fast growing body of literature focusing on tax policy evaluation
using disaggregated computational general equilibrium (CGE) models That such models have become
so popular Is hardly uPrling In fact the CGE approacih has several advantages over more macro
orienled aggregated models or analytical parllal equilibrium analysis First ttre CGE methodology
allows the study 01 differential impacts acnoss sectors of production andocnossconsumer groups
Second U allows one to consider the int~ractions among different sectors and agents so that the
policy evaluation is not biased by ceteris paribus assumptions Third and in a more technical vein
it makes use of flexible computational numerical techniques Analytical tools often become
intractable for disaggregated models Fur1hermore the CGE modeller does not have to be confined to
small changes in parameters This i an Important feature because large changes In policy
parameters ere often nremplaled in most tax reform proposals
Historically the field was fostered by Ihe early pioneering work of Harberger (1959 1962
9SG) Herberger uses a highly aggregated analytical model to focus on the taxation of capital
Income The first fully dlsaggregated computational general equilibrium model was introduced by
ShavenWhalley (1972) to evaluate the effects of differentiallaxalio of Inceme from capital in the
us
Until the early eighties most empir[cal general equiiibrium work for tax policy evaluation
i I I
I I I
I I r
bull
2
static tnvestment passively actjusts to saving The typical instruments in the US tax system are
modelled and the government budget Is balanced This state of affairs Is discussed and surveyed In
Shovn (1983) Fullerton-Henderson-Shoven (1984 and Shoven-WhaJ)ey (19S4) t I
Integration etc - has Induced Important developments There have been attempts to Incorporate
life-eyce behavior intergenerational issues allocation of savings investment decisions and
adjustment costs corporate financial decisions and government deficits In to the CGE models Underlying all of these features is the need to endow the models with a dynamic structure w~houl I which most of the above issues cannot be adequately addressed
IThe first efforts to incorporate dynamics into the CGe tax models are associated with bull
Fullerton-Shoven-Whalley (1978 middot1981 1993) Consumption decisions incorporale some t I
int~rtemporal aspects and an Intertemporal equilibrium path for the economy Is obtained by Ihe 1
Isequencing of static short-nm equilibria The equHibna are connected through the evolution of the I
capital endowments which_ tn turn depend on saving middotbehavior A compr~1e discussion of a recent I i Iversion of Ihis model is provided in BallardmiddotFulierton-Shoven-Whaney (19SS) I
More recently a whole new generation of models has been developed by Andersson (19S7) Auerbach-Kotlikoff (1983 1984 1987) Ballard (1983) Ballard-Goulder (laeS) Bovenberg I (1984 1985 19B5) ErlichmiddotGinsburgh-Heyden (laB7) Fellonstoin (19B4 ISBS) Goulder I (19BS) Jorgenson-Yun (1984) Pereira (leaSh 19B7e ta87d) and Goulder-Summers
(1987)
tThis article focuse~ on slJch efforts Our scope is narrow By concentrating on decentralized i
I i
Mo~e recently j the ~hrust towards better descrIptive power together with the nature of the tax
policy issues under analysis effects of a ponsumption tax capital taxation corporate tax
3
numerical W~ieral equilibrium modelling designed 10 addresS tax polic issues we abstract from
several ether related areas and approaches First we abstract from analytle models along the lines
of Abel-Blanchard (1983) and Judd (1985) Second we abslract from centralized growth-type
megels of laxation like Charnley (1981 1982) Third we abstract from eentralielti models
I
associated with World Bank researchers See for example Adelman-Robinson (1978) and
Dervis-de Mele-Robinson (1981) Finally we abstract from other CIl fields See for example
ShovenVlcalley (1984) for a sUlVey ollotematlonal trade applications DecaluwmiddotMartens
(1985) ane Robinson (1986) for applications In the area of development James (1985) for a bull I I
survey of economic history applications and A Manne (1986) and BOrges (1988) far all I encompassing surveys
The models included in our comparfsion are I
I 1 8allardFuliertan-ShavenmiddotWhalley (198S) l 2 Andersson (1987)
3 AuerbachmiddotKotlikoff (1963 1987)
4 Bailarcl (1983) and BallardmiddotGould (19aS)
5 Bovenbeg (1985 1985)
6 ErlichmiddotGinsburgh-Heyden (1987)
7 FeltenSleln (1984 1985)
S Goulde (1985)
9 GouldermiddotSummers (1987)
10 Jorgenson Yun (19B4)
11 Pereira (1986b 19870)
bull 4 I
The kei features cf these models afe summarized in Table 1 which provides a structure for the
discussion Most of toe models Ihal we examine could be said to be In the Yale tradition That Is
they wer developped oy sudnts oi Herbert Scarf or students of the students of Herber1 Scar This
shyincludes BalladmiddotFuerM-Shoven-Whalley Andersson Ballard-Goulder ~)Venberg FellenSlein I
ibull Go-cider GouldermiddotSJmmers ard Pereira The exceptions from this heritage are Auerbachshybull
Kotlikoff Erllch-GinsburghHeyden and JorgensonmiddotYun I This survey is organized ltis follows rne second section provides a brief overview of I
Inon-dynamic CGE models The third seolion offers a discussion of modelling economic behavior and
equilibrium The fourth section focuses on model implemenlalion and policy ev~uatlon The fifth
Isection summarrzes empirical evidence en selected policy issues The sixth section illustrates the
power and wea~nesses of dynamic modelling with the issue of corporate laxintegration Finafly the last section summarizes the state~of-the-art in terms of dynamic modelling for lax policy
evalutlon and sugg-ess areas of future research
_
--
--- -~ ----- __-shy
_ _ - _- - Con~um9rs_ 8amphavlor
Bajla rd -Fulflonmiddot -Shaven-Whalley
1985 Anderaaon
1087
Auerba ch~ Ko1Uko 1983 1987
BaHElfdmiddotGouldor 1983 1905
Sovanborg 1985 198e
ErHchmiddotGlnaburghmiddot middotHeydeR
1987 Feltensteln 1984 1966
Gouldar 1985
GouldorSummor 1987
Jotgeneon-Yun 1984
Porelr 1966 1991
Count dynamlc dl9agg~ objecth lastes constraint ondog endog OvorlapBnqvMI of datil
year _seltlng labor no savings gonorst mollves
inlertsmporal 12 Income otility eel IntertempOfll1 yes yo no noLS 1973 Ymiddotmiddot
_ groups max ---_
0 Intertnmporal rllcufsivs
Swaden 1984 yes aggrogate utility CES eq of molion 00 yos na no consumer max for weaHh
tntortorrporal Isoeiastlc~ 00-1983
ill yo 55 poundtUS u~illy middot19831 In tcemporal yesmiddot 11El7 yes yo 00
cohorts rn~x CES-19C7 plauslblo
bYPothlC1 inlsrtamporal
19n yes 55 ago uliJity Isoelastlc Intertemporal ybullbull ye ye ysLS BFSWexpanOOd _ cohorts max
0118 inlertemporal recursivo LS plausible ye aggregate ollllly lsoolastic eq ot motion no yes no no
hYPOlhotic consumer max for weallh jnterlemporlll Boigium 1960 1983 ro aggrogate utility St()ne~Gearyjner1emporal no yo no no
consumer max Inlerl(lmporal two periodmiddot two yes
Austrs1i 1991middot1982 yes 2 income ulllly CD yeluly ybullbull tn the flrsl no no 2-perlod) groups max constraint porlod
Intertamporal LS 1973 ybullbull 55 aga utility eel intarhtmporal ye ybullbull ybullbull y
BFSW bullbullponded cohorts max
intert9mpor~1 W 1973 ves aggregate ullily 180085110 intertemporal no ybullbull no no
BFSW bullbulllMded consumer max (I~gt IntertempQ(a~
W 1955-80 ybullbull aggregate ulilily trenslog intertemporal ybullbull yes no no consumer max
Inlerlemporal recursive 1973LS yes a income ulllily CmiddotD eq of motion yos ybullbull 00 no
BFSWxpanded groups max for wealth
middot5middot
bull-- _---- -- _-----shy~ -~
I AlJlt 1 - uynafllc (lIt MOOero fur all runey tVlIIVllllQlI
dynamic dluggI utllng
I Ba~ll-a rd-fullerton
-Shoven-Vlhalley I no 19 sectors 19S5
Andarsson 1987 ye 2 sectors
~ Auerbach~KotHkof no~1gB3
1983 1967 yesmiddot1g87 1 sector
Ballard~Go uld ar 1983 1985 no 19 sectors
Bovenborg 1965 1986 yes 2 sectors
Erllch~GIn5burg hshy-Heyden no 24 seclors
1957 1__ Feltenataln 1964 196e no 30 sectors
GoUldsr 1985 no 3 sectors
~ GoutdarmiddotSumrnamprs
1987 yes 5 sectors
_ Jorgen9ClIO-Yun ona sector
1984 no produces two goods
Poralr 19S~ 1987 yos 4 sectors
~bJecllve
Produc~rs
technology behavior InpIJ1 1-0
matrix adl costs opt Imal
capital stock
sialic cost min
CES in valu$ added
110
capital labor yes 00
np Invest adjusts
to sFvnns
max 01
interlempora flot rash flow
cnsl min-lOS) Imax of valul)
of the firm~1987
ill
CD-ISBa CES-1987
capital labor
captel labor
no
no
YA quadratic in
I nO-191)3J
qutldratic in Ilk-1gB
ye
YAS
static cost min
CES in vafl1oadded
10
capita labor yes no
no invest edjuss
10 savifQs max ot capita yas ~
value ot the CD labor 00 quadratic In yo Ilrm Ilk
static CD in several no global ltlost mln vaiue added
110 ltsp1al
Jabor Ves no Inves adjusts
to MVrHIS
statle oost min
____ slatio
cost min
max of value of the
firm equivalent 10 two~stage
~ost min me) of
CO in value ad~ed
110 CES In
value added 110
CESin shyvalue added
110 dual
trBRslog price function Cobb-Douglas
capUal labor
cepHal labor
capital labor
capital labor
capital
yes
yes
yes
no
no invost adjusts no to S8vinos net of
El-xog 9011 deflcitsl no
no invest adjusts to savings r
yes quadratic In Y
Ilk
no yo I
yes Intartamporal net cash flow
In V Added 110
labof Y quadratic In I
y bullbull
- 6 -
~ --- ~ ~- -- -~------ --- --- - --~--~ ---~ shy
_
- ---
--
--------
- - - --------
H n v I taxes
olllng allowed debt Ballardfullamprton~
dynamic objective constrfnt tupendlttire (IlleUs endogenous
aU major balanced trade yoarly ybullbull rly endogenous -f ~Shovon WtlitUy US taxes with C(in~Hlt00 pubHc uility balancod cQmposilon no no debt
eiasticllv junel l 1985 rnagt eCP-REV~d1 aU fn3jO[AndtH3Son vearly balanco small opan Gcon J Swodish inturn fa to cr raturn1987 no - inaia dlJn axogonolJs 00 Ini ITnininod poundgtPfUV
-I L)~ in~h~ --~ ~~ ~pound~~---1- ----shy
runrac h~KQllfkofi ai gulD
1963 1937 US m1l3syo - intertnmponl tixogancuamp yo ys cos~d ampConoroy bandmiddotfinanced flon-oppoundIla
B iii lIa fe-Gould 9r nil llltl1gtr balanced 1rndayearly yearly crcoge)Jus
US t)X6S wiih conslanl I
1963 1965 no public utility balanced cOlYlposWon~ no nD debt max budnt ExP~RCV alastlcHy lunct
Bovenberg simp~iriad 1985middotclosod econybullbull rl~ no - balanced exogenous no no debt US laxes 19a6~two counlry1985 1986 bodgel model
Ert leh~G tn aburghmiddot yeally simplified Inlert walt funet Ishorl~run ttads Heyden no balanced exogenous 00 no Belgium
taxes _ desloullbtium flltUensleln
1967 budaet EXP=REV 6-1mprifia-dmlm costs of yoarly yo rast of tho world
1984 1985 producing a ExTax Rev exogenous financed with yo Australian ~s a Ihlrd conSUmer public good money and bonds non-optimal taxes group
Goulder yaarly annual and endogenous aU maJor
laquo1()Sed economy1985 publlc utilily In1ertemporal composillon bull yes yes U~ taxesY rna versions IlEFElQgtfpoundV bond-financed (non-opllmal balanced trademiddotGoulder-Summer yeary yoarly endogonous bull all major
1987 00 public utility balMc composition no no US taxes with constant max budGel EXPREV alasticilY fund Jl)rgenaon~Yun yoarly aU major
no balanced exogenous no1964 US taxes closed economy budGet inteftemporal recurslve endogenousPereira all major
yo public utility eq of molion lovel and yo yos1995 1987 US laxes closed economy rna for debt composition bondmiddot financed oplloll
- 7 shy
bull - -__----- _ - ------ -- ------- _-----
---- ---
------
------
ll-ltiLt 1 uynamlc Lut MOOSS fOf I ax YOtlcy tv8luallon
Inanell ~ets and maf~9ts ojumLrtm- ----_ ----c_~
privata do btl GlvldtHlI pL~lc aliocalQflo rls shy
1rlc ypa expGct assets equity retention dobl 01 savlnq$ promlums path
Sal ra rd fu Barton phys cap1 used 10 buy soquence lt
Shavon-Wha1ley no financial exogenous oxogenous no physicBl 00 bull 01 sIalic 1E myopic 1985 assets capital flqrlil -
Andersson phys cap used 10 buy 1987 no financial no no yo physical no dynamic ~ perfect
assets capital Auerbach~KolUko If phys cap usod to buy
1983 1987 neo financial no no physical cap no dyoamlc ~ perfectYmiddotmiddotnSSf)ts and Qovt debt SailBrd-Goutder phys cap
1983 1985 no financ[al assets-shy -------
Bovenberg phys cap 1985 1986 no iinenclaf
asso1s Ertlch-Glnsburgh~ phys cap1
-Heyden no financlaJ 1987 assefs
FeltoRsteln phys cap 1984 1986 Invostment is
bondmiddotfinanced
used to buy
exogenous exogenous nObull physl~al no dynamic ~ perfllcl capital
used to buy
no no no physical capital
no dynamic ~onlinuos lima
PI perfect
used 10 buy no no no physical no dynamIc PF perfect
capital jllNt) porlod) bull used 10 buy
no no ye physical cap no dynamic PI p4)rfect and govt debt (Iwo Iods) bull
Goulder phys cap used to buy 1995 no financial exogenous exogenous yenbullbull physical cap no PFlfE bull IdynamIc perfoct~assels and am dabt slaUc
- -~ - shyGoulder-Summers phy$ cap
1987 bonds and no e)(oganOU$ no used 10 buy ye dynamic PF perfect _ bullbullultv bonds MullY (exoaenous
Jorgensonyun phys cap 1984 no financial
essels -~ - ------
Pereira phys cap 1986 1967 bonds end
equHy
________
no
8xogenoos
no
exogenous
no
ye
used to buy two type$ of physical
copltal used 10 buy
go dobl bonds equity
no
no
dynamIc
-
dynamlc
~
lE
perfect
fIbullbulllble (nonperlecB
middot6middot
~-- _middot~ ___~w_ - - ___ bullbullbull___ -------- shy~
-----
------
- ---
-
IlflpnUItll tit lun pVII-y PIlUAlIV
q6H1brla oqual fax polley ovnluaU-on efUclency derlvatton Ioorllhm
calibration param~hr computation COmelfed llold tndfcaHH eHecls
Bnl jard~FuU6rtQ n base year back soll1lion steady-sta1e yo inlsrgroup inerseel or Iii
VS transilion lump-sum with Interfemporai 1965
~Shovn Whallmiddotv replication wllh exog Morrills nmamo1ers alQorilhm _~ond~middotsta1$ ears Inc CVor EV
i~noerucn sleady-sfala no agg(ogaled lnlalsectora 1967
base year back sobHon SiMNON package replication wHn oxog valiant of VS Iransilion no Indicamptor lnlertamporal
paramelsrs multiple shoo ling 6~~~l~slat8
AU8rt)ach~Kotllkot steady-slato yo Inlerguna-r a lions
1893 1llH
quatitaUve by replication a$Sium~lkml Gauss-Soidel liS Uln1WOn any rule -lilh 1111ermpOrB
sonsililJ1y I _ ~poundI~Jjnnelill $t(lfdvmiddot~luiA J3djlt15S ~~ nnrilW ~--------
a ail r nimiddotGould or bas9 9tH bac~ solution Ballard~ staadymiddotslalo )05 kltHSclo(a~
1983 1985 VS Iransllinn lump~sum dynalTlk CV inlerlemporal steadY slafe parameters alaorlthm replicaUon+ with oxog Gourder
+stead~stato (ers loc inlareel10rat BCvonborg qunmaiive by dynamic version stoadY-$Igtto yes 1985 1985 YS transition some govt eN Inlortempora
S9rHltivilv OnearizaJian method ropllcaUon assumptionl of Johansons
+steady-sfata paoo palh Erllch-Glnsbu rgh~ actual vorsus no aootagalod inlersectoral2 bftse years back solu1ion variant of Negishis
Heyden replication wIth alltogbull optlmizaUon countorfaclUal no ~odjcator inlottempora 1987 paramlJters aocroach
faltonstaln baso yeat back solution actual versus no aggrampgalfld intersElclornl replication wilh exog Merrills counlerfaclutll no indlclltor middotlntamprlempOull1984 19B6
parameters al90rithm 1981middot1982
Gouldar staady~slala Inlar5ecioralbase year back souUon Variant or 1S lronsilion no dynamic CV inlortempora
stoady state parameters Newlon 1985 feplicalion+ with iUog Fair~Taylorl
+slaildy-sla1e Intargeoorat Gtlulder~Summers base year back solution varIant of stlJadymiddotstate fixed no aggregated fntersectoral
vs trarlsition government Indicator Intertempo1al steady state oatameters Newton
repticaUon+ with exog Fair-T aylQf 1967 +ste-adv-stata bullbullbullndlno financial
Jorgenon-Yun econometric steady-stale Money melfic In1ersecloral
- - 8slimalion vs traosUian no social weUare functton intertampotal1984 +~teadlmiddotstatl)
~ shy
qtJalitaUv8 by NPSCt optimal path ye generalized inlersectorai 1986 1987 Perelr
replication 8S$umptlonl optimizatIon V$ lumpmiddotsum dynamic lntertemporat se~~ttilitf ____Jllgorilhm opllmal pah pen inc CVoclV financial
- 9 shy
bull ~ ___ - _______ - ____ bull___ OM bullbull __ bullbull __~____ ~_
--
- - - --------
--------- - - - --------
---
1nltJ1 bull I ~ 1J11ltI1 v I ~
-
Sa Ita rd-Futlertonmiddot intargration consumption lax ~StHlven~Whall~y VAT capila~ or labor tax
Ch~mqflS1985 Andersson
consumpion tax neutral corporals lax system 1987
it ue rb a c h~ i(at II Sltotf COIlSUft1ptian lax (poundpird ItS hibor Mx~on ufocto of taxation on capital fOfrnJilti1
dvnamio fiscal policv effects af OQvernmanl defioit (3tlard-Goulder
1883 1987
allocts 01 the adoption of a 1ge3 1985 pure consumpHon lax
im~rtmce of cons fCreSClht -Bovenborg analytically orianiad approach
1985 1986 effects ot pure conumption tax capllal income taxation in Opo economies
Erllch-Glnburghshy-Heyden reat wage policios in Belgium
1987 Felter1steln 1984 - theorotical
1985 ~ linnncial crcwdlng out1984 1986
Goulder effects of tax- vs bond-financed 1985 chanQes In pubtlc expondllufe
~~--
Gouidr~SUmmampfI torpora1$ lax cvts reduced 1987 ITO increase 10 gasolioe taxes
Jorgenson-Yun
effects ot several programs of tax reform on the allocation of ~~ptai
1984
-- Purelra
corporate la)( Integration effects cf romiddotinlroducinQ imreslamnl tax credits
1986 1967
10 bull
bull -~- ~- --- _ - ---____--_ - -----_-_ --~ ---- ~
1 1
2 Brief overview of non-dynamic CGE models
The origin of most or the models we survey goes back to the work of Scarf (1967 1973) whlgt
first developed a reliablo algorithm to compute equilibrium price for a ArrowmiddotDebreu economy
HiS algorithm used simpicial Bubdivision techniques and can -be shown to be the computational analog
of the fixed point theorems previously used to prove the existence of equilibrium His technique
could soW a model with an arbitrary number of consumers and commodities as long as all agents
were price takers consumers were subject to budget constraints demands were continuous andmiddot
production did not display increasing returns to scale The alQorithm while guaranteed to converge I
I I I
was relatively slow for prcb1ems involving more than S1If 20 dimensions A maior improvement in
computational speed Woo offered by Merrills algorithm (1972) which used tpe same fundamental
ideagt as Scarfs procedure
Scarfs model (as is true for the standard Arrow-Debreu model) does not include a government
sector ~ neither taxes nor public goocs At one of the mosi promising appHciUions of the new
computational techniqu w~s in Iha area of tax policy evaluation Shoven and Whalley (1973)
exlended the general equilibrium model and computational approach to include a wide rry of taxes
and a governl11ampnt spending plan The original ShavenmiddotWhalley model was static (although the
different commodi~ies CQuid be considered simitar goods available at different datest as In
Arrow-Debrau) and government was aumad to run a balanced budget The data are arranged as a
social accounling matrix The models spacification and calibration is checked by solving it in the
presence of the base set of laxes The result should ba exactly tha initial social accounting matrix
After having pasd this replication check the model is solved for a counterlactua equilibrium In
the presence of a new tux design The result is once again a social accounting matrix The two
equilibria are compared in order to tSS($S the impact of the new tax plan The first uses ot this
bull
1 Z
model for lax policy evauajon were ShovenWhalley (1972) Whalley (1975) and Shoven
(1916)
Completely static models as Shaven-Whalleys are unsatisfactory for many tax reform ISsues
These include corporate lax Integration effects of investment tax credits effects of accelerated
depreciations consumer Or e)penditure tzxes~ importance of saving subsidies like IRAs etc These
are essentialy dynamiC ksues They involve not only the allocation of capital across sectQfsbut
perhaps more Important the capital intensity of the economy But the capital accumulation and
capital reallocation take time and may involve adjustment costs Because of these issues Ballardmiddot
Fulierton~Shoven~Whaloy took the first steps towards developing a dynamic model
In this conlexl the ShovenWhalley model has been extended and implemented for Ihe US
-economy In BuIIBrdmiddotFulienonmiddotShovenmiddotWhalley (1985) While Ihis book completely documents the
mo~el it was used in several publicalions beginning in 1978 Their mode) consists of nineteem
production sectors twelvo households and fifteen consumer goods It includes a very detailed set of
taxes including the federal and state personallncome taxes federal and state corporate income taxes
SOCial Security taxes pro1J0rty taxes unemployment Insurance excise taxes sales taxes etc It
has been calibl att-d to reproduce the 1973 US economy Recentty a version corresponding to the
1983 JS economy has been developed
wThe 8alard-Fullenon-Shoven Whalley model is dynamic In the sense that consumers face a
choice between current consumption and leisure versus future consumption (which can be
purchased via s wlngs) The consumer classes act as if they were maximizing a nested CES utility
function over the domain of a CES aggregate of contempo-raneous consumer goods and leisure and
future consumption subject to their income constraint The parameters of those funcHons
determine the shares of income devo1ed to each commodity to saving and to the purchase of leisure
They also determine the two key elasticities in the models ~ the elasticity of labor supply with
bull
13
respecl to tgte rbullbull1 aoer lax wage rae and the elasticity of saving with respecl 10 the real after lax- rate of retum to capital
In the model consumerS have myopic expectations regarding lulure prices and in particular
regarding the future rale of return 10 capilal Ballard (1983) and BallardmiddotGoulder (1985)
inco~ora1e both perfect foresight and limited foresight into this model Future consumption is
acquired by buying a fixed composition portfolio of real investments that offer an infinite annuity
of returns
The production side of Ihe model is compielely static The mod1 Incorporates a constantmiddot
olasicily 01 substitution betweeen prlmary inputs in production (capital and labor) and fixed I coefficients for intermediate inputs The model distinguIshes between Industrial outputs and
eonsumer goods for the simple reason Ihat tile data are classified differenUy Industrial sectol$ I I involve such categories as forestry and fisherle metal mino9 and publishing and printing while
consumers purchas-3 fUfHi1ure 2ulomoblles and books This fact is recognized in the model by Ibull
incorporatirg a secord stage Of production which converts industrial outputs intO consumer goods
This technology is uStally mod9-led as a flxed~coefficient conversion matrix
The BallardmiddotFINton-ShovenWhalley model assumeS that the private sector finance
marginai ioveslment with the same composition of debt and equity as currently exists in each sector
This is the same arsumrtion that Herberger originally used Investors all hold debt and equity in tile
same proportion Thampreiore this ownership can be aggregated into simply capital ownership For
tax purposes however the separate treatment of debt and equity is taken into account at both the
corporate and persofa Io=vel SimllarlYJ the dividend poIicies on corporale equitY are established
exogenously There are no government bonds in the model since there are no government deficits
The modol is solved for a sequence (as many as 100) of temporary equilibria wltll consumers
allocating income between present and future conumption at each point in time The palh for 1he
14
economy ~ a sel of connected equilibria The connectIon is pra~ided by capital accumulation
Capital accumulation is endogenous and determined by saving The model stans with a sociagt
accOunting mi1rix In the bzse case the economy is assumed to be in a steaqy state growth -path (along
which all relative prices ~fe constant) The madel solves for both the naw steady state growth path
and the transltion 10 it zf~er a policy intervention rhe authors have frequently addressed the
question-of howong rt tKes to effectively settle un10 a new steady S1ate growth path
The dynamics of the model are limited however in that future consumption is collapsed into a
composite commodity Aso the absence of government deficits and the lack ofproductlon dynamics
limits the realism of the BallardmiddotFuliertonmiddotShoven-Whatley model for the analysis of dynamic
policy issues (such as the adoption 01 a consumption tax or Ihe elimination of the investment tax
credit) The models we survey try to Improve upon Ihe Ballard-FlIiertonmiddotShaven-Whalley
dynamic modelling The models Jn~orporate more forward lookTng behavior They improve the
dynamics on the conampumrtkm side of the economy and some of them introduce true 1ynamlc behavior
on the part oI the producers We turn now to the issues inY9lv_ed in developing such dynamic models
15 I I ~3 ~ssues in the modelling Of dynamic behavior
We will 100k firt at the speclficalion of economic behavior of consume(s~ producers
goyemment and Ihe r~sl of Ihe world The modemog of corporale and household financial decisions
I is then addr5sed Finally Ihe eencept of equilibrium is discussed
I
i
I 1
31 Consumers behavior
Early efforts 10 build dynamic features into the economic behavior 01 consumers are due to
Ballard (1963) and Auerbach-Kotllkoff (1983 1984) Their work has been closely followed by
several subsequent authors In faet dynamic household behavior Is the single most pervasive aspect
amohg the models surveyed
All the models Inc-O1orale some loro of lila-cycle behavior Household behavior is dewmloed
by the maximization of an addWVely separable time invariant intertemporal utillty function The
utility lune1ion is defined oyer the domain of the consumption goods In the economy In most of the
models leisure is also an argument in utility so that labor supply is optimally determined
Utility maximi4ation is subject to a lifetime intertemporal budget constraint which equalizes
the presen1 value of consumers income and expenditure More reeentiy in Andersson (1987) bull
Bovenberg (1985) and Pereira (1986b 1987d) the constraint i$ defined as a sequence 01
recursive equations of motion on wealth This has the potential advantage 01 accomodating liquidity
constraints However It should be recognized that in the absence 01 liquidity constraints (ie when
consumers are free to completely borrow against future income) the two specificallons of the
househofd ccnstraint are essentiay equivatent Furthermore tn both versions saving is optimally
determined as a way of translaring wealth intertemporally
1 6
Some lodels now have a sophisticated dynamic specification of hoosehokl In Ballard (1983) -
BallardmiddotGoulder (1985) and Auerbach-Kotlikof (1983 1984 1987) consumers behavio is
embedded in an iterQEneraional setting Each of them has S5 age cohorts simulataneously alive
Som-e measure of jntergener1ional altruism is considered in the form of bequest motives In Ballard
(1983) and Ballard-Gouldr (1985) consumers derive utility directly from bequthing parl of
their wealth Therefore bequests assume 1he form of a scrap value of terminal wealth of a
generation In AuerbachmiddotKotlikof (1983 1984 1987) each generation empathyz bullbull with Ihe
utility of future generations ovar bequethd wealth In addition Ballard (1ge3) includes detailed
demographic poiections wfill the intent of incorporating in the policy analysis the effects of the
bull Post World War tI baby-boom
In the real world household decisions also include the optimat allo~atiOn of saving among
shyalternative physical or financial assets Theories of household portfolio behavior are relatively
Iwen~establshed However they involve uncertainty as a crucial element Accordingly they have
middotnot been incorpora~ed in the context of the determinIstic dynamic models we sutyey here We will 1 Come backt-O this Issue below
32 Producer behavior
The efforts to build dynamic features into the economic behavior of producers are more recellt
and less widely adopted The first attempts are due to Bovnberg (1984 1985) and Summers
(1965) Dynamic behavior has been more fully incorporated In the recent models of Andersson
(1987) Auerbach-Kotlikoff (19B7) GouldermiddotSummers (1987) and Pereira (1986b 1987d)
Part of the reason why production side dinamlcs has been more slowty adopted is the weak
supply of accepted theories referring to the dynamic behavior of the firms In the models referred
I 17
to above gynamlc producion and investment b~havior are middotinduced by the existence of capital
adjustment costs and linked to Tobins q theory Adjuslment costs are designed to capture both the
incomplete mobility of capital aoross industries and installation costs ~e the costs of adjusting
capital towards its cpHmai level)
Adjllstment costs can be conceived as internal to the firm and measured in terms of foregone
outpul along the lines of Lucas (1967) Alternatively adjustment costs can be viewed as actual
external cests incurred together with Ihe purchase costs along the Unes of Gould (1969) With the
exception of Pereira (198Gb 1937d) all of the CGE authors follow the former approach
The firms in the economy maximize their market value as the present discountad value of the
future stream of dividends In Andersson (1981) and Pereira (198Gb 1987d) firms are seen as
maximizing the present discounted value of net cash flow Maximization is constrained by Ihe
adjustmenl cosl lechnology and ~n equation of motion describing Ihe evolution of the capital stock
It should be noted that undereogenous divldendrelention rules the two problems are equivalent II
should also be nOled Ihat a salisfectory economic rational for th bullbullistoo of dMdonds with the
present tax code is missing in the profession (Shoven (1986))
The models with static formulation of producers behavior are characterized by passive
investment behavior Investment merely accomodates to saving in the eccnomy With a dynamic
formulation induced b adjustment costs real investmen decisions are forward looking Investmenl
Is endogenously and optimally determined by the firms A fundamental difference between the
shonmiddotrun in which capital stock is given and longmiddotrun in which Ihe level of capital is allowed to be
optimally determined is emphasized
This extra richness of produc1ion dynamics is not without costs A careful look at the summary
tables will clearly show an inverse relation between the adoptlon of dynamic features in production
and the level of disagreggation of the production side of the economy In fact with production
bull
j
1 6
dynamics $e dimension of the problems is immensly increased Let IS be more specific
In a static framework with constant returns to scale production technology the output level ismiddot
indeterminate The optimal aliocation of inputs can be obtained by cost minimization with any
feasible output level generatillg zero profits Such zero profit conditions are used to solve for the
output prices in terms of the factor prices and hencemiddot to reduce the dlmensionality of the problem
from the -number of commodities and factors to the number of factors Thus even if the model deals
with 30 production sectors the computation of an economic equilibrium can take place using only
the dimensionality of the primary inputs in the economy
Now under certain regularity conditions on the production and adjustment costs technologies
leading to enough concavity of the optimality objective the intertemporal output path for the firm is
endogenously optimally and uniquely determined even with constant relums to scale technologies
(see Pereira (1985 1987a) Accordin~y with adjustment costs in the model optimal profits
will in general be non-zero This result has crucial implications for the computatiQn of equilibrium
in the model~ with production dynamics With adjustment costs no reduction of dimensionality is
possible The curse of dimensionality returns as a binding constraint
The introduction of adjustment costs adds another significant complication to the model With
capital being less that perfectly mobile in the economy different rates of return on capital will
exist in different sectors This is a difficult problem to tackle conceptually in the absence of
uncertainty Also in the real world producer maximization choices include also its choice of
financial ratios (debVequity) and payout rates (dividendretained earnings) These financial
subjects are difficult and much Clf this behavior is still taken as exogenous by the modellers We
will come back to these issues below
middot i
19
bull 33 government behavior
middot
Two issues dominate the modelling of government behavIor First govemment behavior has
Ibeen typically seen in Ihe CGE liIerature for lax poliCY evaluation as constrained by yearly balanced
I budsets (see for example 8allard-Fullrton-S~ovn-Whafley 098Sraquo The analysis of
govemment defidlS and public debl in a CelE context requires a dynamic selling Second the level of
government expenditures is either exegenouty given as in Auerbach-KoUikofi (1984 1987)
6ovenbrg (1984 1985) Feltenstein (1984 1986) and Jorgenson-Yun (1984) or
endogenously (but no optimally) determined by the balanced budget conditions as in
6allard-Fullerton-Shoven-Whalley (1985) Andersson (1987) Erllch-Ginshurgh-Heyden
~
(1987) Gouder (19aS) and Gouder-Summers (1987) In the second case the composition of
bull public expenditures is often optimally determined However the leve~ of government expendi1ures
can only 00 endogenously and optimally determined if the government is seen as an optimizing agent
and is allowed to run deficits
The first attempts to deal wiU the government defiCits in CGE tax models are due to
Auerbach-Kollilltoff (1984) and Feensain (1984) In Auerbach-Kotlikof (1984) government
expenditures afe exogenous They grow at the rate of growth of population However given 1he tax
structure and tax revenues yearly deficits and surpluses are atTowed subject to an in1ertempond
constraint that the present value of future tax revenues equals the present value of future
expenditures
In Fellenstein (1984 1986) government expenditures are also exogenously given He also
allows the government to run deficits to finance expenditures in excess to 1ax revenues However
surpluses are returned to CQnsumars in the form cf transfers Accordingly government ts not
subject to any constraint regarding the future repayment of public debt
In GouJder (19851 the 9Vernmenl maximizes a static social welfare ~nctio subject to a-
bull
20 I i
I
I I I
I bull
balanced budget conSirtlL This follows the optimal alocatkm of government expenditures alpng
the lines of Baliard-FHeiton-Shoven-Whalley (1985) Ttle model Is generallzea to allow for
exogenous changes in HIe time path of government expenditures Two financing alternatives are
consjdeft~d and contr8s1ed EldditionaJ tax revenues and bond issuance
Pereira (198Gb 1137d) auempts to address both the incorporation of deficits and the
determination of govement expenditures The path of government expenditures and the path of
delicilSlsurpluses (ano therefore the path lor debt) are endogenously and optimally determined The
government is seen as mltJximizing n intertemporal social welfare function given the 13x structure
Optimization is subject to a sequence of recursive equations of motion reflecting the evolution of the
ptlbllc debt allowing fDr government budget imbalances It should be stressathat this specification
01 the government cOMtcalnt is equivalent to the specificaUon in Auerbaeh-Kotlikoff (1983 1987)
when no liquidity constraints affect -gove~nment behavior
The extra richness in the treatment of government behavior allows tfte examination of
government debt poices in a truly dynamic setting Also financial crowding out effects induced by
government deficits can be analyzed (see Auerbach-Kottlikoff (1987) Feltenstein (1986) and
Perer (198Gb 1987e))
34 Foreign sector
Most of the open economy models follow the assumptions of balanced trade with import and
export net demands characterized by constant elasticities along the line of Ballardmiddot
-Fullerton-Shaven-Whalley (1965) Such is the case of Salfard (1963) Ballard-Goulder
(19S5) and Goulder-Sllmmers (1987)_ In Feltenstin (1985) the rest of the world is treated as
21
an addjlion~ consumer group
Bovcoberg (1986) develops a model In which two economies lIr considered each following
intertempoml perlect foresight paths These economies meet in the international forum Their
trade relaroMhips are characterized by yearly balanced trade accounts
None of the new generation of dynamic cae tax models has yet IncorPorated the lnternational
capital flows as done in aoulder-Shoven-Whalley (1983) for the earlier Ballarrshy
-Fullertonmiddot Shaven-Whalley (1985) model This is unfortunate as there is an important
Intertmporal aspect to international lending The first al1empts along Ihese lines are due to
Andersson (1987) and Erlich-Glnsburgh-Heyden (1987) Andersson (1987) in his model of the
Swedish economy adopts a closed economy appraoch in which rates of return in the domesJic
economy are largely determined by the international capital markets Fidinterest rate induce
Intlnatlonal capital flows which determine and finance the International lrade imbalance In
lurn In ErlichGinsburghHeydenmiddot (1987) foreign trade Is generaled according to an
intertemporal Irade welfare lunction with constant import and export elasticities In the shortmiddotrun
they allow inlernational trade inbalances which generate capilal ftows to the domestic households In
the long run however trade balance is assumed
35 The need for financial markets
A dynamic economic Wucture not only provides the ideal environment 10 model many features of
economic behavior it also permits one to Incorporate the financial side of the economy If the
government is allowed to run deficits the question 01 deficll financing automatically follows If
investment is optimally dewmined and returns to capital are different aClOSS sectors problems of
investment financing arise If there are seeral final1cfal assets In the economymiddot government
4-
I i
ibull
I I I I
I
j
22
bonds privpte bonds and equity (or simply physical capital installed inmiddot different sectors) the
problem of allocation of saving among assets with potentially different returns arises
The papers surveyed var greatly with respect to l~e extent of their attention to the financial
side of the economy At Clne extreme are the models in ~anardmiddotFuliertonmiddotShoven-Whalley (1985)
Andersson (1987) Ballltld (1983) Ballard-Goulder (1985) Bovenberg (1985 1986) and
Erlich-Gjnsburgh-Heyden C19B which are devoted exclusively to the real side of the economy
In turn Auerbach-KotliKoff (1984 1987) Feltenstein (1984 1986) and Goulde (1985)
allow for government debt In these models saving finances changes in government debt and physical
capital Private and publlc 2ssets are perceived by the households as perfect substitutes The
allocation of saving merely adjusts to the relative demands for funds
Feltenstein (1984 1986) is the only model surveyed which introduces money Government deficits are financed by issuing money and bonds according to an exogenously given rule Money is
demanded by consumers for transaction motives and an exogenously given fraction as a store of value
On the other hand government bonds and physical capital are the vehicles for the intertemporal
transfer of wealth
Summers (1985) and GOlldermiddotSummers (1987) introduce a whole menu of financial assets shy
firm specific equity capital Different assets earn different rates of retum However such rates
are equal up to constant and exogenous sector speCific risk premia Therefore the introduction of
constant exogenous risk premia as helpful as it may be in the context of calibration does not solve
the main issue of the non-optimality of the allocation of saving Also talking about risk premia in a
deterministic context is somewhat unsatisfactory
Pereira (1986b 1987d) also introduces a whole menu of assets private and public bonds and
firm specific equity However all the assets are expected to yield the same rate of return and
therefore perc~jved as perfect substitutes The different asset types allow consideration of
23
exogenOU$~8tUequjty triO dividendlrc~enlion rufes and therefore seve-ral sources of investment
financing bond equity and retained earnings
The nonmiddotcptimalHy of the allocation of saving and the absence or exogenelty of corporate
financial rules is _a reftectlon of the limitations of the determInistic approach Either in a context of
perfect anticipation oi the future prices Of in general within the t$atm of point price
expectations all the papers follow a deterministic approach Under such circumstances consumers
either expect diHerent rales of return (inclusive of risk premium) across assets in which case
they wilt buy only one asset (that with highest rate) Of they expect equal rat~s of relurn in which
case Ihey are indifferent about the asset composition of Iheir portfoliO There is no way of
Iradingmiddotoff rales of return and risks to obtain an optimal interiof solution to the problem of the
allocation of saving
The most advanced contribution in the modelling of saving allocation in a CGE setting is due to
Slemrod (1980 1983) in the context of a static one-period model In his model consumers act
according to a two stage separable decision process They Hrsl decide on how much to save Then
they decide on the allocation of saving according to an Indirect uttlity function dependent on the rates
of ratum and variances offered by the different assets in the eoonomy The sourCe of riskiness in the
ampConory comes from an uncertain marginal product of capital On the other hand aside from
portfolio decisions the res of the economy is insulated from uncertainty
The most complete contribution in terms of the treatment of the corporate financial rules is
FulienonmiddotGordon (1983) In a variant 01 Ihe BailardmiddotFulienon-ShovenmiddotWhelley (1985) model
Fullerton and Gordon have capital intensity and optimal financial decisions Jointly deter~ined
through a two siage process The cosl of financing capital is minimized by trading off the tax
advantages of debt against the epected real bankrupcy coSls inherenl with high debVequity ratios
Given the optimal debtequity ratio the level of inveSlment Is chosen such that at Ihe margin the
24 l
bull
rerurn pn ~uity equas the return on bonds plus an exogenous risk premium
36 Market assumptions equilibrium and expectations
Virtually all of the papers arc characterized by Walrasian market clearing assumptions All
markets are perlect1y competi1ive Atomistic competition among agents 1s assumed even thoughmiddot only
a flnile number of agents is considered Virtually no mark~t disequilibria or price stickiness afe
considered The only exception is ErlichmiddotGinsburghmiddotHeyden (1987) bull In thei model of the Belgium
economy I the _wage rate is fixed in the short-run Therefore in the shOrlwrun desequilibrium in
the labor market will generate endogenous unemployment However in lh$ long-run all prices
Including the wage rata are Ilexible and accordingly ali markets ciear
Given the dynamic nature of behavlor in the economy markelmiddotcfearing prices in each period
depend on expectations of fture prices and on tax variables In the economy There are essentially
two ways at interpreting the economic equilibrium in such a dynamic conte~bull If future prices are
perfectly anticipated (io expectations are self-fulfilling) a perfect foresight equilibrium
prevaHs Then future actions are merely the implementation of current decisions for future
periods However jf ptica expecla110ns are not perfect (Le agents make mistakes w1th re$pect 10
future prices) then a temporary or shortrun equilibrium prevails Markets ciear and clearing
prices depend on fu1ure price expectations Current plans about the future are typically not
precisely implemented They will be revised as more or better information becomes available to the
economic agents See Grandmant (1982) for a comprehensive survey 01 the temporary equilibrium
literature
Wilh the exception of Gould (1985) and Pereira (19S6b 19S7d) all the models surveyed
adopt th~ concept of a perfect foresight equilibrium In turn BaliardmiddotGouldr (1965) considers a
J
2S
flexible am9unt of foresight In terms of the number of years over which price movements are ~
foreseen Pereiras model (1986b 1987b)) is flexible in a somewhat different way in that it can
indude any range of foresight from myopia to perfect foresight The choice between the perfct foresight and lemporary equilibrium is ultimately to be made on
philosophic grounds It can be argued that less than perfect expectations Imply that agents are
irralional in some way (see Auerbach-Kctiikoff (1987 p 10) However the reverse argument can
be made One can question wiether agents are reaHy rationar and perfectly knowledgable about
future prices
Recent evidene of Balird (1987) Ballard-Goulder (19851 Goulder (19851 and Pereira
(1986b 1987d) suggests that the choice in modemng expectations is an important one They show
that the degree of foresight inl0 thc future (ranging from perfect foresight to-myopic expectations)
may have dramatic impacts on the pollcy conclusions of Ihe model Accordingly the best research
strategy may b to design models which are flexible enough 10 allow for different rules regarding the
formation of expectations With the exception of the articles just mentioned sensitivity analysis
bullhave previously not been performed along this dimension
In terms of implementation the two concepts of equilibrium - perfect foresight and temporary
equilibrium have different implications The dimensionality of the equ1ibrium soll1ion algorIthm
is involved Suppose we have a model with ten markets to be run for a period of 50 years Aside
from normalization a perfect foresight model implies computing prices In 500 dimensions while a
-t~mporary equilibrium model requires sorving SO equilibria each in ten dimensions Given that
computational speed often varies with the cube of the number of dimensions the lemporary
equilibrium formulation is potentially strikingly more feasible However Ballard (1987) and
Goulder-Ballard (1985) have devoloped techniques to greatly speed the computation of a perfect
foresight equilibrium
26
j
I
4 Jrnplemenlalion and Issues on policy evaluallon
41 Calibration and equilibrium comparlslon
CGE models at typically parameterized by the use of a calibration procedure Some parameters
are exogenously given However some crucial parameters are determined in such a way that the
model replicates the data for a given base year See MansurmiddotWhalleY 11984) for an eXlensive
discussion of 1~i$ issue
CalibratIon in a dynamic context is generaliy interpreted as requiring two properties First
replication of base year data is required Second the model is parameterized to simulate an
intertemporal balanced growth path when Ihe base policy is maintaineq This Is the approach
followed by Ballard (1983) BallardmiddotGoulder (1985) Goulder (1985) and SummersmiddotGoulde
(1986) It follows the practice of BallardmiddotFuliertonmiddotShovenmiddotWhalley (1985) with their
sequential cquilibrium dynamics
Other authors Auerbach-Kotlikof (1983 1984 1987) Boyenberg (1984 1985 1986)
and Pereira (198Gb 1987d) follow whal we cali a qualitalive calibration The slructural
palametars are exogenously chosen so that lh economy follows a reasonable path into Ihe fulure
ThaI has 10 do wllh the fact thaI given the recursive nature 01 the dynamic economy and aside tom
such structural parameters only iniUa stock values are needed to run these models Given initial
conditions on the stocks of say private wealth capltal and government debt agents will optimize
and Ihereby generate a lirst round of net demands and equilibrium conditions In lurn the
equilibrium prices will determine the evolution of the stock variables into the next perIod
There are several potential problems with calibration in the conlext of dynamic models The
assumption of a steadymiddotstate growth ~ath in the base case can be questioned First while
27 I Ibullbull
steadymiddotstaU is a possibity it certainly is not the only meaningfull solution to dynamic models
Even in the case of a perfect foresight equilibrium the model implies an equilibrium path which i may or may not involve colanced growth The model not the modeller should dictate the nature 01 I themiddotbase case path Second in the context of a temporarY equilibrium path a steady state solutlcn is Inol a likely model QutcOn1tl In fact unlessmiddot expectations are stattc short run behavior consistent
with a steady-state evolution will in general not be generated On the other hand if static I expectations are self~fufjili[1g we have in fact a perfect foresight model Third even the base year
replication requirement may cause problems In Ihe cOntexl of temporary equilibrium Any
calibration parameter would be conditional on expectation rules which is probably an undesirable
feature
The nalur of the two-requirement calibration strategy is very muchin the spirit of the bull
traditional design of the comparision of alternative equilibria comparis_ion between a steady-state
base case on one hand and alternative paths including a transition period and a finat sleadyyenstate on
the other hand Pereira (198Gb 1987d) compare different (not necessarily steadystate)
equilibrium paths The arguments against such a procedure are based on the idea that the impact of
policy changes can be observed most easily since all departures from the steadymiddotstate can be
attributed to the altemalive polioy On the other hand the base case so dEfined as a steadymiddotstate is
consist~nl with previous work in a less dynamtc setting and Iherefore allows a common standard
for comparing model results
42 Computation algorithms
We are still al a stage in which basically each author uses a different computational technique
Th development of dynamic models - in parlicular wilh adjustment costs andior perfect foresight
bull
28
has corresponded with the decline in the use of fixed point algorithms In -fact given the relative
arge dimensions inevitably involved such algorithms lend to be very inefficient at the best and
chen prohibitively stow See Stone (1985) and Preckel (1965) for a comparative assessment of
different computation techniques Among the models surveyed only Ball1rdFulle(ton~
-ShavenmiddotWhalley (1985) and Feltensteln 1985 use Merrills variant of the fixed point
algorithm 1echnique
AuerbachmiddotKoHikoff (1963 1984 1967) follow a three stage procedure They first compute a
base case steady-state then a revised case steady state and finally transition path for the economy
berveen these two steady-states tn all stages a Gauss-Seidel iterative procedure is used
Ballard (1982) Ballard-Goulder (1965) Goulder (HiSS) and Summers-Goulder (1986)
~
use a method developed by Ballard and Goutder which is similar in many awects to the FairmiddotTaylQr bull
(1983) algorl1hm Short-run equilibria are calculated (using Merrills algorithm) parametric on
nrice expeC~lions The model is then iterated to generate self-fulfilling intertemporaf expectations
and the corresponding perfect foresight equilibrium In a relatively similar approach Andersson
1987 uses a simulation program SIMNON developed In the University of Lund This program
can handle two-paint boundary problems In a fashion consistent with the multiple shooting
algcrlthm (see Lipton-Poterba-Sachs-Eummers (1982))
Boyenbergs computational approach (1985 1985) differs from the other models surveyed in
that he relies heavily on analytical techniques Computations are done by using a dynamic version of
Johansons linearization method Being essentially determined by the continuous time nature of the
I I
f
model thiS linearization model has the disadvantage of confining the analysls to infinltesinal changes
around the base case equilibrium (see Bovenberg (1985) p 53) Pereira (1986b 1987d) uses an optimization algorithm NPSOL - developped by
rGillMurray-Saunders-Wrlght (1986)) This algorithm is used 10 compute the sequence of
I
29
Shortmiddotrun temporary equilibrium which makes up the intertemporal equilibrium path The
equilibrium conditions are seen as nonlinear equality constraints in the minimization of an artlcial
objective function The prices are normalized to the unit simplex by an additional linear equality I I 1 bull constraint
II ErlichmiddotGinsburgh-Heyden (1987) follow a unique approach in that they use a variant of the
optimization technique introduced by Negishl (1960) The economic equilibrium can be generated i
as a solution of a mathemallcal program Ihe objective function of which is a weighted sum of the
utility functions of the various agents while tho constraints set consists of the market dearing 1
conditions Ginsburgh-Heyden (1985) have extended Negishis resulllO tho case of downward price
lrigidities I
Finally the paper by Jorgonson-Yun (1984) is also unique in tIlat it is the only I
econometrically estimated model Different blocks for the consumption and production Side of the
-
model are separately estimated to provide the necessary structural parameters i
I The diversity of computation techniques is yet another indicator of the exploratory nature of the
body of literature surveyed in this article
43 Equal yield comparisions
The link between a base case and counteriactual simulations Is usually provided by the concept of
equal yield Shaven-Whaley (1977) discuss the meaning 01 equal yield in a general equilibrium
context when government is not allowed to run defICits Equal yield is interpreted to moan constant
public utility Government base case utility is maintained in the counferfactual experiments With
balanced budgets the concepl of equal yield is unambiguous The new equUibrium prices and the
balanced budge condition will del ermine the minimal expendilure and taxes needed to maintain base
30
case publi~ utility Accordingly in general equal yield is inconsistent with equal nominal tax
revenue Some change in tax revenue Is necessary Different tax replacement schemes ate
considered to assure that enough tax revenue is collected This is the approach essentially followed
by most 01 the papers surveyed Only Feltenstein (1985) Goulder (198) and Jorgenson-Yun
(1984) chose not to follow an equal yield strategy
The question is of how to Interprete lila concept of equal yield when the government is alowed to
fUn deflcLts The optimal leve of expenditure for base ease public trliUty can now be taAinanced
bond financed or financed by a mix of bonds and taxation We have several versions of equal yield In
particular equal yield may now be consistent with equal tax revenUE Furthermore some meaSure
ltgtf financial crowding out effects of government deficns can be inferred from the comparislon of the
several equal yield alternatives These issues are extensively discussed ~Pereira (1987b)
44 Price expectations and the dynamic generalization of compensation
Indicators
The sum of equivalent and compensation variations over households is the most widely used
aggregated measure of efficiency gains or losses In the context of $teady~state eomparisions andor
when complete future markets exist andlor perfect foresighl is assumed there are no difficulties
associated with the use of the standard Hicksian indicators In fact correct future prices are known
in these cases
If expectations are not selfmiddot fulfilling andlor future markels are not available a dynamic
generalization is necessary 8allardmiddotGoulder (1985) provide some steps in that direction by
defining an indicator that accomodates periods far into the future when households do not have
perfect foresight but a steady~state prevails On he other hand this issue is more fundamental in
i
[
I
i I I I I [
I I
bullI
31
the contex~ of a Gner1 lemporary equilibrium ftamawor~ In such Circumstances Pereira
(1986a) dev~lops a cnamlc generalization which is obtained as the present discounted value of a i i
sequence of short-fUn optimal expenditure functions consistent with a base case expected future I stream of IlWities I
32
5 Empirical evidenee from selected policy Issues
Dynamic tax models have generated several important results which escaped Sialic modellers
The following is a seleCpoundd set of issuos whIch stress the marginal benefits of dynam~ modelHng of
economic behavior in innovltlttive areas of analysis The discussion of a consumption tax emphasizes
the benefits of dynamic tusehold behavior and intergenerational aspects The study of the
flliminatlon or the re~intToduction of the investsnemt tax credits is made meaningful by the dynamic
modelling of production betlavor Modelling governmenl in a dynamic conlexts allows the modeller
to study the irrrpact of government deflcits Finally a dynamic framework lets us appreCiate the I
relevance of consumer expectations in terms of the evaluation of policy alternatives
i
i
51 Consumption Tax
t
By Its very nature a consumption tax can not be adequately investigated with a static model bull
Fullerion-Shoven-Whaicy (1983) use Ihe model of the US economy s described in
Ballard-Fullerton-Shoven-Whalley (1935) to evaluate the movement from the currenl US tax Isys1em to a progressive consumption tax Since their modeJ Incorporates a Jaborleisure choice
where leisure is an untaxed commodlty their results reflect the fact that both the consumption tax
and Ihe present system are dslortionary
Concentrating on the rntertemporaf distortions they show that sheltering more saving from lhe
current US income tax could improve econornic efficiency even if marginal tax rate increases are
necessary in order to maintain government fevenue~ At first you have a revenue shortfall
However the economy moves to a higher sustatned growth path and uitimatefy lower tax rates
cangenerate the same revenue path Also wages increase in the long run with this policy The
bull
bullbull ~
33 shyt
i
I
I
present value of welfare gi[lS for a potiey of complete savin9 deduction wilh marginal tate
adjustments (consumption lax) is simulated to be around $500 billion to $600 billion 01 1973
dollars
They also investiga1e th 1ngth of ttme it takes th~ economy to adjust to these policy changes
Roughly they estimate the long run to b thirty years although this figure Is very sensitive to the
specification of the savings elasticity - a crucial parameler in this model
52 Investment T~x Credit
Goulder-$ummers (1 S87) address the impact of eliminating the lnvestmenl Tax Cradit (ITC) on
intersectoral capital formation and on economic growth Their model isYa1ticularly adequate to
addies$ this issu~ in that is postulates a forward looking investment m9we with adjustment costs
They show that the eliminating the ITC causes a reduction in the rale of investment Inveslment is
estimated to fall by about 7 in the short-run and by about 12 in Ihe new long run steady state
In lurn a previous policy anncuncement lowers Ihe overall al1ractiveness of Investment and leads 10
a downward shift of the Investrrent profile On the other hand the combined effect of a revenue
neutral simultaneous efimiraron of the ITC and reduction of the corporate tax rates is a long run
reduction ef the capital Sieck by 35 This pattern suggests that a revenue neutral increament in
both Ihe corporale tax rates and the inveSlment lax credits would be preferable 10 the formula
adopted in Ihe US Tax Reform Act of 1986
Pereira (1987d) addresses the impact of the ITe from Ihe oland point of Ihe curTenllax syslem
under the Tax Reform Act of 1986 Given Ihe concern over the potential depressiVe effects upon
savings and investment of the new tax system in conjunction with deficit reduC1ion mechanisms of
the GrammmiddotRudman type a pertinent question is should Ih lTC be e-introduced In the context
bull
34
-of his mod~1 of the US economy Pereira focuses on the trade-off between the distortion in the
intersectoral allocation of capital induced by the lTC the lower tax revenues it generates and
potential financial crowding-out effects on one hand and the positive effects of lowering the relative
price of new capital goods on the other hand Preliminary results confirm the qualitative results in
Goulder-Summers (1987) suggestting a welfare gain of about 2 of the present value of GNP in
the best scenario of absence of replacement taxes Furthermore preliminary results show an
important time pattern to welfare gains with the average benefits increasing the further you look
into the future This pattern is due to the presence of constraints to intersectoral mobility of capital
and inertia in the adjustment towards the optimal capital levels as reflected by the presence of
adjustment costs
I
I
bullAuerbach-Katlikaff (1gB) in the context of their intergenerational model of the US economy
consider the impact on savings capital formation and interest rates of deficit policies and balanced
budget increases in government consumption They conclude that deficit finance and government
consumption can significa-ntly crowd out capital formation and lower the welfare of future
generations However crowding out from deficit finance is a very slow process because it results
from increased government spending over potentially long horizons Also deficit policies may
substantially influence the longmiddotrun interest rates while leaving the short-run rates essentially
unaffected Finally and in opposition to the central role adjustment costs seem to play in both
53 Financial crowding outeff~cts of government deficits
Goulder-Summers and Pereiras models Auerbach-Kotlikoff suggest that the time path of interest
rates induced by a policy of deficit financing seems to be insensitive to the presence of adjustment
costs
bull
35
54 The role 01 consumer expectatlons
The expectations analysis has uncovered one of the benefits of dynamic modelling
Ballard-Goulder (1985) find that the appeal of adopting a consumption tax depends on the level of
foresight possessed by the consumers Furthermore additional foresight may be welfare worsening
This is a second best type of result This tends 10 occur under policies that lead to capital deepening
and declining rate of return to capital over time To the extent that consumers have more foresight
they will be beller equippod to anticipate the fall in the rental price of capital Consequently if a
saving incentive is enacted people save more with myopia than with perfect foresight Given the
exi~tence of taxes on capita and the discrepancy between the private an~--sccial returns to capital
the greater savings with myopia is beller socially Ballard-Goulder find that the welfare gains
from a consumption tax are reduced by about 10 when we move from myopia to farleet foresight
Therefore the attractivenness of adopting a consumption tax seems fairly [obust across these
foresight specifications
I
bull
36
6 Tbe power and weaknesses 01 dynamic modelling the example 01 corporale
tax Integrallon
The Issue of corporate tax integration allows us to sh~w the stre-ngths and weaknesses or the
dynamic approach Empirical evidence using CGE techniques indicates that depending on the precise
scheme of integration and the 1ax replacement methods integratiOn may have substantial effects In
Ihe work of FuliertonmiddotKingmiddotShoven-Whalley (1980 1981) total integration was found to yield an bull
annual static eHielency gain of $4 to $8 billion in 1973 dollars Simulated dynamic gains may be as
large as $695 billion or abcu 14 of the present value of tulUre consumption and leisure in the
US economy These gains result primarily from interindustry reallocalions of Investment and an
improved inter~emporal allocation of consumption
Mora recent work emphasizes hOw consumers asset portfolio decisions and firms finanCial
decisions affect the efficiency gains from integration Slemrod (1980) focusing on consumers
asset ponfoio deCisions finds static efficiency gains which are about twice -as Iarge as those
repOrted iiy Fulierton-King-ShovenWhalley (1961) FuUrlon-Gurdon (1983) focus on the
firms financial decisions They roporl efficiency gains of 6 of GNP from the elimination of the
tax distor1ions favoring debt However when they eliminate the corporate tax and repiace it with
increaSed persona income taxes addlHonal dls10t1ions are created in the optimal labot~leisure
decisions These distortions tend to dominate the analysis slJ(h that the overall effects of complete
integration are very modest Galpr-LuckmiddotTodr (1986) address simultaneously consumers
asset pOrlfollo decisions and the firms financial decisions They also find very modest efficiency
gains from integralion
The above results are important but may be severely biased There are severa aspects of
economic behavior and modelling crucial for the study of Income tax Integration whiCh have no been
i 37
captured innny of the above papers One aspect is the absence of government deficitsmiddot a balanced
budget is assumed It is true Ihal resource crowding OUI is caplured In these mOdels but financial
crowding out induced by government spending is nOI Second investment is not derived from
optimtztlcn behavior Investment behavior passively accomodates endogenous saving decisions As
a consequence the differential impacts of policies in the incentives to save and to invest are not
captured Aso full capital mobility across secors is assumed with instantaneous capital
adjustmenlS towards optimal levels This assumption rules qut different costs of capital across
seclors and therefore differentiated reactions to tax policies changes Finallyhe modelling of both
gov9rnmerlt deficits and of endogenous real and financial investment decisions necessitates the
~nsideraiion of a dynamilt framework and the introduction of financiaJ assets govemment bonds and
iprivate financial assets A dynamic framework also highlights the efficiency effects of Integration on r
ithe ptimal intertemporal decisions The above models are either sIalic (Slemrods and
GalpermiddotltJckemiddotTode(sj or a dynamic sequ~nce of otherwise Slatic model bullbull I Pereir a (1986b) develops a dynamic applied general equilibrium model bull to study the tbull
efficiency effects of integration on the growth and allocation of investmenl across sectors Special
Iattention is paid to the structural affeelS of resource and financial crowding oul induced by
gove mect spending and deficits and to the real and financial Investment reactions across industries
to both tax integration and flnancial crowding out His mode departs from previous work on income
lax Integraticn and for that matler from most 01 the CGIE literalure for tax policy evaluation in
several direcllons It encompasses an endogenous sequential equilibrium struelure founded on
dynamic behavior with flexible expectations Government deficits are optlmaUy determined
Investment decisions are tward looking and the resull of optimizing behavior Several financial
assetsmiddot public and privatemiddot are considered
Simulation resulls Indicale that the welfare gains from Integration under large deficits are at
32
best lJlodest when measured in terms of the GNP The elimination of the corporate tax and its I replacemenl by increased income lax rates yields long run benefits which are never larger than
2 of the presenl yalue ct telr consumption and leisure onder the previous tax regime and 1
under the current tax regime However the short run w~lare effects of full integration tend 10 be
very low and in in some scenarios even negative This is a flew intertemporar pattern of efficiency Ieffects which reflects an adjJsiment lag in the interindustry investment decisions due to the
existence of costs of adjustment On the of her hand partial integration achieved by excluding I dividends from the corporale lax base syslematically yields negalive efficlencyeffeclS In Ihe long Inm these can be as high as 3 of present value of the future value of consumption and leisure This
is a new second best effect suggesllng Ihat less than complete integration may have perverse I efficiency effects
The dynamic slructure of PerclrBs model snowing for an improved freatment of real investment
declsions and government deficts provides a potential setting for a more accurate measure of the
costs and benefits of inlegration The simulation resullS suggest lower benefits and an intertempora Ibull
pattern of growing benefits
Simulation results also clearly suggest robust negative effects from partial integration How
reliable is this result If dJvidends were deductible from the corporate lax base (as are interest
payments) the preferenlial Ireatmant of debl over equily would be ellminaled Corporallons should
be expecled 10 react by decreasing the optimal deWeqully rallo However corporale financial
gecisions are exogenously set In Pereiras model as in aft of the other models surveyed that go as far
as dealing with the issue
Also withdrawing dividends from Ihe corporate tax base is a way of eliminating Ihe double
taxation of dividends at belh Ihe corporate and personnal Income levels How will the optimal
aliocellon of household saving accomodate that change in Ihe relalive return to corporale equity
39 I
t However hOusehold portfolio decisions are exogenouly set in Pereiras model as in all of the other
models surveyeltj go as far as dealing with the issue
IIt is sate to say that the evaluation 01 the benefits 01 partial Integration as defined above are
sevrey underestima1eO Meanwhile the distortionSoenerated by the increase in the marginal
personal tax rates designed to ralse the revenue foregone by the dividend exclusion are fully
aCCltlunted for A good measure of the costs of integration together With a poor evaluation of the
benefils may very well be the reason why partial integration is Simulated 10 yield neg alive
eHikiency gains
On a different vein as most of Ihe model surveyed here Pereiras is a closed economy model It
is legitimate to wonder how the resullS would change If international capital flows were anowed
IThis almost certainly would affect financial crowding out since a good d~al of the capital inflows
could be used to finance public debt
For these various reasons the results should be treated as preliminary but they strongly I suggosl that the dynamic structure provides valuable new insights into the corporate tax integration SSU2
40
7 Concluding remarks I I I i
What has been acccmplished Ihis far The success of Ihe CGE research in laelding the challenge
of dynamic modelling can not be denied
I iGreat progress in the modelling of household behavior has been achieved Promising
developm~nts in the modeUing of production and government behavior have also been accomplished
Interesting innovations in the concept of equilibrium and computation techniques were discussed
The policy analysis was greatly enriched by considering transitional effects together with longer
run steadymiddotstate equilibrium paths generalized equal yield strategies accomodating different
financial crowding out impacts and generalized dynamic policy evaluation indicators
An important set of issues have been addresed by the models surveyed Traditional issues
focusing on the effects of capital taxation and consumption taxes have been pursued In terms of
capital taxation for example the intertemporal nature of the issue was enhanced by allowing an
optimal evolution of capital stock in the economy and forward looking optimal investment decisions
In Andersson (1987) GouldermiddotSummers (1987) and Pereira (1986b 1987d) this is coupled
with a improved treatment of several tax provisions like investment tax credits and depreciation
allowances In terms of the consumption taxes developments in the specification of Intertemporal
household behavior the introduction of overlapping generations and the modelling of
intergenerational links via bequest motives as well as a closer attention to demographic evolution
provided a much improved economic setting for the understanding of the several aspects of the
problem
Furthermore dynamic modelling has permitted the addressing of a set of new issues Policy
Issues ranging from dynamic tax policy analysis to the evaluation of exogenous changes in
government expenditures to the impact of different methods of financing government expenditure to
4 1
the importance of tinanciai croHrii1g out to 1he relevance of consumorS expectations in tne policy
results have been addressed
Despile all of the progress and in part due to such progress several avenues are wide open for
much neded additional research The following seem to be the most Interesting and promising
areasmiddot First the specincation of liquidity constraints to household behavior may help to obtain a
better understanding of policy chenges that affect directly the interest rates in the economy Second
major attention needs to be paid to the incorporafion of a roraign sector (with open capital markets)
Third some efforts are needed in terms of finding adequate computation techniques now that
dynamic modelling eally makes the curse of dimensionality worse than eVOrbull
The modelling of financial markets is the single most unsatisfactory speet of the dynamic
models surveyed here~ The problems associated with provlding a mo~ adequate treatment of
financial markets and in general endogenous and optimal financial deci$ions bull both at the corporate
and at the personal levelsmiddot have 10 be addressed That necessarily requires Introducing uncertainty
into the models To b adequate this must go well beyond the mere assumption of point expeetalions
with the whole array of modelling implementalion and evaluation problems thai generates
bull I i
42
REFERENCES
Abel A and O Blanchard 1983 An Intertemporal Model of Saving and Interest Econometrica
51 pp 675-92
~ Andersson K 1987 Tar-ation of Capital in Sweden - A General Equilibrium Model Ph D
Dissenation UnIversity of Lund Sweder
Auerbach A L KotlikoH 1983 National Savings Economic Welfare and the Structure of
Taxation in M Feldstein ed Behavioral Simulation Methods in Tax Policy- Analysis Chicago
University of Chicago Press
Auerbach A l Kotlikoff 1984 Social Security and the Economics of Demographic-
Transition in H Aaron and G Burtless eds Retirement and Economic Behavior Washington DC
Brookings Institution
Auerbach A L Kotlikoff 1987 Dynamic Fiscal Policy Cambridge Cambridge University
Press
Auerbach A L Kotlikof and J Skinner 1983 The Efficiency Gains from Dynamic Tax
Reform Internatiooaj Economic Reyiew 24 pp 81middot100
Ballard C 1983 Evaluation of the Consumption Tax with Dynamic General Equilibrium
Models Ph D Dissertation Stanford University
Ballard C 1987 Tax Policy and Consumer Foresight A General Equilibrium Simulation
Study Economic Inquiry 25 pp 267-284
Ballard C and L Goulder 1985 Consumption Taxes ForeSight and Welfare a Computatonal
General Equilibrium Analysis in J Piggott and J Whalley eds New Deyelooments in Apolied
General Equilibrium Analysis Cambridge Cambridge University Press
Ballard C D Fullerton J Shoven and J Whalley 1985 A General Equilibrium Model for
t 43
Tax Eoliey Eyalua~iQO NBER University or Chicago Press
Borges A (1987) bull A Survey of Compula1ional General Equilibnum Models CECD Economic
Slud~ 8
Bovenbig L 19B3 imperfect Capital Mobility in a Perfect Foresight Herberger Model
Mimea UC Br~eley
Bovenberg L 1984 Capitel Accumulation and Capital Immobility Qmiddottheory In a Dynamic General Equilibrium Framework PhD dissertation University of California at Berkeley t
Bovenberg L 19850 Dynamic General Equilibrium Tax Models with Adjustment Costs in A iManne ed bull 1985 EtQnomrc ECUjilbrium Model Formulation and Solution Mathematical
lProgramming Siudy 23 Arnsterdam NorthmiddotHoliand
Bovenberg L t 1986 Capital Income Taxation in Growing Open Eeondmies Journa of Pubtic
Eamp(omlcs 3t pp 347middot376 IBovenberg L ald W Keller 1981 Dynamics and Nonlinearlties in Applied General
Equilibrium Models Internal Report Mlmiddot81middot208 Department for Statistical Methods CBS bull
Vooiburg The Netherlands
Charnley C 1981 The Welfare Cost of C~pltal income Taxation In a Growing Economy
IJournal of Political ECQnQm~ 89 3 pp 468middot496 I
Charnley C 19810 Efficient Tax Reform in a Dynamic Model of General Equilibrium I Mimeo Yale University I Charnley C 1982 On the Optimal Taxation In Stylized Models 01 Inlartemporal General
Equilibrium Mimeo Yale University I Decaluw B and A Manens 1985 Pays en Oeveloppement at Modeles Calculables OEquilibre
General - Une Revue de al litterature empirique Monographie Sene 9 Volume 3 Centre de
Recherche et Developpemont Economlque Unrslt de Montreal
44
bull
I I
I
I
1
Erlich $ V Ginsburgh and L van der Heyden 1987 Wher~ do Real Wage Policies Lead
Belgium bull A General Equilibrium Analysis European Economic Reyiew (forthcoming)
air R and J Taylor 1983 Solution and Maximum Likelihood Estimation of Dynamic
Nonlinear Ratioant Expeclations Models Ecooometrica 51 pp 1169middot1186
Feltenstein A 1984 Money and Bonds In a Disaggregated Ec~nonYin Searl 1-1 J Shoven
Eds Aoolied ~eoeral EqJiliDCiulD tioalYsis Cambridge University Press Cambridge
Feltenstein A 19B6 A Dynamic General Equllibrlu Analysis of Financial Crowding Out
Theory with an Application to Australia Journal of Public Economics 31 No1
Fullerlon D R Gorden 1983 A Reexamination of Tax Distorllons in General Equilibrium
Models in M Feldstein Ed Behayora Simulation Methods in Tax pOlicy Analysis Chicago
University of Chipgo Press
Fvllerlon D J Shaven and J Whalley 1983 Replacing the US Income Tax with a
Progressive Consumption Tax JoulOal of Public ECQnomics 20 pp3middot23
Fullerton D Y Henderson J Shaven 1984 A COmparision of Methodologies in Empirical
General Equilibrium Models of Taxation in Searl H J Shoven Eds ~pQlied Geoaral Jguilibrium
lIoalyss Cambridge University Press Cambridge
Fullerton 0 A King J Shaven and J Whaney 1981 Corporate Tax Integration in the United
Statesmiddot a General Equiiibdufi1 Approach American Economic Review 71 pp~ 677~691
Galper H R Lucke E Tader 1985 Taxation portlor Choice and the Allocation of Capital A
General Equilibrium Approach Mimeo
Gill p W Murray M Saunders M Wright 1986 Users Guide for NPSOL (Version 40) A
Fortran Package for Nonlinear Programming T echnicsl Report SOL 86middot2 Deparlment of
Operations Research Stanford University
Glnsbur~h V and L van der Heyden 1985 General Equilibrium with Wage Rigidities An
45
AppJication to Belgium in A Manne ed t Economic EQumprit1m~ Made Formulation and $oJioo
Mathematical Programming Study 23 Amsterdam NorthmiddotHolland ~
Goulet J 1968 1 Adjustrent CO$1S in the Theory o(the Firm Reyiew of EconomiC Studies 35 bull
pp47middot55
Goulder L 1985 Tax-Financed versus BondmiddotFinancedGhanges in Governent Spending
bull Mimeo Harva(d UniversilYmiddot
Goulder l J Shaven and J Whalley 1983 Domestic Tax Policy and the Foreign Sector The
importance of altemative foreign policy formulations to results from a general equilibrium tax
analysis model In ~n methods in Ibe antico of income from caoia1 ed Martin
Feldstein Chicago University of Chicago riSS
Goulder l L Summers 1987 Tax Policy Asset Prices and 3rowtly ~
A General Equilibrium Analysis NBER Working Pper No 212B
Grandmont J 1982 Temporary General Equilibrium Theory Ch 20 in Handbook 01
Mathemalical Eccnornics Vol II NorthmiddotHoliand
Harberger A 1959 The Corporate Income Tax An empirical Appraisal In lax Revision
Comoendjum Voll House Cornmitee on Ways and Means Washington DC Govemment Printing
Office
Harberger A 1962 The Incidence of the Corporate Income Tax JOULn1 of Political
Economy 70 pp 215middot240
Harberger A 1964 Taxation Resource Aliocalrlon and Welfare fn lb Role of Dir bullbulll and
Indjresl Taxes io tbe Fede[al Reserve s~I~Ill Princeton Princeton Unlveresity Press
James J 1985 New Developments in the Applications of General Equilibrium models to
Economic History in Piggott aand Whalley Eds
Jorgenson D and K Yun 1984 Tax Policy and Capital Allocation Harvard Inslitute of
bull
I
I f
Economic Research WP 1107
Judd K 1965 ShonmiddotRun Analysis of Fiscal Policy in a Simple Perfect Foresight Model
Journal of poJiHcal 5CCJiW1l pp 298middot319 r
Upton D J Potbc J Sachsa nd L Summers 1982 Multiple Shooting in Rational
Expectations Models fuoometrlc 50 1329middot1333
Manne A ed 1985 fconQrnjc Eoullibrium Model FOrIDujaljQo and Solution Mathematical
Pogrammlng Study 23 Amsterdam NorthmiddotHolland
Manne A 1985 Or the Formulation and Solution 01 Economic Equilibrium Models In A
Manne ed ECQOQlli-Jr~ibmiddotum ~tQdel EQlJulailon and Solution Malhematical Programming
Study 23 Amsterdam NorthmiddotHolland
Mansur A and J Whalley 1984 Numerical Specification of Applied General Equilibrium
Models in Scarf H J Shoven Eds Aoplied General Eouillbrium AnaiysectIsect Cambridge Universily
Press Cambridge
Merrill 0 1972 Applications and Extensions of an Algorithm thaI Computes FixedmiddotPoints of
Cenain Upper Semimiddotcominuou Point to Set Mappings PhD Disserlation Unlvel$ity 01 Michigan
Negishi T 1960 Welfar Economics and Exislence I an Equnlbrlum for a Competitive
Economy MelroeCOOQ11ka 12 92middot97
Pereira A 1985 On the Existence and Uniqueness of Optimal Ou1put Path lor CRTS Firms in
Adjustment Costs Technologies Mimeo Stanford University
Pereira A 1985a Consumer Expectations and Ihe COSI 01 UllUty In an Intertemporat
Framework Mlmeo Stanford University
Pereira A 1985b Corporale Tax Integrallon and Government Deficits A Dynamic General
Equilibrium Anaiysls Mlmeo Stanford University
Pereir A 1987 On Ihe Computation of the Users Costs of Stocks Ecooomi (forthcoming)
47
Pereira A bull 1987b Equal Yield Alternatives and Government Oeficlts Mimeof Stanford
University
Pereira A 1987c A Dynamic Applied General ~uHlbrium Model for Tax Policy Evaluation
Comlletamp Docurnent2tion Mimeo University of California San Diego (in progress)
Pereira A 1987d Should the investment Tax Credit be Rei(ltroduced Mlmeo University
of California San Diego (ir progreSs)
Piggott J and J Whalley eds 1985 tJew QeyeQpments in Applied General EQuilibrium
bullenalysjs Cambridge Cambridge University Press
Preckel Pbull 1985 Alternative algorithms for computing economic equilibria In A Manne Ed
Radnor R 19amp3 Equllibrium under Uncertainty In K Arrow M Intrilligaor ods
I ibull
Robinson S 1 g86 Mullisectorai Models of Oeveloping CountrieS a Survey Oivision of
Agricullure and Resources UC Borkeley WP 401
Scarf H 1967 The approximation of fixed points of a continuous mapping ~MM Jouwal of
epplid Mathornallcs 15 pp 1328middot43
Scarf H with T Hansen 1973 00 the Computation of EconomiC Eguilibria New Haven Yale
University Press
Scarf H J Shoven Eds 1984 61l1llied Genera Elujlibrium 6nalyssect Cambridge University
~ Press Cambridge
Shoven J 1976 The Incidence and Efficiency Effects of Taxe on Income from Capital
Jurnal of ramie1 Economy 84 pp 1261-83
Shoven J 1983 Applied General Equilibrium Tax Modemng IMF Slaff PaoerS 30
Shoven J and LB Simon 1986 Share Repurchases and Acquisitions An Analysis of which
Firms Participate NBER Working Paper No 2243
bull
48
IShaven J and J Whalley 1972 ft General Equmbrium Calculation of lhe Effects of
Differential Taxation of Incoma from Capital in the US Jauroe of public Econgmics 1
pp281middot321 (
Shoven J bull and J Whalley 1973 General Equilibrium with Tax$s a computation procedure
and an existence pr00f ReviEhi of Economic Studies 40 pp 475middot490
Shoven J J Whalley 1977 Equal Yield Tax Alternatives A General Equilibrium
Computa~ui0nal Tfchnique J~HJrnal of PUblic EconomIcs 8 pp 211~224 bull
Shaven J J Whalley 1984 Applied GeneralmiddotEquilibrlum MOdels of Taxalion and
International Tr8de An Introduction and Survey Journal of Economic literature 23t pp
1007middot1051
Slemrod J 1980 A General Equilibrium Model of Capital Income Taxation Ph D bullDissertation Harvard Univnrsity
Slemrod J 1983 A General Equilibrium Model of Taxation with Endogenous Financial
Behavior in M Feldstein edbull Behavioral Simulation Methods In Tax Policy Analysis Chicago
UniVerSity of Chicago Press
Stone J 1985 Sequential Optimization and Complementarily Techniques for computing
Economic equilibrium in A Manne Ed
Summers L 1981 Taxation and Corporate Investment a q-Theory Approach BroQkinas
Summers L 1985 Taxation and the Size and ComposHion 01 the Capital Stock An Asset Price
Approach Mlmeo Harvard University
Whalley J 1975 A Geneal Equiloibrium Assessment of the 1973 United Kingdom tax
reform Economica 42 pp 139middot61
1JLTIllOS JIQll~)ltG PAPERS PUBLICADOS bull
I
I I
j
112 70 - LUCENA iJiCIIJo ~ IIEnvironrilent Honitoring and Orqanization Structure T (Setellibro 1987)
nI 71 - L~ENA Diogo A Note on the Representation of Information Stxueture (SeteJObro 1987)
nQ 72 - LUCENA D1CgO Envuorment 111)nitoring and Organization
Structure II (Setelilhro 1987)
ril 73 - AlmmE3 Antonio Pais e CAsPAR Yitor bull Tributacao Inoentivos e InvesUlllentos Analise Qualitativa noYellibro 19S7)
n2 74 - liAlltOSA rose Pedro Opt1lllal lIage R1gidity A SUggestedtlthodology to Teet tb Theory and an Application (Outubro 1987)
nI 75 - CABRAL L 11 B Three lIotes on SYJIIlletric Gemes witt itsYJIIlletric Equilibri (NoveJObro 1987)
n2 76 - LUCENlt D10gOTo Search or Not to Search (Fevereiro1967)
nQ 17 - SA Jorge asconcelos euroI A Theory of SynergyB (Fevereiro1987)
n2 76 - VILARES tanuei Jose Os Be IntenaMios IlIIportados ColIIO Factor de Produao (Julho 1987)
nQ 79 - SA J~rge VasCQncel03 e HoY to Copete and Co~unicate in tature Industrial Products_ (aio 1988)
n2 80 - ROIl Rafael Learning and Capacity Expansion in a New larket Under Uncertainty (Fevereiro 1988)
n2 81 PEREIRA Alfredo llarvilo Survey of Dynamic Computational General EquilibriUlll Hodel for Tax Policy Evaluation (Outubro 1987)
Qualquer in ormacao sabre o 1IIorlltlng Papers ja publlcados sera prestada pelo Secretariado de Apoio aos Docentes podendo os ~e3OSf
ser adquirido5 ne Secyo de Vlndas da Faculdade de EconollLia UNL na Travessa Esteviio Pinto Cltpolide - 1000 LISllOA
bull
1
Survey cl Dynamic Ccmputational General Equilibrium Models
for Tax Pelley Evaluation
1 Introduction
There is a well established and fast growing body of literature focusing on tax policy evaluation
using disaggregated computational general equilibrium (CGE) models That such models have become
so popular Is hardly uPrling In fact the CGE approacih has several advantages over more macro
orienled aggregated models or analytical parllal equilibrium analysis First ttre CGE methodology
allows the study 01 differential impacts acnoss sectors of production andocnossconsumer groups
Second U allows one to consider the int~ractions among different sectors and agents so that the
policy evaluation is not biased by ceteris paribus assumptions Third and in a more technical vein
it makes use of flexible computational numerical techniques Analytical tools often become
intractable for disaggregated models Fur1hermore the CGE modeller does not have to be confined to
small changes in parameters This i an Important feature because large changes In policy
parameters ere often nremplaled in most tax reform proposals
Historically the field was fostered by Ihe early pioneering work of Harberger (1959 1962
9SG) Herberger uses a highly aggregated analytical model to focus on the taxation of capital
Income The first fully dlsaggregated computational general equilibrium model was introduced by
ShavenWhalley (1972) to evaluate the effects of differentiallaxalio of Inceme from capital in the
us
Until the early eighties most empir[cal general equiiibrium work for tax policy evaluation
i I I
I I I
I I r
bull
2
static tnvestment passively actjusts to saving The typical instruments in the US tax system are
modelled and the government budget Is balanced This state of affairs Is discussed and surveyed In
Shovn (1983) Fullerton-Henderson-Shoven (1984 and Shoven-WhaJ)ey (19S4) t I
Integration etc - has Induced Important developments There have been attempts to Incorporate
life-eyce behavior intergenerational issues allocation of savings investment decisions and
adjustment costs corporate financial decisions and government deficits In to the CGE models Underlying all of these features is the need to endow the models with a dynamic structure w~houl I which most of the above issues cannot be adequately addressed
IThe first efforts to incorporate dynamics into the CGe tax models are associated with bull
Fullerton-Shoven-Whalley (1978 middot1981 1993) Consumption decisions incorporale some t I
int~rtemporal aspects and an Intertemporal equilibrium path for the economy Is obtained by Ihe 1
Isequencing of static short-nm equilibria The equHibna are connected through the evolution of the I
capital endowments which_ tn turn depend on saving middotbehavior A compr~1e discussion of a recent I i Iversion of Ihis model is provided in BallardmiddotFulierton-Shoven-Whaney (19SS) I
More recently a whole new generation of models has been developed by Andersson (19S7) Auerbach-Kotlikoff (1983 1984 1987) Ballard (1983) Ballard-Goulder (laeS) Bovenberg I (1984 1985 19B5) ErlichmiddotGinsburgh-Heyden (laB7) Fellonstoin (19B4 ISBS) Goulder I (19BS) Jorgenson-Yun (1984) Pereira (leaSh 19B7e ta87d) and Goulder-Summers
(1987)
tThis article focuse~ on slJch efforts Our scope is narrow By concentrating on decentralized i
I i
Mo~e recently j the ~hrust towards better descrIptive power together with the nature of the tax
policy issues under analysis effects of a ponsumption tax capital taxation corporate tax
3
numerical W~ieral equilibrium modelling designed 10 addresS tax polic issues we abstract from
several ether related areas and approaches First we abstract from analytle models along the lines
of Abel-Blanchard (1983) and Judd (1985) Second we abslract from centralized growth-type
megels of laxation like Charnley (1981 1982) Third we abstract from eentralielti models
I
associated with World Bank researchers See for example Adelman-Robinson (1978) and
Dervis-de Mele-Robinson (1981) Finally we abstract from other CIl fields See for example
ShovenVlcalley (1984) for a sUlVey ollotematlonal trade applications DecaluwmiddotMartens
(1985) ane Robinson (1986) for applications In the area of development James (1985) for a bull I I
survey of economic history applications and A Manne (1986) and BOrges (1988) far all I encompassing surveys
The models included in our comparfsion are I
I 1 8allardFuliertan-ShavenmiddotWhalley (198S) l 2 Andersson (1987)
3 AuerbachmiddotKotlikoff (1963 1987)
4 Bailarcl (1983) and BallardmiddotGould (19aS)
5 Bovenbeg (1985 1985)
6 ErlichmiddotGinsburgh-Heyden (1987)
7 FeltenSleln (1984 1985)
S Goulde (1985)
9 GouldermiddotSummers (1987)
10 Jorgenson Yun (19B4)
11 Pereira (1986b 19870)
bull 4 I
The kei features cf these models afe summarized in Table 1 which provides a structure for the
discussion Most of toe models Ihal we examine could be said to be In the Yale tradition That Is
they wer developped oy sudnts oi Herbert Scarf or students of the students of Herber1 Scar This
shyincludes BalladmiddotFuerM-Shoven-Whalley Andersson Ballard-Goulder ~)Venberg FellenSlein I
ibull Go-cider GouldermiddotSJmmers ard Pereira The exceptions from this heritage are Auerbachshybull
Kotlikoff Erllch-GinsburghHeyden and JorgensonmiddotYun I This survey is organized ltis follows rne second section provides a brief overview of I
Inon-dynamic CGE models The third seolion offers a discussion of modelling economic behavior and
equilibrium The fourth section focuses on model implemenlalion and policy ev~uatlon The fifth
Isection summarrzes empirical evidence en selected policy issues The sixth section illustrates the
power and wea~nesses of dynamic modelling with the issue of corporate laxintegration Finafly the last section summarizes the state~of-the-art in terms of dynamic modelling for lax policy
evalutlon and sugg-ess areas of future research
_
--
--- -~ ----- __-shy
_ _ - _- - Con~um9rs_ 8amphavlor
Bajla rd -Fulflonmiddot -Shaven-Whalley
1985 Anderaaon
1087
Auerba ch~ Ko1Uko 1983 1987
BaHElfdmiddotGouldor 1983 1905
Sovanborg 1985 198e
ErHchmiddotGlnaburghmiddot middotHeydeR
1987 Feltensteln 1984 1966
Gouldar 1985
GouldorSummor 1987
Jotgeneon-Yun 1984
Porelr 1966 1991
Count dynamlc dl9agg~ objecth lastes constraint ondog endog OvorlapBnqvMI of datil
year _seltlng labor no savings gonorst mollves
inlertsmporal 12 Income otility eel IntertempOfll1 yes yo no noLS 1973 Ymiddotmiddot
_ groups max ---_
0 Intertnmporal rllcufsivs
Swaden 1984 yes aggrogate utility CES eq of molion 00 yos na no consumer max for weaHh
tntortorrporal Isoeiastlc~ 00-1983
ill yo 55 poundtUS u~illy middot19831 In tcemporal yesmiddot 11El7 yes yo 00
cohorts rn~x CES-19C7 plauslblo
bYPothlC1 inlsrtamporal
19n yes 55 ago uliJity Isoelastlc Intertemporal ybullbull ye ye ysLS BFSWexpanOOd _ cohorts max
0118 inlertemporal recursivo LS plausible ye aggregate ollllly lsoolastic eq ot motion no yes no no
hYPOlhotic consumer max for weallh jnterlemporlll Boigium 1960 1983 ro aggrogate utility St()ne~Gearyjner1emporal no yo no no
consumer max Inlerl(lmporal two periodmiddot two yes
Austrs1i 1991middot1982 yes 2 income ulllly CD yeluly ybullbull tn the flrsl no no 2-perlod) groups max constraint porlod
Intertamporal LS 1973 ybullbull 55 aga utility eel intarhtmporal ye ybullbull ybullbull y
BFSW bullbullponded cohorts max
intert9mpor~1 W 1973 ves aggregate ullily 180085110 intertemporal no ybullbull no no
BFSW bullbulllMded consumer max (I~gt IntertempQ(a~
W 1955-80 ybullbull aggregate ulilily trenslog intertemporal ybullbull yes no no consumer max
Inlerlemporal recursive 1973LS yes a income ulllily CmiddotD eq of motion yos ybullbull 00 no
BFSWxpanded groups max for wealth
middot5middot
bull-- _---- -- _-----shy~ -~
I AlJlt 1 - uynafllc (lIt MOOero fur all runey tVlIIVllllQlI
dynamic dluggI utllng
I Ba~ll-a rd-fullerton
-Shoven-Vlhalley I no 19 sectors 19S5
Andarsson 1987 ye 2 sectors
~ Auerbach~KotHkof no~1gB3
1983 1967 yesmiddot1g87 1 sector
Ballard~Go uld ar 1983 1985 no 19 sectors
Bovenborg 1965 1986 yes 2 sectors
Erllch~GIn5burg hshy-Heyden no 24 seclors
1957 1__ Feltenataln 1964 196e no 30 sectors
GoUldsr 1985 no 3 sectors
~ GoutdarmiddotSumrnamprs
1987 yes 5 sectors
_ Jorgen9ClIO-Yun ona sector
1984 no produces two goods
Poralr 19S~ 1987 yos 4 sectors
~bJecllve
Produc~rs
technology behavior InpIJ1 1-0
matrix adl costs opt Imal
capital stock
sialic cost min
CES in valu$ added
110
capital labor yes 00
np Invest adjusts
to sFvnns
max 01
interlempora flot rash flow
cnsl min-lOS) Imax of valul)
of the firm~1987
ill
CD-ISBa CES-1987
capital labor
captel labor
no
no
YA quadratic in
I nO-191)3J
qutldratic in Ilk-1gB
ye
YAS
static cost min
CES in vafl1oadded
10
capita labor yes no
no invest edjuss
10 savifQs max ot capita yas ~
value ot the CD labor 00 quadratic In yo Ilrm Ilk
static CD in several no global ltlost mln vaiue added
110 ltsp1al
Jabor Ves no Inves adjusts
to MVrHIS
statle oost min
____ slatio
cost min
max of value of the
firm equivalent 10 two~stage
~ost min me) of
CO in value ad~ed
110 CES In
value added 110
CESin shyvalue added
110 dual
trBRslog price function Cobb-Douglas
capUal labor
cepHal labor
capital labor
capital labor
capital
yes
yes
yes
no
no invost adjusts no to S8vinos net of
El-xog 9011 deflcitsl no
no invest adjusts to savings r
yes quadratic In Y
Ilk
no yo I
yes Intartamporal net cash flow
In V Added 110
labof Y quadratic In I
y bullbull
- 6 -
~ --- ~ ~- -- -~------ --- --- - --~--~ ---~ shy
_
- ---
--
--------
- - - --------
H n v I taxes
olllng allowed debt Ballardfullamprton~
dynamic objective constrfnt tupendlttire (IlleUs endogenous
aU major balanced trade yoarly ybullbull rly endogenous -f ~Shovon WtlitUy US taxes with C(in~Hlt00 pubHc uility balancod cQmposilon no no debt
eiasticllv junel l 1985 rnagt eCP-REV~d1 aU fn3jO[AndtH3Son vearly balanco small opan Gcon J Swodish inturn fa to cr raturn1987 no - inaia dlJn axogonolJs 00 Ini ITnininod poundgtPfUV
-I L)~ in~h~ --~ ~~ ~pound~~---1- ----shy
runrac h~KQllfkofi ai gulD
1963 1937 US m1l3syo - intertnmponl tixogancuamp yo ys cos~d ampConoroy bandmiddotfinanced flon-oppoundIla
B iii lIa fe-Gould 9r nil llltl1gtr balanced 1rndayearly yearly crcoge)Jus
US t)X6S wiih conslanl I
1963 1965 no public utility balanced cOlYlposWon~ no nD debt max budnt ExP~RCV alastlcHy lunct
Bovenberg simp~iriad 1985middotclosod econybullbull rl~ no - balanced exogenous no no debt US laxes 19a6~two counlry1985 1986 bodgel model
Ert leh~G tn aburghmiddot yeally simplified Inlert walt funet Ishorl~run ttads Heyden no balanced exogenous 00 no Belgium
taxes _ desloullbtium flltUensleln
1967 budaet EXP=REV 6-1mprifia-dmlm costs of yoarly yo rast of tho world
1984 1985 producing a ExTax Rev exogenous financed with yo Australian ~s a Ihlrd conSUmer public good money and bonds non-optimal taxes group
Goulder yaarly annual and endogenous aU maJor
laquo1()Sed economy1985 publlc utilily In1ertemporal composillon bull yes yes U~ taxesY rna versions IlEFElQgtfpoundV bond-financed (non-opllmal balanced trademiddotGoulder-Summer yeary yoarly endogonous bull all major
1987 00 public utility balMc composition no no US taxes with constant max budGel EXPREV alasticilY fund Jl)rgenaon~Yun yoarly aU major
no balanced exogenous no1964 US taxes closed economy budGet inteftemporal recurslve endogenousPereira all major
yo public utility eq of molion lovel and yo yos1995 1987 US laxes closed economy rna for debt composition bondmiddot financed oplloll
- 7 shy
bull - -__----- _ - ------ -- ------- _-----
---- ---
------
------
ll-ltiLt 1 uynamlc Lut MOOSS fOf I ax YOtlcy tv8luallon
Inanell ~ets and maf~9ts ojumLrtm- ----_ ----c_~
privata do btl GlvldtHlI pL~lc aliocalQflo rls shy
1rlc ypa expGct assets equity retention dobl 01 savlnq$ promlums path
Sal ra rd fu Barton phys cap1 used 10 buy soquence lt
Shavon-Wha1ley no financial exogenous oxogenous no physicBl 00 bull 01 sIalic 1E myopic 1985 assets capital flqrlil -
Andersson phys cap used 10 buy 1987 no financial no no yo physical no dynamic ~ perfect
assets capital Auerbach~KolUko If phys cap usod to buy
1983 1987 neo financial no no physical cap no dyoamlc ~ perfectYmiddotmiddotnSSf)ts and Qovt debt SailBrd-Goutder phys cap
1983 1985 no financ[al assets-shy -------
Bovenberg phys cap 1985 1986 no iinenclaf
asso1s Ertlch-Glnsburgh~ phys cap1
-Heyden no financlaJ 1987 assefs
FeltoRsteln phys cap 1984 1986 Invostment is
bondmiddotfinanced
used to buy
exogenous exogenous nObull physl~al no dynamic ~ perfllcl capital
used to buy
no no no physical capital
no dynamic ~onlinuos lima
PI perfect
used 10 buy no no no physical no dynamIc PF perfect
capital jllNt) porlod) bull used 10 buy
no no ye physical cap no dynamic PI p4)rfect and govt debt (Iwo Iods) bull
Goulder phys cap used to buy 1995 no financial exogenous exogenous yenbullbull physical cap no PFlfE bull IdynamIc perfoct~assels and am dabt slaUc
- -~ - shyGoulder-Summers phy$ cap
1987 bonds and no e)(oganOU$ no used 10 buy ye dynamic PF perfect _ bullbullultv bonds MullY (exoaenous
Jorgensonyun phys cap 1984 no financial
essels -~ - ------
Pereira phys cap 1986 1967 bonds end
equHy
________
no
8xogenoos
no
exogenous
no
ye
used to buy two type$ of physical
copltal used 10 buy
go dobl bonds equity
no
no
dynamIc
-
dynamlc
~
lE
perfect
fIbullbulllble (nonperlecB
middot6middot
~-- _middot~ ___~w_ - - ___ bullbullbull___ -------- shy~
-----
------
- ---
-
IlflpnUItll tit lun pVII-y PIlUAlIV
q6H1brla oqual fax polley ovnluaU-on efUclency derlvatton Ioorllhm
calibration param~hr computation COmelfed llold tndfcaHH eHecls
Bnl jard~FuU6rtQ n base year back soll1lion steady-sta1e yo inlsrgroup inerseel or Iii
VS transilion lump-sum with Interfemporai 1965
~Shovn Whallmiddotv replication wllh exog Morrills nmamo1ers alQorilhm _~ond~middotsta1$ ears Inc CVor EV
i~noerucn sleady-sfala no agg(ogaled lnlalsectora 1967
base year back sobHon SiMNON package replication wHn oxog valiant of VS Iransilion no Indicamptor lnlertamporal
paramelsrs multiple shoo ling 6~~~l~slat8
AU8rt)ach~Kotllkot steady-slato yo Inlerguna-r a lions
1893 1llH
quatitaUve by replication a$Sium~lkml Gauss-Soidel liS Uln1WOn any rule -lilh 1111ermpOrB
sonsililJ1y I _ ~poundI~Jjnnelill $t(lfdvmiddot~luiA J3djlt15S ~~ nnrilW ~--------
a ail r nimiddotGould or bas9 9tH bac~ solution Ballard~ staadymiddotslalo )05 kltHSclo(a~
1983 1985 VS Iransllinn lump~sum dynalTlk CV inlerlemporal steadY slafe parameters alaorlthm replicaUon+ with oxog Gourder
+stead~stato (ers loc inlareel10rat BCvonborg qunmaiive by dynamic version stoadY-$Igtto yes 1985 1985 YS transition some govt eN Inlortempora
S9rHltivilv OnearizaJian method ropllcaUon assumptionl of Johansons
+steady-sfata paoo palh Erllch-Glnsbu rgh~ actual vorsus no aootagalod inlersectoral2 bftse years back solu1ion variant of Negishis
Heyden replication wIth alltogbull optlmizaUon countorfaclUal no ~odjcator inlottempora 1987 paramlJters aocroach
faltonstaln baso yeat back solution actual versus no aggrampgalfld intersElclornl replication wilh exog Merrills counlerfaclutll no indlclltor middotlntamprlempOull1984 19B6
parameters al90rithm 1981middot1982
Gouldar staady~slala Inlar5ecioralbase year back souUon Variant or 1S lronsilion no dynamic CV inlortempora
stoady state parameters Newlon 1985 feplicalion+ with iUog Fair~Taylorl
+slaildy-sla1e Intargeoorat Gtlulder~Summers base year back solution varIant of stlJadymiddotstate fixed no aggregated fntersectoral
vs trarlsition government Indicator Intertempo1al steady state oatameters Newton
repticaUon+ with exog Fair-T aylQf 1967 +ste-adv-stata bullbullbullndlno financial
Jorgenon-Yun econometric steady-stale Money melfic In1ersecloral
- - 8slimalion vs traosUian no social weUare functton intertampotal1984 +~teadlmiddotstatl)
~ shy
qtJalitaUv8 by NPSCt optimal path ye generalized inlersectorai 1986 1987 Perelr
replication 8S$umptlonl optimizatIon V$ lumpmiddotsum dynamic lntertemporat se~~ttilitf ____Jllgorilhm opllmal pah pen inc CVoclV financial
- 9 shy
bull ~ ___ - _______ - ____ bull___ OM bullbull __ bullbull __~____ ~_
--
- - - --------
--------- - - - --------
---
1nltJ1 bull I ~ 1J11ltI1 v I ~
-
Sa Ita rd-Futlertonmiddot intargration consumption lax ~StHlven~Whall~y VAT capila~ or labor tax
Ch~mqflS1985 Andersson
consumpion tax neutral corporals lax system 1987
it ue rb a c h~ i(at II Sltotf COIlSUft1ptian lax (poundpird ItS hibor Mx~on ufocto of taxation on capital fOfrnJilti1
dvnamio fiscal policv effects af OQvernmanl defioit (3tlard-Goulder
1883 1987
allocts 01 the adoption of a 1ge3 1985 pure consumpHon lax
im~rtmce of cons fCreSClht -Bovenborg analytically orianiad approach
1985 1986 effects ot pure conumption tax capllal income taxation in Opo economies
Erllch-Glnburghshy-Heyden reat wage policios in Belgium
1987 Felter1steln 1984 - theorotical
1985 ~ linnncial crcwdlng out1984 1986
Goulder effects of tax- vs bond-financed 1985 chanQes In pubtlc expondllufe
~~--
Gouidr~SUmmampfI torpora1$ lax cvts reduced 1987 ITO increase 10 gasolioe taxes
Jorgenson-Yun
effects ot several programs of tax reform on the allocation of ~~ptai
1984
-- Purelra
corporate la)( Integration effects cf romiddotinlroducinQ imreslamnl tax credits
1986 1967
10 bull
bull -~- ~- --- _ - ---____--_ - -----_-_ --~ ---- ~
1 1
2 Brief overview of non-dynamic CGE models
The origin of most or the models we survey goes back to the work of Scarf (1967 1973) whlgt
first developed a reliablo algorithm to compute equilibrium price for a ArrowmiddotDebreu economy
HiS algorithm used simpicial Bubdivision techniques and can -be shown to be the computational analog
of the fixed point theorems previously used to prove the existence of equilibrium His technique
could soW a model with an arbitrary number of consumers and commodities as long as all agents
were price takers consumers were subject to budget constraints demands were continuous andmiddot
production did not display increasing returns to scale The alQorithm while guaranteed to converge I
I I I
was relatively slow for prcb1ems involving more than S1If 20 dimensions A maior improvement in
computational speed Woo offered by Merrills algorithm (1972) which used tpe same fundamental
ideagt as Scarfs procedure
Scarfs model (as is true for the standard Arrow-Debreu model) does not include a government
sector ~ neither taxes nor public goocs At one of the mosi promising appHciUions of the new
computational techniqu w~s in Iha area of tax policy evaluation Shoven and Whalley (1973)
exlended the general equilibrium model and computational approach to include a wide rry of taxes
and a governl11ampnt spending plan The original ShavenmiddotWhalley model was static (although the
different commodi~ies CQuid be considered simitar goods available at different datest as In
Arrow-Debrau) and government was aumad to run a balanced budget The data are arranged as a
social accounling matrix The models spacification and calibration is checked by solving it in the
presence of the base set of laxes The result should ba exactly tha initial social accounting matrix
After having pasd this replication check the model is solved for a counterlactua equilibrium In
the presence of a new tux design The result is once again a social accounting matrix The two
equilibria are compared in order to tSS($S the impact of the new tax plan The first uses ot this
bull
1 Z
model for lax policy evauajon were ShovenWhalley (1972) Whalley (1975) and Shoven
(1916)
Completely static models as Shaven-Whalleys are unsatisfactory for many tax reform ISsues
These include corporate lax Integration effects of investment tax credits effects of accelerated
depreciations consumer Or e)penditure tzxes~ importance of saving subsidies like IRAs etc These
are essentialy dynamiC ksues They involve not only the allocation of capital across sectQfsbut
perhaps more Important the capital intensity of the economy But the capital accumulation and
capital reallocation take time and may involve adjustment costs Because of these issues Ballardmiddot
Fulierton~Shoven~Whaloy took the first steps towards developing a dynamic model
In this conlexl the ShovenWhalley model has been extended and implemented for Ihe US
-economy In BuIIBrdmiddotFulienonmiddotShovenmiddotWhalley (1985) While Ihis book completely documents the
mo~el it was used in several publicalions beginning in 1978 Their mode) consists of nineteem
production sectors twelvo households and fifteen consumer goods It includes a very detailed set of
taxes including the federal and state personallncome taxes federal and state corporate income taxes
SOCial Security taxes pro1J0rty taxes unemployment Insurance excise taxes sales taxes etc It
has been calibl att-d to reproduce the 1973 US economy Recentty a version corresponding to the
1983 JS economy has been developed
wThe 8alard-Fullenon-Shoven Whalley model is dynamic In the sense that consumers face a
choice between current consumption and leisure versus future consumption (which can be
purchased via s wlngs) The consumer classes act as if they were maximizing a nested CES utility
function over the domain of a CES aggregate of contempo-raneous consumer goods and leisure and
future consumption subject to their income constraint The parameters of those funcHons
determine the shares of income devo1ed to each commodity to saving and to the purchase of leisure
They also determine the two key elasticities in the models ~ the elasticity of labor supply with
bull
13
respecl to tgte rbullbull1 aoer lax wage rae and the elasticity of saving with respecl 10 the real after lax- rate of retum to capital
In the model consumerS have myopic expectations regarding lulure prices and in particular
regarding the future rale of return 10 capilal Ballard (1983) and BallardmiddotGoulder (1985)
inco~ora1e both perfect foresight and limited foresight into this model Future consumption is
acquired by buying a fixed composition portfolio of real investments that offer an infinite annuity
of returns
The production side of Ihe model is compielely static The mod1 Incorporates a constantmiddot
olasicily 01 substitution betweeen prlmary inputs in production (capital and labor) and fixed I coefficients for intermediate inputs The model distinguIshes between Industrial outputs and
eonsumer goods for the simple reason Ihat tile data are classified differenUy Industrial sectol$ I I involve such categories as forestry and fisherle metal mino9 and publishing and printing while
consumers purchas-3 fUfHi1ure 2ulomoblles and books This fact is recognized in the model by Ibull
incorporatirg a secord stage Of production which converts industrial outputs intO consumer goods
This technology is uStally mod9-led as a flxed~coefficient conversion matrix
The BallardmiddotFINton-ShovenWhalley model assumeS that the private sector finance
marginai ioveslment with the same composition of debt and equity as currently exists in each sector
This is the same arsumrtion that Herberger originally used Investors all hold debt and equity in tile
same proportion Thampreiore this ownership can be aggregated into simply capital ownership For
tax purposes however the separate treatment of debt and equity is taken into account at both the
corporate and persofa Io=vel SimllarlYJ the dividend poIicies on corporale equitY are established
exogenously There are no government bonds in the model since there are no government deficits
The modol is solved for a sequence (as many as 100) of temporary equilibria wltll consumers
allocating income between present and future conumption at each point in time The palh for 1he
14
economy ~ a sel of connected equilibria The connectIon is pra~ided by capital accumulation
Capital accumulation is endogenous and determined by saving The model stans with a sociagt
accOunting mi1rix In the bzse case the economy is assumed to be in a steaqy state growth -path (along
which all relative prices ~fe constant) The madel solves for both the naw steady state growth path
and the transltion 10 it zf~er a policy intervention rhe authors have frequently addressed the
question-of howong rt tKes to effectively settle un10 a new steady S1ate growth path
The dynamics of the model are limited however in that future consumption is collapsed into a
composite commodity Aso the absence of government deficits and the lack ofproductlon dynamics
limits the realism of the BallardmiddotFuliertonmiddotShoven-Whatley model for the analysis of dynamic
policy issues (such as the adoption 01 a consumption tax or Ihe elimination of the investment tax
credit) The models we survey try to Improve upon Ihe Ballard-FlIiertonmiddotShaven-Whalley
dynamic modelling The models Jn~orporate more forward lookTng behavior They improve the
dynamics on the conampumrtkm side of the economy and some of them introduce true 1ynamlc behavior
on the part oI the producers We turn now to the issues inY9lv_ed in developing such dynamic models
15 I I ~3 ~ssues in the modelling Of dynamic behavior
We will 100k firt at the speclficalion of economic behavior of consume(s~ producers
goyemment and Ihe r~sl of Ihe world The modemog of corporale and household financial decisions
I is then addr5sed Finally Ihe eencept of equilibrium is discussed
I
i
I 1
31 Consumers behavior
Early efforts 10 build dynamic features into the economic behavior 01 consumers are due to
Ballard (1963) and Auerbach-Kotllkoff (1983 1984) Their work has been closely followed by
several subsequent authors In faet dynamic household behavior Is the single most pervasive aspect
amohg the models surveyed
All the models Inc-O1orale some loro of lila-cycle behavior Household behavior is dewmloed
by the maximization of an addWVely separable time invariant intertemporal utillty function The
utility lune1ion is defined oyer the domain of the consumption goods In the economy In most of the
models leisure is also an argument in utility so that labor supply is optimally determined
Utility maximi4ation is subject to a lifetime intertemporal budget constraint which equalizes
the presen1 value of consumers income and expenditure More reeentiy in Andersson (1987) bull
Bovenberg (1985) and Pereira (1986b 1987d) the constraint i$ defined as a sequence 01
recursive equations of motion on wealth This has the potential advantage 01 accomodating liquidity
constraints However It should be recognized that in the absence 01 liquidity constraints (ie when
consumers are free to completely borrow against future income) the two specificallons of the
househofd ccnstraint are essentiay equivatent Furthermore tn both versions saving is optimally
determined as a way of translaring wealth intertemporally
1 6
Some lodels now have a sophisticated dynamic specification of hoosehokl In Ballard (1983) -
BallardmiddotGoulder (1985) and Auerbach-Kotlikof (1983 1984 1987) consumers behavio is
embedded in an iterQEneraional setting Each of them has S5 age cohorts simulataneously alive
Som-e measure of jntergener1ional altruism is considered in the form of bequest motives In Ballard
(1983) and Ballard-Gouldr (1985) consumers derive utility directly from bequthing parl of
their wealth Therefore bequests assume 1he form of a scrap value of terminal wealth of a
generation In AuerbachmiddotKotlikof (1983 1984 1987) each generation empathyz bullbull with Ihe
utility of future generations ovar bequethd wealth In addition Ballard (1ge3) includes detailed
demographic poiections wfill the intent of incorporating in the policy analysis the effects of the
bull Post World War tI baby-boom
In the real world household decisions also include the optimat allo~atiOn of saving among
shyalternative physical or financial assets Theories of household portfolio behavior are relatively
Iwen~establshed However they involve uncertainty as a crucial element Accordingly they have
middotnot been incorpora~ed in the context of the determinIstic dynamic models we sutyey here We will 1 Come backt-O this Issue below
32 Producer behavior
The efforts to build dynamic features into the economic behavior of producers are more recellt
and less widely adopted The first attempts are due to Bovnberg (1984 1985) and Summers
(1965) Dynamic behavior has been more fully incorporated In the recent models of Andersson
(1987) Auerbach-Kotlikoff (19B7) GouldermiddotSummers (1987) and Pereira (1986b 1987d)
Part of the reason why production side dinamlcs has been more slowty adopted is the weak
supply of accepted theories referring to the dynamic behavior of the firms In the models referred
I 17
to above gynamlc producion and investment b~havior are middotinduced by the existence of capital
adjustment costs and linked to Tobins q theory Adjuslment costs are designed to capture both the
incomplete mobility of capital aoross industries and installation costs ~e the costs of adjusting
capital towards its cpHmai level)
Adjllstment costs can be conceived as internal to the firm and measured in terms of foregone
outpul along the lines of Lucas (1967) Alternatively adjustment costs can be viewed as actual
external cests incurred together with Ihe purchase costs along the Unes of Gould (1969) With the
exception of Pereira (198Gb 1937d) all of the CGE authors follow the former approach
The firms in the economy maximize their market value as the present discountad value of the
future stream of dividends In Andersson (1981) and Pereira (198Gb 1987d) firms are seen as
maximizing the present discounted value of net cash flow Maximization is constrained by Ihe
adjustmenl cosl lechnology and ~n equation of motion describing Ihe evolution of the capital stock
It should be noted that undereogenous divldendrelention rules the two problems are equivalent II
should also be nOled Ihat a salisfectory economic rational for th bullbullistoo of dMdonds with the
present tax code is missing in the profession (Shoven (1986))
The models with static formulation of producers behavior are characterized by passive
investment behavior Investment merely accomodates to saving in the eccnomy With a dynamic
formulation induced b adjustment costs real investmen decisions are forward looking Investmenl
Is endogenously and optimally determined by the firms A fundamental difference between the
shonmiddotrun in which capital stock is given and longmiddotrun in which Ihe level of capital is allowed to be
optimally determined is emphasized
This extra richness of produc1ion dynamics is not without costs A careful look at the summary
tables will clearly show an inverse relation between the adoptlon of dynamic features in production
and the level of disagreggation of the production side of the economy In fact with production
bull
j
1 6
dynamics $e dimension of the problems is immensly increased Let IS be more specific
In a static framework with constant returns to scale production technology the output level ismiddot
indeterminate The optimal aliocation of inputs can be obtained by cost minimization with any
feasible output level generatillg zero profits Such zero profit conditions are used to solve for the
output prices in terms of the factor prices and hencemiddot to reduce the dlmensionality of the problem
from the -number of commodities and factors to the number of factors Thus even if the model deals
with 30 production sectors the computation of an economic equilibrium can take place using only
the dimensionality of the primary inputs in the economy
Now under certain regularity conditions on the production and adjustment costs technologies
leading to enough concavity of the optimality objective the intertemporal output path for the firm is
endogenously optimally and uniquely determined even with constant relums to scale technologies
(see Pereira (1985 1987a) Accordin~y with adjustment costs in the model optimal profits
will in general be non-zero This result has crucial implications for the computatiQn of equilibrium
in the model~ with production dynamics With adjustment costs no reduction of dimensionality is
possible The curse of dimensionality returns as a binding constraint
The introduction of adjustment costs adds another significant complication to the model With
capital being less that perfectly mobile in the economy different rates of return on capital will
exist in different sectors This is a difficult problem to tackle conceptually in the absence of
uncertainty Also in the real world producer maximization choices include also its choice of
financial ratios (debVequity) and payout rates (dividendretained earnings) These financial
subjects are difficult and much Clf this behavior is still taken as exogenous by the modellers We
will come back to these issues below
middot i
19
bull 33 government behavior
middot
Two issues dominate the modelling of government behavIor First govemment behavior has
Ibeen typically seen in Ihe CGE liIerature for lax poliCY evaluation as constrained by yearly balanced
I budsets (see for example 8allard-Fullrton-S~ovn-Whafley 098Sraquo The analysis of
govemment defidlS and public debl in a CelE context requires a dynamic selling Second the level of
government expenditures is either exegenouty given as in Auerbach-KoUikofi (1984 1987)
6ovenbrg (1984 1985) Feltenstein (1984 1986) and Jorgenson-Yun (1984) or
endogenously (but no optimally) determined by the balanced budget conditions as in
6allard-Fullerton-Shoven-Whalley (1985) Andersson (1987) Erllch-Ginshurgh-Heyden
~
(1987) Gouder (19aS) and Gouder-Summers (1987) In the second case the composition of
bull public expenditures is often optimally determined However the leve~ of government expendi1ures
can only 00 endogenously and optimally determined if the government is seen as an optimizing agent
and is allowed to run deficits
The first attempts to deal wiU the government defiCits in CGE tax models are due to
Auerbach-Kollilltoff (1984) and Feensain (1984) In Auerbach-Kotlikof (1984) government
expenditures afe exogenous They grow at the rate of growth of population However given 1he tax
structure and tax revenues yearly deficits and surpluses are atTowed subject to an in1ertempond
constraint that the present value of future tax revenues equals the present value of future
expenditures
In Fellenstein (1984 1986) government expenditures are also exogenously given He also
allows the government to run deficits to finance expenditures in excess to 1ax revenues However
surpluses are returned to CQnsumars in the form cf transfers Accordingly government ts not
subject to any constraint regarding the future repayment of public debt
In GouJder (19851 the 9Vernmenl maximizes a static social welfare ~nctio subject to a-
bull
20 I i
I
I I I
I bull
balanced budget conSirtlL This follows the optimal alocatkm of government expenditures alpng
the lines of Baliard-FHeiton-Shoven-Whalley (1985) Ttle model Is generallzea to allow for
exogenous changes in HIe time path of government expenditures Two financing alternatives are
consjdeft~d and contr8s1ed EldditionaJ tax revenues and bond issuance
Pereira (198Gb 1137d) auempts to address both the incorporation of deficits and the
determination of govement expenditures The path of government expenditures and the path of
delicilSlsurpluses (ano therefore the path lor debt) are endogenously and optimally determined The
government is seen as mltJximizing n intertemporal social welfare function given the 13x structure
Optimization is subject to a sequence of recursive equations of motion reflecting the evolution of the
ptlbllc debt allowing fDr government budget imbalances It should be stressathat this specification
01 the government cOMtcalnt is equivalent to the specificaUon in Auerbaeh-Kotlikoff (1983 1987)
when no liquidity constraints affect -gove~nment behavior
The extra richness in the treatment of government behavior allows tfte examination of
government debt poices in a truly dynamic setting Also financial crowding out effects induced by
government deficits can be analyzed (see Auerbach-Kottlikoff (1987) Feltenstein (1986) and
Perer (198Gb 1987e))
34 Foreign sector
Most of the open economy models follow the assumptions of balanced trade with import and
export net demands characterized by constant elasticities along the line of Ballardmiddot
-Fullerton-Shaven-Whalley (1965) Such is the case of Salfard (1963) Ballard-Goulder
(19S5) and Goulder-Sllmmers (1987)_ In Feltenstin (1985) the rest of the world is treated as
21
an addjlion~ consumer group
Bovcoberg (1986) develops a model In which two economies lIr considered each following
intertempoml perlect foresight paths These economies meet in the international forum Their
trade relaroMhips are characterized by yearly balanced trade accounts
None of the new generation of dynamic cae tax models has yet IncorPorated the lnternational
capital flows as done in aoulder-Shoven-Whalley (1983) for the earlier Ballarrshy
-Fullertonmiddot Shaven-Whalley (1985) model This is unfortunate as there is an important
Intertmporal aspect to international lending The first al1empts along Ihese lines are due to
Andersson (1987) and Erlich-Glnsburgh-Heyden (1987) Andersson (1987) in his model of the
Swedish economy adopts a closed economy appraoch in which rates of return in the domesJic
economy are largely determined by the international capital markets Fidinterest rate induce
Intlnatlonal capital flows which determine and finance the International lrade imbalance In
lurn In ErlichGinsburghHeydenmiddot (1987) foreign trade Is generaled according to an
intertemporal Irade welfare lunction with constant import and export elasticities In the shortmiddotrun
they allow inlernational trade inbalances which generate capilal ftows to the domestic households In
the long run however trade balance is assumed
35 The need for financial markets
A dynamic economic Wucture not only provides the ideal environment 10 model many features of
economic behavior it also permits one to Incorporate the financial side of the economy If the
government is allowed to run deficits the question 01 deficll financing automatically follows If
investment is optimally dewmined and returns to capital are different aClOSS sectors problems of
investment financing arise If there are seeral final1cfal assets In the economymiddot government
4-
I i
ibull
I I I I
I
j
22
bonds privpte bonds and equity (or simply physical capital installed inmiddot different sectors) the
problem of allocation of saving among assets with potentially different returns arises
The papers surveyed var greatly with respect to l~e extent of their attention to the financial
side of the economy At Clne extreme are the models in ~anardmiddotFuliertonmiddotShoven-Whalley (1985)
Andersson (1987) Ballltld (1983) Ballard-Goulder (1985) Bovenberg (1985 1986) and
Erlich-Gjnsburgh-Heyden C19B which are devoted exclusively to the real side of the economy
In turn Auerbach-KotliKoff (1984 1987) Feltenstein (1984 1986) and Goulde (1985)
allow for government debt In these models saving finances changes in government debt and physical
capital Private and publlc 2ssets are perceived by the households as perfect substitutes The
allocation of saving merely adjusts to the relative demands for funds
Feltenstein (1984 1986) is the only model surveyed which introduces money Government deficits are financed by issuing money and bonds according to an exogenously given rule Money is
demanded by consumers for transaction motives and an exogenously given fraction as a store of value
On the other hand government bonds and physical capital are the vehicles for the intertemporal
transfer of wealth
Summers (1985) and GOlldermiddotSummers (1987) introduce a whole menu of financial assets shy
firm specific equity capital Different assets earn different rates of retum However such rates
are equal up to constant and exogenous sector speCific risk premia Therefore the introduction of
constant exogenous risk premia as helpful as it may be in the context of calibration does not solve
the main issue of the non-optimality of the allocation of saving Also talking about risk premia in a
deterministic context is somewhat unsatisfactory
Pereira (1986b 1987d) also introduces a whole menu of assets private and public bonds and
firm specific equity However all the assets are expected to yield the same rate of return and
therefore perc~jved as perfect substitutes The different asset types allow consideration of
23
exogenOU$~8tUequjty triO dividendlrc~enlion rufes and therefore seve-ral sources of investment
financing bond equity and retained earnings
The nonmiddotcptimalHy of the allocation of saving and the absence or exogenelty of corporate
financial rules is _a reftectlon of the limitations of the determInistic approach Either in a context of
perfect anticipation oi the future prices Of in general within the t$atm of point price
expectations all the papers follow a deterministic approach Under such circumstances consumers
either expect diHerent rales of return (inclusive of risk premium) across assets in which case
they wilt buy only one asset (that with highest rate) Of they expect equal rat~s of relurn in which
case Ihey are indifferent about the asset composition of Iheir portfoliO There is no way of
Iradingmiddotoff rales of return and risks to obtain an optimal interiof solution to the problem of the
allocation of saving
The most advanced contribution in the modelling of saving allocation in a CGE setting is due to
Slemrod (1980 1983) in the context of a static one-period model In his model consumers act
according to a two stage separable decision process They Hrsl decide on how much to save Then
they decide on the allocation of saving according to an Indirect uttlity function dependent on the rates
of ratum and variances offered by the different assets in the eoonomy The sourCe of riskiness in the
ampConory comes from an uncertain marginal product of capital On the other hand aside from
portfolio decisions the res of the economy is insulated from uncertainty
The most complete contribution in terms of the treatment of the corporate financial rules is
FulienonmiddotGordon (1983) In a variant 01 Ihe BailardmiddotFulienon-ShovenmiddotWhelley (1985) model
Fullerton and Gordon have capital intensity and optimal financial decisions Jointly deter~ined
through a two siage process The cosl of financing capital is minimized by trading off the tax
advantages of debt against the epected real bankrupcy coSls inherenl with high debVequity ratios
Given the optimal debtequity ratio the level of inveSlment Is chosen such that at Ihe margin the
24 l
bull
rerurn pn ~uity equas the return on bonds plus an exogenous risk premium
36 Market assumptions equilibrium and expectations
Virtually all of the papers arc characterized by Walrasian market clearing assumptions All
markets are perlect1y competi1ive Atomistic competition among agents 1s assumed even thoughmiddot only
a flnile number of agents is considered Virtually no mark~t disequilibria or price stickiness afe
considered The only exception is ErlichmiddotGinsburghmiddotHeyden (1987) bull In thei model of the Belgium
economy I the _wage rate is fixed in the short-run Therefore in the shOrlwrun desequilibrium in
the labor market will generate endogenous unemployment However in lh$ long-run all prices
Including the wage rata are Ilexible and accordingly ali markets ciear
Given the dynamic nature of behavlor in the economy markelmiddotcfearing prices in each period
depend on expectations of fture prices and on tax variables In the economy There are essentially
two ways at interpreting the economic equilibrium in such a dynamic conte~bull If future prices are
perfectly anticipated (io expectations are self-fulfilling) a perfect foresight equilibrium
prevaHs Then future actions are merely the implementation of current decisions for future
periods However jf ptica expecla110ns are not perfect (Le agents make mistakes w1th re$pect 10
future prices) then a temporary or shortrun equilibrium prevails Markets ciear and clearing
prices depend on fu1ure price expectations Current plans about the future are typically not
precisely implemented They will be revised as more or better information becomes available to the
economic agents See Grandmant (1982) for a comprehensive survey 01 the temporary equilibrium
literature
Wilh the exception of Gould (1985) and Pereira (19S6b 19S7d) all the models surveyed
adopt th~ concept of a perfect foresight equilibrium In turn BaliardmiddotGouldr (1965) considers a
J
2S
flexible am9unt of foresight In terms of the number of years over which price movements are ~
foreseen Pereiras model (1986b 1987b)) is flexible in a somewhat different way in that it can
indude any range of foresight from myopia to perfect foresight The choice between the perfct foresight and lemporary equilibrium is ultimately to be made on
philosophic grounds It can be argued that less than perfect expectations Imply that agents are
irralional in some way (see Auerbach-Kctiikoff (1987 p 10) However the reverse argument can
be made One can question wiether agents are reaHy rationar and perfectly knowledgable about
future prices
Recent evidene of Balird (1987) Ballard-Goulder (19851 Goulder (19851 and Pereira
(1986b 1987d) suggests that the choice in modemng expectations is an important one They show
that the degree of foresight inl0 thc future (ranging from perfect foresight to-myopic expectations)
may have dramatic impacts on the pollcy conclusions of Ihe model Accordingly the best research
strategy may b to design models which are flexible enough 10 allow for different rules regarding the
formation of expectations With the exception of the articles just mentioned sensitivity analysis
bullhave previously not been performed along this dimension
In terms of implementation the two concepts of equilibrium - perfect foresight and temporary
equilibrium have different implications The dimensionality of the equ1ibrium soll1ion algorIthm
is involved Suppose we have a model with ten markets to be run for a period of 50 years Aside
from normalization a perfect foresight model implies computing prices In 500 dimensions while a
-t~mporary equilibrium model requires sorving SO equilibria each in ten dimensions Given that
computational speed often varies with the cube of the number of dimensions the lemporary
equilibrium formulation is potentially strikingly more feasible However Ballard (1987) and
Goulder-Ballard (1985) have devoloped techniques to greatly speed the computation of a perfect
foresight equilibrium
26
j
I
4 Jrnplemenlalion and Issues on policy evaluallon
41 Calibration and equilibrium comparlslon
CGE models at typically parameterized by the use of a calibration procedure Some parameters
are exogenously given However some crucial parameters are determined in such a way that the
model replicates the data for a given base year See MansurmiddotWhalleY 11984) for an eXlensive
discussion of 1~i$ issue
CalibratIon in a dynamic context is generaliy interpreted as requiring two properties First
replication of base year data is required Second the model is parameterized to simulate an
intertemporal balanced growth path when Ihe base policy is maintaineq This Is the approach
followed by Ballard (1983) BallardmiddotGoulder (1985) Goulder (1985) and SummersmiddotGoulde
(1986) It follows the practice of BallardmiddotFuliertonmiddotShovenmiddotWhalley (1985) with their
sequential cquilibrium dynamics
Other authors Auerbach-Kotlikof (1983 1984 1987) Boyenberg (1984 1985 1986)
and Pereira (198Gb 1987d) follow whal we cali a qualitalive calibration The slructural
palametars are exogenously chosen so that lh economy follows a reasonable path into Ihe fulure
ThaI has 10 do wllh the fact thaI given the recursive nature 01 the dynamic economy and aside tom
such structural parameters only iniUa stock values are needed to run these models Given initial
conditions on the stocks of say private wealth capltal and government debt agents will optimize
and Ihereby generate a lirst round of net demands and equilibrium conditions In lurn the
equilibrium prices will determine the evolution of the stock variables into the next perIod
There are several potential problems with calibration in the conlext of dynamic models The
assumption of a steadymiddotstate growth ~ath in the base case can be questioned First while
27 I Ibullbull
steadymiddotstaU is a possibity it certainly is not the only meaningfull solution to dynamic models
Even in the case of a perfect foresight equilibrium the model implies an equilibrium path which i may or may not involve colanced growth The model not the modeller should dictate the nature 01 I themiddotbase case path Second in the context of a temporarY equilibrium path a steady state solutlcn is Inol a likely model QutcOn1tl In fact unlessmiddot expectations are stattc short run behavior consistent
with a steady-state evolution will in general not be generated On the other hand if static I expectations are self~fufjili[1g we have in fact a perfect foresight model Third even the base year
replication requirement may cause problems In Ihe cOntexl of temporary equilibrium Any
calibration parameter would be conditional on expectation rules which is probably an undesirable
feature
The nalur of the two-requirement calibration strategy is very muchin the spirit of the bull
traditional design of the comparision of alternative equilibria comparis_ion between a steady-state
base case on one hand and alternative paths including a transition period and a finat sleadyyenstate on
the other hand Pereira (198Gb 1987d) compare different (not necessarily steadystate)
equilibrium paths The arguments against such a procedure are based on the idea that the impact of
policy changes can be observed most easily since all departures from the steadymiddotstate can be
attributed to the altemalive polioy On the other hand the base case so dEfined as a steadymiddotstate is
consist~nl with previous work in a less dynamtc setting and Iherefore allows a common standard
for comparing model results
42 Computation algorithms
We are still al a stage in which basically each author uses a different computational technique
Th development of dynamic models - in parlicular wilh adjustment costs andior perfect foresight
bull
28
has corresponded with the decline in the use of fixed point algorithms In -fact given the relative
arge dimensions inevitably involved such algorithms lend to be very inefficient at the best and
chen prohibitively stow See Stone (1985) and Preckel (1965) for a comparative assessment of
different computation techniques Among the models surveyed only Ball1rdFulle(ton~
-ShavenmiddotWhalley (1985) and Feltensteln 1985 use Merrills variant of the fixed point
algorithm 1echnique
AuerbachmiddotKoHikoff (1963 1984 1967) follow a three stage procedure They first compute a
base case steady-state then a revised case steady state and finally transition path for the economy
berveen these two steady-states tn all stages a Gauss-Seidel iterative procedure is used
Ballard (1982) Ballard-Goulder (1965) Goulder (HiSS) and Summers-Goulder (1986)
~
use a method developed by Ballard and Goutder which is similar in many awects to the FairmiddotTaylQr bull
(1983) algorl1hm Short-run equilibria are calculated (using Merrills algorithm) parametric on
nrice expeC~lions The model is then iterated to generate self-fulfilling intertemporaf expectations
and the corresponding perfect foresight equilibrium In a relatively similar approach Andersson
1987 uses a simulation program SIMNON developed In the University of Lund This program
can handle two-paint boundary problems In a fashion consistent with the multiple shooting
algcrlthm (see Lipton-Poterba-Sachs-Eummers (1982))
Boyenbergs computational approach (1985 1985) differs from the other models surveyed in
that he relies heavily on analytical techniques Computations are done by using a dynamic version of
Johansons linearization method Being essentially determined by the continuous time nature of the
I I
f
model thiS linearization model has the disadvantage of confining the analysls to infinltesinal changes
around the base case equilibrium (see Bovenberg (1985) p 53) Pereira (1986b 1987d) uses an optimization algorithm NPSOL - developped by
rGillMurray-Saunders-Wrlght (1986)) This algorithm is used 10 compute the sequence of
I
29
Shortmiddotrun temporary equilibrium which makes up the intertemporal equilibrium path The
equilibrium conditions are seen as nonlinear equality constraints in the minimization of an artlcial
objective function The prices are normalized to the unit simplex by an additional linear equality I I 1 bull constraint
II ErlichmiddotGinsburgh-Heyden (1987) follow a unique approach in that they use a variant of the
optimization technique introduced by Negishl (1960) The economic equilibrium can be generated i
as a solution of a mathemallcal program Ihe objective function of which is a weighted sum of the
utility functions of the various agents while tho constraints set consists of the market dearing 1
conditions Ginsburgh-Heyden (1985) have extended Negishis resulllO tho case of downward price
lrigidities I
Finally the paper by Jorgonson-Yun (1984) is also unique in tIlat it is the only I
econometrically estimated model Different blocks for the consumption and production Side of the
-
model are separately estimated to provide the necessary structural parameters i
I The diversity of computation techniques is yet another indicator of the exploratory nature of the
body of literature surveyed in this article
43 Equal yield comparisions
The link between a base case and counteriactual simulations Is usually provided by the concept of
equal yield Shaven-Whaley (1977) discuss the meaning 01 equal yield in a general equilibrium
context when government is not allowed to run defICits Equal yield is interpreted to moan constant
public utility Government base case utility is maintained in the counferfactual experiments With
balanced budgets the concepl of equal yield is unambiguous The new equUibrium prices and the
balanced budge condition will del ermine the minimal expendilure and taxes needed to maintain base
30
case publi~ utility Accordingly in general equal yield is inconsistent with equal nominal tax
revenue Some change in tax revenue Is necessary Different tax replacement schemes ate
considered to assure that enough tax revenue is collected This is the approach essentially followed
by most 01 the papers surveyed Only Feltenstein (1985) Goulder (198) and Jorgenson-Yun
(1984) chose not to follow an equal yield strategy
The question is of how to Interprete lila concept of equal yield when the government is alowed to
fUn deflcLts The optimal leve of expenditure for base ease public trliUty can now be taAinanced
bond financed or financed by a mix of bonds and taxation We have several versions of equal yield In
particular equal yield may now be consistent with equal tax revenUE Furthermore some meaSure
ltgtf financial crowding out effects of government deficns can be inferred from the comparislon of the
several equal yield alternatives These issues are extensively discussed ~Pereira (1987b)
44 Price expectations and the dynamic generalization of compensation
Indicators
The sum of equivalent and compensation variations over households is the most widely used
aggregated measure of efficiency gains or losses In the context of $teady~state eomparisions andor
when complete future markets exist andlor perfect foresighl is assumed there are no difficulties
associated with the use of the standard Hicksian indicators In fact correct future prices are known
in these cases
If expectations are not selfmiddot fulfilling andlor future markels are not available a dynamic
generalization is necessary 8allardmiddotGoulder (1985) provide some steps in that direction by
defining an indicator that accomodates periods far into the future when households do not have
perfect foresight but a steady~state prevails On he other hand this issue is more fundamental in
i
[
I
i I I I I [
I I
bullI
31
the contex~ of a Gner1 lemporary equilibrium ftamawor~ In such Circumstances Pereira
(1986a) dev~lops a cnamlc generalization which is obtained as the present discounted value of a i i
sequence of short-fUn optimal expenditure functions consistent with a base case expected future I stream of IlWities I
32
5 Empirical evidenee from selected policy Issues
Dynamic tax models have generated several important results which escaped Sialic modellers
The following is a seleCpoundd set of issuos whIch stress the marginal benefits of dynam~ modelHng of
economic behavior in innovltlttive areas of analysis The discussion of a consumption tax emphasizes
the benefits of dynamic tusehold behavior and intergenerational aspects The study of the
flliminatlon or the re~intToduction of the investsnemt tax credits is made meaningful by the dynamic
modelling of production betlavor Modelling governmenl in a dynamic conlexts allows the modeller
to study the irrrpact of government deflcits Finally a dynamic framework lets us appreCiate the I
relevance of consumer expectations in terms of the evaluation of policy alternatives
i
i
51 Consumption Tax
t
By Its very nature a consumption tax can not be adequately investigated with a static model bull
Fullerion-Shoven-Whaicy (1983) use Ihe model of the US economy s described in
Ballard-Fullerton-Shoven-Whalley (1935) to evaluate the movement from the currenl US tax Isys1em to a progressive consumption tax Since their modeJ Incorporates a Jaborleisure choice
where leisure is an untaxed commodlty their results reflect the fact that both the consumption tax
and Ihe present system are dslortionary
Concentrating on the rntertemporaf distortions they show that sheltering more saving from lhe
current US income tax could improve econornic efficiency even if marginal tax rate increases are
necessary in order to maintain government fevenue~ At first you have a revenue shortfall
However the economy moves to a higher sustatned growth path and uitimatefy lower tax rates
cangenerate the same revenue path Also wages increase in the long run with this policy The
bull
bullbull ~
33 shyt
i
I
I
present value of welfare gi[lS for a potiey of complete savin9 deduction wilh marginal tate
adjustments (consumption lax) is simulated to be around $500 billion to $600 billion 01 1973
dollars
They also investiga1e th 1ngth of ttme it takes th~ economy to adjust to these policy changes
Roughly they estimate the long run to b thirty years although this figure Is very sensitive to the
specification of the savings elasticity - a crucial parameler in this model
52 Investment T~x Credit
Goulder-$ummers (1 S87) address the impact of eliminating the lnvestmenl Tax Cradit (ITC) on
intersectoral capital formation and on economic growth Their model isYa1ticularly adequate to
addies$ this issu~ in that is postulates a forward looking investment m9we with adjustment costs
They show that the eliminating the ITC causes a reduction in the rale of investment Inveslment is
estimated to fall by about 7 in the short-run and by about 12 in Ihe new long run steady state
In lurn a previous policy anncuncement lowers Ihe overall al1ractiveness of Investment and leads 10
a downward shift of the Investrrent profile On the other hand the combined effect of a revenue
neutral simultaneous efimiraron of the ITC and reduction of the corporate tax rates is a long run
reduction ef the capital Sieck by 35 This pattern suggests that a revenue neutral increament in
both Ihe corporale tax rates and the inveSlment lax credits would be preferable 10 the formula
adopted in Ihe US Tax Reform Act of 1986
Pereira (1987d) addresses the impact of the ITe from Ihe oland point of Ihe curTenllax syslem
under the Tax Reform Act of 1986 Given Ihe concern over the potential depressiVe effects upon
savings and investment of the new tax system in conjunction with deficit reduC1ion mechanisms of
the GrammmiddotRudman type a pertinent question is should Ih lTC be e-introduced In the context
bull
34
-of his mod~1 of the US economy Pereira focuses on the trade-off between the distortion in the
intersectoral allocation of capital induced by the lTC the lower tax revenues it generates and
potential financial crowding-out effects on one hand and the positive effects of lowering the relative
price of new capital goods on the other hand Preliminary results confirm the qualitative results in
Goulder-Summers (1987) suggestting a welfare gain of about 2 of the present value of GNP in
the best scenario of absence of replacement taxes Furthermore preliminary results show an
important time pattern to welfare gains with the average benefits increasing the further you look
into the future This pattern is due to the presence of constraints to intersectoral mobility of capital
and inertia in the adjustment towards the optimal capital levels as reflected by the presence of
adjustment costs
I
I
bullAuerbach-Katlikaff (1gB) in the context of their intergenerational model of the US economy
consider the impact on savings capital formation and interest rates of deficit policies and balanced
budget increases in government consumption They conclude that deficit finance and government
consumption can significa-ntly crowd out capital formation and lower the welfare of future
generations However crowding out from deficit finance is a very slow process because it results
from increased government spending over potentially long horizons Also deficit policies may
substantially influence the longmiddotrun interest rates while leaving the short-run rates essentially
unaffected Finally and in opposition to the central role adjustment costs seem to play in both
53 Financial crowding outeff~cts of government deficits
Goulder-Summers and Pereiras models Auerbach-Kotlikoff suggest that the time path of interest
rates induced by a policy of deficit financing seems to be insensitive to the presence of adjustment
costs
bull
35
54 The role 01 consumer expectatlons
The expectations analysis has uncovered one of the benefits of dynamic modelling
Ballard-Goulder (1985) find that the appeal of adopting a consumption tax depends on the level of
foresight possessed by the consumers Furthermore additional foresight may be welfare worsening
This is a second best type of result This tends 10 occur under policies that lead to capital deepening
and declining rate of return to capital over time To the extent that consumers have more foresight
they will be beller equippod to anticipate the fall in the rental price of capital Consequently if a
saving incentive is enacted people save more with myopia than with perfect foresight Given the
exi~tence of taxes on capita and the discrepancy between the private an~--sccial returns to capital
the greater savings with myopia is beller socially Ballard-Goulder find that the welfare gains
from a consumption tax are reduced by about 10 when we move from myopia to farleet foresight
Therefore the attractivenness of adopting a consumption tax seems fairly [obust across these
foresight specifications
I
bull
36
6 Tbe power and weaknesses 01 dynamic modelling the example 01 corporale
tax Integrallon
The Issue of corporate tax integration allows us to sh~w the stre-ngths and weaknesses or the
dynamic approach Empirical evidence using CGE techniques indicates that depending on the precise
scheme of integration and the 1ax replacement methods integratiOn may have substantial effects In
Ihe work of FuliertonmiddotKingmiddotShoven-Whalley (1980 1981) total integration was found to yield an bull
annual static eHielency gain of $4 to $8 billion in 1973 dollars Simulated dynamic gains may be as
large as $695 billion or abcu 14 of the present value of tulUre consumption and leisure in the
US economy These gains result primarily from interindustry reallocalions of Investment and an
improved inter~emporal allocation of consumption
Mora recent work emphasizes hOw consumers asset portfolio decisions and firms finanCial
decisions affect the efficiency gains from integration Slemrod (1980) focusing on consumers
asset ponfoio deCisions finds static efficiency gains which are about twice -as Iarge as those
repOrted iiy Fulierton-King-ShovenWhalley (1961) FuUrlon-Gurdon (1983) focus on the
firms financial decisions They roporl efficiency gains of 6 of GNP from the elimination of the
tax distor1ions favoring debt However when they eliminate the corporate tax and repiace it with
increaSed persona income taxes addlHonal dls10t1ions are created in the optimal labot~leisure
decisions These distortions tend to dominate the analysis slJ(h that the overall effects of complete
integration are very modest Galpr-LuckmiddotTodr (1986) address simultaneously consumers
asset pOrlfollo decisions and the firms financial decisions They also find very modest efficiency
gains from integralion
The above results are important but may be severely biased There are severa aspects of
economic behavior and modelling crucial for the study of Income tax Integration whiCh have no been
i 37
captured innny of the above papers One aspect is the absence of government deficitsmiddot a balanced
budget is assumed It is true Ihal resource crowding OUI is caplured In these mOdels but financial
crowding out induced by government spending is nOI Second investment is not derived from
optimtztlcn behavior Investment behavior passively accomodates endogenous saving decisions As
a consequence the differential impacts of policies in the incentives to save and to invest are not
captured Aso full capital mobility across secors is assumed with instantaneous capital
adjustmenlS towards optimal levels This assumption rules qut different costs of capital across
seclors and therefore differentiated reactions to tax policies changes Finallyhe modelling of both
gov9rnmerlt deficits and of endogenous real and financial investment decisions necessitates the
~nsideraiion of a dynamilt framework and the introduction of financiaJ assets govemment bonds and
iprivate financial assets A dynamic framework also highlights the efficiency effects of Integration on r
ithe ptimal intertemporal decisions The above models are either sIalic (Slemrods and
GalpermiddotltJckemiddotTode(sj or a dynamic sequ~nce of otherwise Slatic model bullbull I Pereir a (1986b) develops a dynamic applied general equilibrium model bull to study the tbull
efficiency effects of integration on the growth and allocation of investmenl across sectors Special
Iattention is paid to the structural affeelS of resource and financial crowding oul induced by
gove mect spending and deficits and to the real and financial Investment reactions across industries
to both tax integration and flnancial crowding out His mode departs from previous work on income
lax Integraticn and for that matler from most 01 the CGIE literalure for tax policy evaluation in
several direcllons It encompasses an endogenous sequential equilibrium struelure founded on
dynamic behavior with flexible expectations Government deficits are optlmaUy determined
Investment decisions are tward looking and the resull of optimizing behavior Several financial
assetsmiddot public and privatemiddot are considered
Simulation resulls Indicale that the welfare gains from Integration under large deficits are at
32
best lJlodest when measured in terms of the GNP The elimination of the corporate tax and its I replacemenl by increased income lax rates yields long run benefits which are never larger than
2 of the presenl yalue ct telr consumption and leisure onder the previous tax regime and 1
under the current tax regime However the short run w~lare effects of full integration tend 10 be
very low and in in some scenarios even negative This is a flew intertemporar pattern of efficiency Ieffects which reflects an adjJsiment lag in the interindustry investment decisions due to the
existence of costs of adjustment On the of her hand partial integration achieved by excluding I dividends from the corporale lax base syslematically yields negalive efficlencyeffeclS In Ihe long Inm these can be as high as 3 of present value of the future value of consumption and leisure This
is a new second best effect suggesllng Ihat less than complete integration may have perverse I efficiency effects
The dynamic slructure of PerclrBs model snowing for an improved freatment of real investment
declsions and government deficts provides a potential setting for a more accurate measure of the
costs and benefits of inlegration The simulation resullS suggest lower benefits and an intertempora Ibull
pattern of growing benefits
Simulation results also clearly suggest robust negative effects from partial integration How
reliable is this result If dJvidends were deductible from the corporate lax base (as are interest
payments) the preferenlial Ireatmant of debl over equily would be ellminaled Corporallons should
be expecled 10 react by decreasing the optimal deWeqully rallo However corporale financial
gecisions are exogenously set In Pereiras model as in aft of the other models surveyed that go as far
as dealing with the issue
Also withdrawing dividends from Ihe corporate tax base is a way of eliminating Ihe double
taxation of dividends at belh Ihe corporate and personnal Income levels How will the optimal
aliocellon of household saving accomodate that change in Ihe relalive return to corporale equity
39 I
t However hOusehold portfolio decisions are exogenouly set in Pereiras model as in all of the other
models surveyeltj go as far as dealing with the issue
IIt is sate to say that the evaluation 01 the benefits 01 partial Integration as defined above are
sevrey underestima1eO Meanwhile the distortionSoenerated by the increase in the marginal
personal tax rates designed to ralse the revenue foregone by the dividend exclusion are fully
aCCltlunted for A good measure of the costs of integration together With a poor evaluation of the
benefils may very well be the reason why partial integration is Simulated 10 yield neg alive
eHikiency gains
On a different vein as most of Ihe model surveyed here Pereiras is a closed economy model It
is legitimate to wonder how the resullS would change If international capital flows were anowed
IThis almost certainly would affect financial crowding out since a good d~al of the capital inflows
could be used to finance public debt
For these various reasons the results should be treated as preliminary but they strongly I suggosl that the dynamic structure provides valuable new insights into the corporate tax integration SSU2
40
7 Concluding remarks I I I i
What has been acccmplished Ihis far The success of Ihe CGE research in laelding the challenge
of dynamic modelling can not be denied
I iGreat progress in the modelling of household behavior has been achieved Promising
developm~nts in the modeUing of production and government behavior have also been accomplished
Interesting innovations in the concept of equilibrium and computation techniques were discussed
The policy analysis was greatly enriched by considering transitional effects together with longer
run steadymiddotstate equilibrium paths generalized equal yield strategies accomodating different
financial crowding out impacts and generalized dynamic policy evaluation indicators
An important set of issues have been addresed by the models surveyed Traditional issues
focusing on the effects of capital taxation and consumption taxes have been pursued In terms of
capital taxation for example the intertemporal nature of the issue was enhanced by allowing an
optimal evolution of capital stock in the economy and forward looking optimal investment decisions
In Andersson (1987) GouldermiddotSummers (1987) and Pereira (1986b 1987d) this is coupled
with a improved treatment of several tax provisions like investment tax credits and depreciation
allowances In terms of the consumption taxes developments in the specification of Intertemporal
household behavior the introduction of overlapping generations and the modelling of
intergenerational links via bequest motives as well as a closer attention to demographic evolution
provided a much improved economic setting for the understanding of the several aspects of the
problem
Furthermore dynamic modelling has permitted the addressing of a set of new issues Policy
Issues ranging from dynamic tax policy analysis to the evaluation of exogenous changes in
government expenditures to the impact of different methods of financing government expenditure to
4 1
the importance of tinanciai croHrii1g out to 1he relevance of consumorS expectations in tne policy
results have been addressed
Despile all of the progress and in part due to such progress several avenues are wide open for
much neded additional research The following seem to be the most Interesting and promising
areasmiddot First the specincation of liquidity constraints to household behavior may help to obtain a
better understanding of policy chenges that affect directly the interest rates in the economy Second
major attention needs to be paid to the incorporafion of a roraign sector (with open capital markets)
Third some efforts are needed in terms of finding adequate computation techniques now that
dynamic modelling eally makes the curse of dimensionality worse than eVOrbull
The modelling of financial markets is the single most unsatisfactory speet of the dynamic
models surveyed here~ The problems associated with provlding a mo~ adequate treatment of
financial markets and in general endogenous and optimal financial deci$ions bull both at the corporate
and at the personal levelsmiddot have 10 be addressed That necessarily requires Introducing uncertainty
into the models To b adequate this must go well beyond the mere assumption of point expeetalions
with the whole array of modelling implementalion and evaluation problems thai generates
bull I i
42
REFERENCES
Abel A and O Blanchard 1983 An Intertemporal Model of Saving and Interest Econometrica
51 pp 675-92
~ Andersson K 1987 Tar-ation of Capital in Sweden - A General Equilibrium Model Ph D
Dissenation UnIversity of Lund Sweder
Auerbach A L KotlikoH 1983 National Savings Economic Welfare and the Structure of
Taxation in M Feldstein ed Behavioral Simulation Methods in Tax Policy- Analysis Chicago
University of Chicago Press
Auerbach A l Kotlikoff 1984 Social Security and the Economics of Demographic-
Transition in H Aaron and G Burtless eds Retirement and Economic Behavior Washington DC
Brookings Institution
Auerbach A L Kotlikoff 1987 Dynamic Fiscal Policy Cambridge Cambridge University
Press
Auerbach A L Kotlikof and J Skinner 1983 The Efficiency Gains from Dynamic Tax
Reform Internatiooaj Economic Reyiew 24 pp 81middot100
Ballard C 1983 Evaluation of the Consumption Tax with Dynamic General Equilibrium
Models Ph D Dissertation Stanford University
Ballard C 1987 Tax Policy and Consumer Foresight A General Equilibrium Simulation
Study Economic Inquiry 25 pp 267-284
Ballard C and L Goulder 1985 Consumption Taxes ForeSight and Welfare a Computatonal
General Equilibrium Analysis in J Piggott and J Whalley eds New Deyelooments in Apolied
General Equilibrium Analysis Cambridge Cambridge University Press
Ballard C D Fullerton J Shoven and J Whalley 1985 A General Equilibrium Model for
t 43
Tax Eoliey Eyalua~iQO NBER University or Chicago Press
Borges A (1987) bull A Survey of Compula1ional General Equilibnum Models CECD Economic
Slud~ 8
Bovenbig L 19B3 imperfect Capital Mobility in a Perfect Foresight Herberger Model
Mimea UC Br~eley
Bovenberg L 1984 Capitel Accumulation and Capital Immobility Qmiddottheory In a Dynamic General Equilibrium Framework PhD dissertation University of California at Berkeley t
Bovenberg L 19850 Dynamic General Equilibrium Tax Models with Adjustment Costs in A iManne ed bull 1985 EtQnomrc ECUjilbrium Model Formulation and Solution Mathematical
lProgramming Siudy 23 Arnsterdam NorthmiddotHoliand
Bovenberg L t 1986 Capital Income Taxation in Growing Open Eeondmies Journa of Pubtic
Eamp(omlcs 3t pp 347middot376 IBovenberg L ald W Keller 1981 Dynamics and Nonlinearlties in Applied General
Equilibrium Models Internal Report Mlmiddot81middot208 Department for Statistical Methods CBS bull
Vooiburg The Netherlands
Charnley C 1981 The Welfare Cost of C~pltal income Taxation In a Growing Economy
IJournal of Political ECQnQm~ 89 3 pp 468middot496 I
Charnley C 19810 Efficient Tax Reform in a Dynamic Model of General Equilibrium I Mimeo Yale University I Charnley C 1982 On the Optimal Taxation In Stylized Models 01 Inlartemporal General
Equilibrium Mimeo Yale University I Decaluw B and A Manens 1985 Pays en Oeveloppement at Modeles Calculables OEquilibre
General - Une Revue de al litterature empirique Monographie Sene 9 Volume 3 Centre de
Recherche et Developpemont Economlque Unrslt de Montreal
44
bull
I I
I
I
1
Erlich $ V Ginsburgh and L van der Heyden 1987 Wher~ do Real Wage Policies Lead
Belgium bull A General Equilibrium Analysis European Economic Reyiew (forthcoming)
air R and J Taylor 1983 Solution and Maximum Likelihood Estimation of Dynamic
Nonlinear Ratioant Expeclations Models Ecooometrica 51 pp 1169middot1186
Feltenstein A 1984 Money and Bonds In a Disaggregated Ec~nonYin Searl 1-1 J Shoven
Eds Aoolied ~eoeral EqJiliDCiulD tioalYsis Cambridge University Press Cambridge
Feltenstein A 19B6 A Dynamic General Equllibrlu Analysis of Financial Crowding Out
Theory with an Application to Australia Journal of Public Economics 31 No1
Fullerlon D R Gorden 1983 A Reexamination of Tax Distorllons in General Equilibrium
Models in M Feldstein Ed Behayora Simulation Methods in Tax pOlicy Analysis Chicago
University of Chipgo Press
Fvllerlon D J Shaven and J Whalley 1983 Replacing the US Income Tax with a
Progressive Consumption Tax JoulOal of Public ECQnomics 20 pp3middot23
Fullerton D Y Henderson J Shaven 1984 A COmparision of Methodologies in Empirical
General Equilibrium Models of Taxation in Searl H J Shoven Eds ~pQlied Geoaral Jguilibrium
lIoalyss Cambridge University Press Cambridge
Fullerton 0 A King J Shaven and J Whaney 1981 Corporate Tax Integration in the United
Statesmiddot a General Equiiibdufi1 Approach American Economic Review 71 pp~ 677~691
Galper H R Lucke E Tader 1985 Taxation portlor Choice and the Allocation of Capital A
General Equilibrium Approach Mimeo
Gill p W Murray M Saunders M Wright 1986 Users Guide for NPSOL (Version 40) A
Fortran Package for Nonlinear Programming T echnicsl Report SOL 86middot2 Deparlment of
Operations Research Stanford University
Glnsbur~h V and L van der Heyden 1985 General Equilibrium with Wage Rigidities An
45
AppJication to Belgium in A Manne ed t Economic EQumprit1m~ Made Formulation and $oJioo
Mathematical Programming Study 23 Amsterdam NorthmiddotHolland ~
Goulet J 1968 1 Adjustrent CO$1S in the Theory o(the Firm Reyiew of EconomiC Studies 35 bull
pp47middot55
Goulder L 1985 Tax-Financed versus BondmiddotFinancedGhanges in Governent Spending
bull Mimeo Harva(d UniversilYmiddot
Goulder l J Shaven and J Whalley 1983 Domestic Tax Policy and the Foreign Sector The
importance of altemative foreign policy formulations to results from a general equilibrium tax
analysis model In ~n methods in Ibe antico of income from caoia1 ed Martin
Feldstein Chicago University of Chicago riSS
Goulder l L Summers 1987 Tax Policy Asset Prices and 3rowtly ~
A General Equilibrium Analysis NBER Working Pper No 212B
Grandmont J 1982 Temporary General Equilibrium Theory Ch 20 in Handbook 01
Mathemalical Eccnornics Vol II NorthmiddotHoliand
Harberger A 1959 The Corporate Income Tax An empirical Appraisal In lax Revision
Comoendjum Voll House Cornmitee on Ways and Means Washington DC Govemment Printing
Office
Harberger A 1962 The Incidence of the Corporate Income Tax JOULn1 of Political
Economy 70 pp 215middot240
Harberger A 1964 Taxation Resource Aliocalrlon and Welfare fn lb Role of Dir bullbulll and
Indjresl Taxes io tbe Fede[al Reserve s~I~Ill Princeton Princeton Unlveresity Press
James J 1985 New Developments in the Applications of General Equilibrium models to
Economic History in Piggott aand Whalley Eds
Jorgenson D and K Yun 1984 Tax Policy and Capital Allocation Harvard Inslitute of
bull
I
I f
Economic Research WP 1107
Judd K 1965 ShonmiddotRun Analysis of Fiscal Policy in a Simple Perfect Foresight Model
Journal of poJiHcal 5CCJiW1l pp 298middot319 r
Upton D J Potbc J Sachsa nd L Summers 1982 Multiple Shooting in Rational
Expectations Models fuoometrlc 50 1329middot1333
Manne A ed 1985 fconQrnjc Eoullibrium Model FOrIDujaljQo and Solution Mathematical
Pogrammlng Study 23 Amsterdam NorthmiddotHolland
Manne A 1985 Or the Formulation and Solution 01 Economic Equilibrium Models In A
Manne ed ECQOQlli-Jr~ibmiddotum ~tQdel EQlJulailon and Solution Malhematical Programming
Study 23 Amsterdam NorthmiddotHolland
Mansur A and J Whalley 1984 Numerical Specification of Applied General Equilibrium
Models in Scarf H J Shoven Eds Aoplied General Eouillbrium AnaiysectIsect Cambridge Universily
Press Cambridge
Merrill 0 1972 Applications and Extensions of an Algorithm thaI Computes FixedmiddotPoints of
Cenain Upper Semimiddotcominuou Point to Set Mappings PhD Disserlation Unlvel$ity 01 Michigan
Negishi T 1960 Welfar Economics and Exislence I an Equnlbrlum for a Competitive
Economy MelroeCOOQ11ka 12 92middot97
Pereira A 1985 On the Existence and Uniqueness of Optimal Ou1put Path lor CRTS Firms in
Adjustment Costs Technologies Mimeo Stanford University
Pereira A 1985a Consumer Expectations and Ihe COSI 01 UllUty In an Intertemporat
Framework Mlmeo Stanford University
Pereira A 1985b Corporale Tax Integrallon and Government Deficits A Dynamic General
Equilibrium Anaiysls Mlmeo Stanford University
Pereir A 1987 On Ihe Computation of the Users Costs of Stocks Ecooomi (forthcoming)
47
Pereira A bull 1987b Equal Yield Alternatives and Government Oeficlts Mimeof Stanford
University
Pereira A 1987c A Dynamic Applied General ~uHlbrium Model for Tax Policy Evaluation
Comlletamp Docurnent2tion Mimeo University of California San Diego (in progress)
Pereira A 1987d Should the investment Tax Credit be Rei(ltroduced Mlmeo University
of California San Diego (ir progreSs)
Piggott J and J Whalley eds 1985 tJew QeyeQpments in Applied General EQuilibrium
bullenalysjs Cambridge Cambridge University Press
Preckel Pbull 1985 Alternative algorithms for computing economic equilibria In A Manne Ed
Radnor R 19amp3 Equllibrium under Uncertainty In K Arrow M Intrilligaor ods
I ibull
Robinson S 1 g86 Mullisectorai Models of Oeveloping CountrieS a Survey Oivision of
Agricullure and Resources UC Borkeley WP 401
Scarf H 1967 The approximation of fixed points of a continuous mapping ~MM Jouwal of
epplid Mathornallcs 15 pp 1328middot43
Scarf H with T Hansen 1973 00 the Computation of EconomiC Eguilibria New Haven Yale
University Press
Scarf H J Shoven Eds 1984 61l1llied Genera Elujlibrium 6nalyssect Cambridge University
~ Press Cambridge
Shoven J 1976 The Incidence and Efficiency Effects of Taxe on Income from Capital
Jurnal of ramie1 Economy 84 pp 1261-83
Shoven J 1983 Applied General Equilibrium Tax Modemng IMF Slaff PaoerS 30
Shoven J and LB Simon 1986 Share Repurchases and Acquisitions An Analysis of which
Firms Participate NBER Working Paper No 2243
bull
48
IShaven J and J Whalley 1972 ft General Equmbrium Calculation of lhe Effects of
Differential Taxation of Incoma from Capital in the US Jauroe of public Econgmics 1
pp281middot321 (
Shoven J bull and J Whalley 1973 General Equilibrium with Tax$s a computation procedure
and an existence pr00f ReviEhi of Economic Studies 40 pp 475middot490
Shoven J J Whalley 1977 Equal Yield Tax Alternatives A General Equilibrium
Computa~ui0nal Tfchnique J~HJrnal of PUblic EconomIcs 8 pp 211~224 bull
Shaven J J Whalley 1984 Applied GeneralmiddotEquilibrlum MOdels of Taxalion and
International Tr8de An Introduction and Survey Journal of Economic literature 23t pp
1007middot1051
Slemrod J 1980 A General Equilibrium Model of Capital Income Taxation Ph D bullDissertation Harvard Univnrsity
Slemrod J 1983 A General Equilibrium Model of Taxation with Endogenous Financial
Behavior in M Feldstein edbull Behavioral Simulation Methods In Tax Policy Analysis Chicago
UniVerSity of Chicago Press
Stone J 1985 Sequential Optimization and Complementarily Techniques for computing
Economic equilibrium in A Manne Ed
Summers L 1981 Taxation and Corporate Investment a q-Theory Approach BroQkinas
Summers L 1985 Taxation and the Size and ComposHion 01 the Capital Stock An Asset Price
Approach Mlmeo Harvard University
Whalley J 1975 A Geneal Equiloibrium Assessment of the 1973 United Kingdom tax
reform Economica 42 pp 139middot61
1JLTIllOS JIQll~)ltG PAPERS PUBLICADOS bull
I
I I
j
112 70 - LUCENA iJiCIIJo ~ IIEnvironrilent Honitoring and Orqanization Structure T (Setellibro 1987)
nI 71 - L~ENA Diogo A Note on the Representation of Information Stxueture (SeteJObro 1987)
nQ 72 - LUCENA D1CgO Envuorment 111)nitoring and Organization
Structure II (Setelilhro 1987)
ril 73 - AlmmE3 Antonio Pais e CAsPAR Yitor bull Tributacao Inoentivos e InvesUlllentos Analise Qualitativa noYellibro 19S7)
n2 74 - liAlltOSA rose Pedro Opt1lllal lIage R1gidity A SUggestedtlthodology to Teet tb Theory and an Application (Outubro 1987)
nI 75 - CABRAL L 11 B Three lIotes on SYJIIlletric Gemes witt itsYJIIlletric Equilibri (NoveJObro 1987)
n2 76 - LUCENlt D10gOTo Search or Not to Search (Fevereiro1967)
nQ 17 - SA Jorge asconcelos euroI A Theory of SynergyB (Fevereiro1987)
n2 76 - VILARES tanuei Jose Os Be IntenaMios IlIIportados ColIIO Factor de Produao (Julho 1987)
nQ 79 - SA J~rge VasCQncel03 e HoY to Copete and Co~unicate in tature Industrial Products_ (aio 1988)
n2 80 - ROIl Rafael Learning and Capacity Expansion in a New larket Under Uncertainty (Fevereiro 1988)
n2 81 PEREIRA Alfredo llarvilo Survey of Dynamic Computational General EquilibriUlll Hodel for Tax Policy Evaluation (Outubro 1987)
Qualquer in ormacao sabre o 1IIorlltlng Papers ja publlcados sera prestada pelo Secretariado de Apoio aos Docentes podendo os ~e3OSf
ser adquirido5 ne Secyo de Vlndas da Faculdade de EconollLia UNL na Travessa Esteviio Pinto Cltpolide - 1000 LISllOA
bull
2
static tnvestment passively actjusts to saving The typical instruments in the US tax system are
modelled and the government budget Is balanced This state of affairs Is discussed and surveyed In
Shovn (1983) Fullerton-Henderson-Shoven (1984 and Shoven-WhaJ)ey (19S4) t I
Integration etc - has Induced Important developments There have been attempts to Incorporate
life-eyce behavior intergenerational issues allocation of savings investment decisions and
adjustment costs corporate financial decisions and government deficits In to the CGE models Underlying all of these features is the need to endow the models with a dynamic structure w~houl I which most of the above issues cannot be adequately addressed
IThe first efforts to incorporate dynamics into the CGe tax models are associated with bull
Fullerton-Shoven-Whalley (1978 middot1981 1993) Consumption decisions incorporale some t I
int~rtemporal aspects and an Intertemporal equilibrium path for the economy Is obtained by Ihe 1
Isequencing of static short-nm equilibria The equHibna are connected through the evolution of the I
capital endowments which_ tn turn depend on saving middotbehavior A compr~1e discussion of a recent I i Iversion of Ihis model is provided in BallardmiddotFulierton-Shoven-Whaney (19SS) I
More recently a whole new generation of models has been developed by Andersson (19S7) Auerbach-Kotlikoff (1983 1984 1987) Ballard (1983) Ballard-Goulder (laeS) Bovenberg I (1984 1985 19B5) ErlichmiddotGinsburgh-Heyden (laB7) Fellonstoin (19B4 ISBS) Goulder I (19BS) Jorgenson-Yun (1984) Pereira (leaSh 19B7e ta87d) and Goulder-Summers
(1987)
tThis article focuse~ on slJch efforts Our scope is narrow By concentrating on decentralized i
I i
Mo~e recently j the ~hrust towards better descrIptive power together with the nature of the tax
policy issues under analysis effects of a ponsumption tax capital taxation corporate tax
3
numerical W~ieral equilibrium modelling designed 10 addresS tax polic issues we abstract from
several ether related areas and approaches First we abstract from analytle models along the lines
of Abel-Blanchard (1983) and Judd (1985) Second we abslract from centralized growth-type
megels of laxation like Charnley (1981 1982) Third we abstract from eentralielti models
I
associated with World Bank researchers See for example Adelman-Robinson (1978) and
Dervis-de Mele-Robinson (1981) Finally we abstract from other CIl fields See for example
ShovenVlcalley (1984) for a sUlVey ollotematlonal trade applications DecaluwmiddotMartens
(1985) ane Robinson (1986) for applications In the area of development James (1985) for a bull I I
survey of economic history applications and A Manne (1986) and BOrges (1988) far all I encompassing surveys
The models included in our comparfsion are I
I 1 8allardFuliertan-ShavenmiddotWhalley (198S) l 2 Andersson (1987)
3 AuerbachmiddotKotlikoff (1963 1987)
4 Bailarcl (1983) and BallardmiddotGould (19aS)
5 Bovenbeg (1985 1985)
6 ErlichmiddotGinsburgh-Heyden (1987)
7 FeltenSleln (1984 1985)
S Goulde (1985)
9 GouldermiddotSummers (1987)
10 Jorgenson Yun (19B4)
11 Pereira (1986b 19870)
bull 4 I
The kei features cf these models afe summarized in Table 1 which provides a structure for the
discussion Most of toe models Ihal we examine could be said to be In the Yale tradition That Is
they wer developped oy sudnts oi Herbert Scarf or students of the students of Herber1 Scar This
shyincludes BalladmiddotFuerM-Shoven-Whalley Andersson Ballard-Goulder ~)Venberg FellenSlein I
ibull Go-cider GouldermiddotSJmmers ard Pereira The exceptions from this heritage are Auerbachshybull
Kotlikoff Erllch-GinsburghHeyden and JorgensonmiddotYun I This survey is organized ltis follows rne second section provides a brief overview of I
Inon-dynamic CGE models The third seolion offers a discussion of modelling economic behavior and
equilibrium The fourth section focuses on model implemenlalion and policy ev~uatlon The fifth
Isection summarrzes empirical evidence en selected policy issues The sixth section illustrates the
power and wea~nesses of dynamic modelling with the issue of corporate laxintegration Finafly the last section summarizes the state~of-the-art in terms of dynamic modelling for lax policy
evalutlon and sugg-ess areas of future research
_
--
--- -~ ----- __-shy
_ _ - _- - Con~um9rs_ 8amphavlor
Bajla rd -Fulflonmiddot -Shaven-Whalley
1985 Anderaaon
1087
Auerba ch~ Ko1Uko 1983 1987
BaHElfdmiddotGouldor 1983 1905
Sovanborg 1985 198e
ErHchmiddotGlnaburghmiddot middotHeydeR
1987 Feltensteln 1984 1966
Gouldar 1985
GouldorSummor 1987
Jotgeneon-Yun 1984
Porelr 1966 1991
Count dynamlc dl9agg~ objecth lastes constraint ondog endog OvorlapBnqvMI of datil
year _seltlng labor no savings gonorst mollves
inlertsmporal 12 Income otility eel IntertempOfll1 yes yo no noLS 1973 Ymiddotmiddot
_ groups max ---_
0 Intertnmporal rllcufsivs
Swaden 1984 yes aggrogate utility CES eq of molion 00 yos na no consumer max for weaHh
tntortorrporal Isoeiastlc~ 00-1983
ill yo 55 poundtUS u~illy middot19831 In tcemporal yesmiddot 11El7 yes yo 00
cohorts rn~x CES-19C7 plauslblo
bYPothlC1 inlsrtamporal
19n yes 55 ago uliJity Isoelastlc Intertemporal ybullbull ye ye ysLS BFSWexpanOOd _ cohorts max
0118 inlertemporal recursivo LS plausible ye aggregate ollllly lsoolastic eq ot motion no yes no no
hYPOlhotic consumer max for weallh jnterlemporlll Boigium 1960 1983 ro aggrogate utility St()ne~Gearyjner1emporal no yo no no
consumer max Inlerl(lmporal two periodmiddot two yes
Austrs1i 1991middot1982 yes 2 income ulllly CD yeluly ybullbull tn the flrsl no no 2-perlod) groups max constraint porlod
Intertamporal LS 1973 ybullbull 55 aga utility eel intarhtmporal ye ybullbull ybullbull y
BFSW bullbullponded cohorts max
intert9mpor~1 W 1973 ves aggregate ullily 180085110 intertemporal no ybullbull no no
BFSW bullbulllMded consumer max (I~gt IntertempQ(a~
W 1955-80 ybullbull aggregate ulilily trenslog intertemporal ybullbull yes no no consumer max
Inlerlemporal recursive 1973LS yes a income ulllily CmiddotD eq of motion yos ybullbull 00 no
BFSWxpanded groups max for wealth
middot5middot
bull-- _---- -- _-----shy~ -~
I AlJlt 1 - uynafllc (lIt MOOero fur all runey tVlIIVllllQlI
dynamic dluggI utllng
I Ba~ll-a rd-fullerton
-Shoven-Vlhalley I no 19 sectors 19S5
Andarsson 1987 ye 2 sectors
~ Auerbach~KotHkof no~1gB3
1983 1967 yesmiddot1g87 1 sector
Ballard~Go uld ar 1983 1985 no 19 sectors
Bovenborg 1965 1986 yes 2 sectors
Erllch~GIn5burg hshy-Heyden no 24 seclors
1957 1__ Feltenataln 1964 196e no 30 sectors
GoUldsr 1985 no 3 sectors
~ GoutdarmiddotSumrnamprs
1987 yes 5 sectors
_ Jorgen9ClIO-Yun ona sector
1984 no produces two goods
Poralr 19S~ 1987 yos 4 sectors
~bJecllve
Produc~rs
technology behavior InpIJ1 1-0
matrix adl costs opt Imal
capital stock
sialic cost min
CES in valu$ added
110
capital labor yes 00
np Invest adjusts
to sFvnns
max 01
interlempora flot rash flow
cnsl min-lOS) Imax of valul)
of the firm~1987
ill
CD-ISBa CES-1987
capital labor
captel labor
no
no
YA quadratic in
I nO-191)3J
qutldratic in Ilk-1gB
ye
YAS
static cost min
CES in vafl1oadded
10
capita labor yes no
no invest edjuss
10 savifQs max ot capita yas ~
value ot the CD labor 00 quadratic In yo Ilrm Ilk
static CD in several no global ltlost mln vaiue added
110 ltsp1al
Jabor Ves no Inves adjusts
to MVrHIS
statle oost min
____ slatio
cost min
max of value of the
firm equivalent 10 two~stage
~ost min me) of
CO in value ad~ed
110 CES In
value added 110
CESin shyvalue added
110 dual
trBRslog price function Cobb-Douglas
capUal labor
cepHal labor
capital labor
capital labor
capital
yes
yes
yes
no
no invost adjusts no to S8vinos net of
El-xog 9011 deflcitsl no
no invest adjusts to savings r
yes quadratic In Y
Ilk
no yo I
yes Intartamporal net cash flow
In V Added 110
labof Y quadratic In I
y bullbull
- 6 -
~ --- ~ ~- -- -~------ --- --- - --~--~ ---~ shy
_
- ---
--
--------
- - - --------
H n v I taxes
olllng allowed debt Ballardfullamprton~
dynamic objective constrfnt tupendlttire (IlleUs endogenous
aU major balanced trade yoarly ybullbull rly endogenous -f ~Shovon WtlitUy US taxes with C(in~Hlt00 pubHc uility balancod cQmposilon no no debt
eiasticllv junel l 1985 rnagt eCP-REV~d1 aU fn3jO[AndtH3Son vearly balanco small opan Gcon J Swodish inturn fa to cr raturn1987 no - inaia dlJn axogonolJs 00 Ini ITnininod poundgtPfUV
-I L)~ in~h~ --~ ~~ ~pound~~---1- ----shy
runrac h~KQllfkofi ai gulD
1963 1937 US m1l3syo - intertnmponl tixogancuamp yo ys cos~d ampConoroy bandmiddotfinanced flon-oppoundIla
B iii lIa fe-Gould 9r nil llltl1gtr balanced 1rndayearly yearly crcoge)Jus
US t)X6S wiih conslanl I
1963 1965 no public utility balanced cOlYlposWon~ no nD debt max budnt ExP~RCV alastlcHy lunct
Bovenberg simp~iriad 1985middotclosod econybullbull rl~ no - balanced exogenous no no debt US laxes 19a6~two counlry1985 1986 bodgel model
Ert leh~G tn aburghmiddot yeally simplified Inlert walt funet Ishorl~run ttads Heyden no balanced exogenous 00 no Belgium
taxes _ desloullbtium flltUensleln
1967 budaet EXP=REV 6-1mprifia-dmlm costs of yoarly yo rast of tho world
1984 1985 producing a ExTax Rev exogenous financed with yo Australian ~s a Ihlrd conSUmer public good money and bonds non-optimal taxes group
Goulder yaarly annual and endogenous aU maJor
laquo1()Sed economy1985 publlc utilily In1ertemporal composillon bull yes yes U~ taxesY rna versions IlEFElQgtfpoundV bond-financed (non-opllmal balanced trademiddotGoulder-Summer yeary yoarly endogonous bull all major
1987 00 public utility balMc composition no no US taxes with constant max budGel EXPREV alasticilY fund Jl)rgenaon~Yun yoarly aU major
no balanced exogenous no1964 US taxes closed economy budGet inteftemporal recurslve endogenousPereira all major
yo public utility eq of molion lovel and yo yos1995 1987 US laxes closed economy rna for debt composition bondmiddot financed oplloll
- 7 shy
bull - -__----- _ - ------ -- ------- _-----
---- ---
------
------
ll-ltiLt 1 uynamlc Lut MOOSS fOf I ax YOtlcy tv8luallon
Inanell ~ets and maf~9ts ojumLrtm- ----_ ----c_~
privata do btl GlvldtHlI pL~lc aliocalQflo rls shy
1rlc ypa expGct assets equity retention dobl 01 savlnq$ promlums path
Sal ra rd fu Barton phys cap1 used 10 buy soquence lt
Shavon-Wha1ley no financial exogenous oxogenous no physicBl 00 bull 01 sIalic 1E myopic 1985 assets capital flqrlil -
Andersson phys cap used 10 buy 1987 no financial no no yo physical no dynamic ~ perfect
assets capital Auerbach~KolUko If phys cap usod to buy
1983 1987 neo financial no no physical cap no dyoamlc ~ perfectYmiddotmiddotnSSf)ts and Qovt debt SailBrd-Goutder phys cap
1983 1985 no financ[al assets-shy -------
Bovenberg phys cap 1985 1986 no iinenclaf
asso1s Ertlch-Glnsburgh~ phys cap1
-Heyden no financlaJ 1987 assefs
FeltoRsteln phys cap 1984 1986 Invostment is
bondmiddotfinanced
used to buy
exogenous exogenous nObull physl~al no dynamic ~ perfllcl capital
used to buy
no no no physical capital
no dynamic ~onlinuos lima
PI perfect
used 10 buy no no no physical no dynamIc PF perfect
capital jllNt) porlod) bull used 10 buy
no no ye physical cap no dynamic PI p4)rfect and govt debt (Iwo Iods) bull
Goulder phys cap used to buy 1995 no financial exogenous exogenous yenbullbull physical cap no PFlfE bull IdynamIc perfoct~assels and am dabt slaUc
- -~ - shyGoulder-Summers phy$ cap
1987 bonds and no e)(oganOU$ no used 10 buy ye dynamic PF perfect _ bullbullultv bonds MullY (exoaenous
Jorgensonyun phys cap 1984 no financial
essels -~ - ------
Pereira phys cap 1986 1967 bonds end
equHy
________
no
8xogenoos
no
exogenous
no
ye
used to buy two type$ of physical
copltal used 10 buy
go dobl bonds equity
no
no
dynamIc
-
dynamlc
~
lE
perfect
fIbullbulllble (nonperlecB
middot6middot
~-- _middot~ ___~w_ - - ___ bullbullbull___ -------- shy~
-----
------
- ---
-
IlflpnUItll tit lun pVII-y PIlUAlIV
q6H1brla oqual fax polley ovnluaU-on efUclency derlvatton Ioorllhm
calibration param~hr computation COmelfed llold tndfcaHH eHecls
Bnl jard~FuU6rtQ n base year back soll1lion steady-sta1e yo inlsrgroup inerseel or Iii
VS transilion lump-sum with Interfemporai 1965
~Shovn Whallmiddotv replication wllh exog Morrills nmamo1ers alQorilhm _~ond~middotsta1$ ears Inc CVor EV
i~noerucn sleady-sfala no agg(ogaled lnlalsectora 1967
base year back sobHon SiMNON package replication wHn oxog valiant of VS Iransilion no Indicamptor lnlertamporal
paramelsrs multiple shoo ling 6~~~l~slat8
AU8rt)ach~Kotllkot steady-slato yo Inlerguna-r a lions
1893 1llH
quatitaUve by replication a$Sium~lkml Gauss-Soidel liS Uln1WOn any rule -lilh 1111ermpOrB
sonsililJ1y I _ ~poundI~Jjnnelill $t(lfdvmiddot~luiA J3djlt15S ~~ nnrilW ~--------
a ail r nimiddotGould or bas9 9tH bac~ solution Ballard~ staadymiddotslalo )05 kltHSclo(a~
1983 1985 VS Iransllinn lump~sum dynalTlk CV inlerlemporal steadY slafe parameters alaorlthm replicaUon+ with oxog Gourder
+stead~stato (ers loc inlareel10rat BCvonborg qunmaiive by dynamic version stoadY-$Igtto yes 1985 1985 YS transition some govt eN Inlortempora
S9rHltivilv OnearizaJian method ropllcaUon assumptionl of Johansons
+steady-sfata paoo palh Erllch-Glnsbu rgh~ actual vorsus no aootagalod inlersectoral2 bftse years back solu1ion variant of Negishis
Heyden replication wIth alltogbull optlmizaUon countorfaclUal no ~odjcator inlottempora 1987 paramlJters aocroach
faltonstaln baso yeat back solution actual versus no aggrampgalfld intersElclornl replication wilh exog Merrills counlerfaclutll no indlclltor middotlntamprlempOull1984 19B6
parameters al90rithm 1981middot1982
Gouldar staady~slala Inlar5ecioralbase year back souUon Variant or 1S lronsilion no dynamic CV inlortempora
stoady state parameters Newlon 1985 feplicalion+ with iUog Fair~Taylorl
+slaildy-sla1e Intargeoorat Gtlulder~Summers base year back solution varIant of stlJadymiddotstate fixed no aggregated fntersectoral
vs trarlsition government Indicator Intertempo1al steady state oatameters Newton
repticaUon+ with exog Fair-T aylQf 1967 +ste-adv-stata bullbullbullndlno financial
Jorgenon-Yun econometric steady-stale Money melfic In1ersecloral
- - 8slimalion vs traosUian no social weUare functton intertampotal1984 +~teadlmiddotstatl)
~ shy
qtJalitaUv8 by NPSCt optimal path ye generalized inlersectorai 1986 1987 Perelr
replication 8S$umptlonl optimizatIon V$ lumpmiddotsum dynamic lntertemporat se~~ttilitf ____Jllgorilhm opllmal pah pen inc CVoclV financial
- 9 shy
bull ~ ___ - _______ - ____ bull___ OM bullbull __ bullbull __~____ ~_
--
- - - --------
--------- - - - --------
---
1nltJ1 bull I ~ 1J11ltI1 v I ~
-
Sa Ita rd-Futlertonmiddot intargration consumption lax ~StHlven~Whall~y VAT capila~ or labor tax
Ch~mqflS1985 Andersson
consumpion tax neutral corporals lax system 1987
it ue rb a c h~ i(at II Sltotf COIlSUft1ptian lax (poundpird ItS hibor Mx~on ufocto of taxation on capital fOfrnJilti1
dvnamio fiscal policv effects af OQvernmanl defioit (3tlard-Goulder
1883 1987
allocts 01 the adoption of a 1ge3 1985 pure consumpHon lax
im~rtmce of cons fCreSClht -Bovenborg analytically orianiad approach
1985 1986 effects ot pure conumption tax capllal income taxation in Opo economies
Erllch-Glnburghshy-Heyden reat wage policios in Belgium
1987 Felter1steln 1984 - theorotical
1985 ~ linnncial crcwdlng out1984 1986
Goulder effects of tax- vs bond-financed 1985 chanQes In pubtlc expondllufe
~~--
Gouidr~SUmmampfI torpora1$ lax cvts reduced 1987 ITO increase 10 gasolioe taxes
Jorgenson-Yun
effects ot several programs of tax reform on the allocation of ~~ptai
1984
-- Purelra
corporate la)( Integration effects cf romiddotinlroducinQ imreslamnl tax credits
1986 1967
10 bull
bull -~- ~- --- _ - ---____--_ - -----_-_ --~ ---- ~
1 1
2 Brief overview of non-dynamic CGE models
The origin of most or the models we survey goes back to the work of Scarf (1967 1973) whlgt
first developed a reliablo algorithm to compute equilibrium price for a ArrowmiddotDebreu economy
HiS algorithm used simpicial Bubdivision techniques and can -be shown to be the computational analog
of the fixed point theorems previously used to prove the existence of equilibrium His technique
could soW a model with an arbitrary number of consumers and commodities as long as all agents
were price takers consumers were subject to budget constraints demands were continuous andmiddot
production did not display increasing returns to scale The alQorithm while guaranteed to converge I
I I I
was relatively slow for prcb1ems involving more than S1If 20 dimensions A maior improvement in
computational speed Woo offered by Merrills algorithm (1972) which used tpe same fundamental
ideagt as Scarfs procedure
Scarfs model (as is true for the standard Arrow-Debreu model) does not include a government
sector ~ neither taxes nor public goocs At one of the mosi promising appHciUions of the new
computational techniqu w~s in Iha area of tax policy evaluation Shoven and Whalley (1973)
exlended the general equilibrium model and computational approach to include a wide rry of taxes
and a governl11ampnt spending plan The original ShavenmiddotWhalley model was static (although the
different commodi~ies CQuid be considered simitar goods available at different datest as In
Arrow-Debrau) and government was aumad to run a balanced budget The data are arranged as a
social accounling matrix The models spacification and calibration is checked by solving it in the
presence of the base set of laxes The result should ba exactly tha initial social accounting matrix
After having pasd this replication check the model is solved for a counterlactua equilibrium In
the presence of a new tux design The result is once again a social accounting matrix The two
equilibria are compared in order to tSS($S the impact of the new tax plan The first uses ot this
bull
1 Z
model for lax policy evauajon were ShovenWhalley (1972) Whalley (1975) and Shoven
(1916)
Completely static models as Shaven-Whalleys are unsatisfactory for many tax reform ISsues
These include corporate lax Integration effects of investment tax credits effects of accelerated
depreciations consumer Or e)penditure tzxes~ importance of saving subsidies like IRAs etc These
are essentialy dynamiC ksues They involve not only the allocation of capital across sectQfsbut
perhaps more Important the capital intensity of the economy But the capital accumulation and
capital reallocation take time and may involve adjustment costs Because of these issues Ballardmiddot
Fulierton~Shoven~Whaloy took the first steps towards developing a dynamic model
In this conlexl the ShovenWhalley model has been extended and implemented for Ihe US
-economy In BuIIBrdmiddotFulienonmiddotShovenmiddotWhalley (1985) While Ihis book completely documents the
mo~el it was used in several publicalions beginning in 1978 Their mode) consists of nineteem
production sectors twelvo households and fifteen consumer goods It includes a very detailed set of
taxes including the federal and state personallncome taxes federal and state corporate income taxes
SOCial Security taxes pro1J0rty taxes unemployment Insurance excise taxes sales taxes etc It
has been calibl att-d to reproduce the 1973 US economy Recentty a version corresponding to the
1983 JS economy has been developed
wThe 8alard-Fullenon-Shoven Whalley model is dynamic In the sense that consumers face a
choice between current consumption and leisure versus future consumption (which can be
purchased via s wlngs) The consumer classes act as if they were maximizing a nested CES utility
function over the domain of a CES aggregate of contempo-raneous consumer goods and leisure and
future consumption subject to their income constraint The parameters of those funcHons
determine the shares of income devo1ed to each commodity to saving and to the purchase of leisure
They also determine the two key elasticities in the models ~ the elasticity of labor supply with
bull
13
respecl to tgte rbullbull1 aoer lax wage rae and the elasticity of saving with respecl 10 the real after lax- rate of retum to capital
In the model consumerS have myopic expectations regarding lulure prices and in particular
regarding the future rale of return 10 capilal Ballard (1983) and BallardmiddotGoulder (1985)
inco~ora1e both perfect foresight and limited foresight into this model Future consumption is
acquired by buying a fixed composition portfolio of real investments that offer an infinite annuity
of returns
The production side of Ihe model is compielely static The mod1 Incorporates a constantmiddot
olasicily 01 substitution betweeen prlmary inputs in production (capital and labor) and fixed I coefficients for intermediate inputs The model distinguIshes between Industrial outputs and
eonsumer goods for the simple reason Ihat tile data are classified differenUy Industrial sectol$ I I involve such categories as forestry and fisherle metal mino9 and publishing and printing while
consumers purchas-3 fUfHi1ure 2ulomoblles and books This fact is recognized in the model by Ibull
incorporatirg a secord stage Of production which converts industrial outputs intO consumer goods
This technology is uStally mod9-led as a flxed~coefficient conversion matrix
The BallardmiddotFINton-ShovenWhalley model assumeS that the private sector finance
marginai ioveslment with the same composition of debt and equity as currently exists in each sector
This is the same arsumrtion that Herberger originally used Investors all hold debt and equity in tile
same proportion Thampreiore this ownership can be aggregated into simply capital ownership For
tax purposes however the separate treatment of debt and equity is taken into account at both the
corporate and persofa Io=vel SimllarlYJ the dividend poIicies on corporale equitY are established
exogenously There are no government bonds in the model since there are no government deficits
The modol is solved for a sequence (as many as 100) of temporary equilibria wltll consumers
allocating income between present and future conumption at each point in time The palh for 1he
14
economy ~ a sel of connected equilibria The connectIon is pra~ided by capital accumulation
Capital accumulation is endogenous and determined by saving The model stans with a sociagt
accOunting mi1rix In the bzse case the economy is assumed to be in a steaqy state growth -path (along
which all relative prices ~fe constant) The madel solves for both the naw steady state growth path
and the transltion 10 it zf~er a policy intervention rhe authors have frequently addressed the
question-of howong rt tKes to effectively settle un10 a new steady S1ate growth path
The dynamics of the model are limited however in that future consumption is collapsed into a
composite commodity Aso the absence of government deficits and the lack ofproductlon dynamics
limits the realism of the BallardmiddotFuliertonmiddotShoven-Whatley model for the analysis of dynamic
policy issues (such as the adoption 01 a consumption tax or Ihe elimination of the investment tax
credit) The models we survey try to Improve upon Ihe Ballard-FlIiertonmiddotShaven-Whalley
dynamic modelling The models Jn~orporate more forward lookTng behavior They improve the
dynamics on the conampumrtkm side of the economy and some of them introduce true 1ynamlc behavior
on the part oI the producers We turn now to the issues inY9lv_ed in developing such dynamic models
15 I I ~3 ~ssues in the modelling Of dynamic behavior
We will 100k firt at the speclficalion of economic behavior of consume(s~ producers
goyemment and Ihe r~sl of Ihe world The modemog of corporale and household financial decisions
I is then addr5sed Finally Ihe eencept of equilibrium is discussed
I
i
I 1
31 Consumers behavior
Early efforts 10 build dynamic features into the economic behavior 01 consumers are due to
Ballard (1963) and Auerbach-Kotllkoff (1983 1984) Their work has been closely followed by
several subsequent authors In faet dynamic household behavior Is the single most pervasive aspect
amohg the models surveyed
All the models Inc-O1orale some loro of lila-cycle behavior Household behavior is dewmloed
by the maximization of an addWVely separable time invariant intertemporal utillty function The
utility lune1ion is defined oyer the domain of the consumption goods In the economy In most of the
models leisure is also an argument in utility so that labor supply is optimally determined
Utility maximi4ation is subject to a lifetime intertemporal budget constraint which equalizes
the presen1 value of consumers income and expenditure More reeentiy in Andersson (1987) bull
Bovenberg (1985) and Pereira (1986b 1987d) the constraint i$ defined as a sequence 01
recursive equations of motion on wealth This has the potential advantage 01 accomodating liquidity
constraints However It should be recognized that in the absence 01 liquidity constraints (ie when
consumers are free to completely borrow against future income) the two specificallons of the
househofd ccnstraint are essentiay equivatent Furthermore tn both versions saving is optimally
determined as a way of translaring wealth intertemporally
1 6
Some lodels now have a sophisticated dynamic specification of hoosehokl In Ballard (1983) -
BallardmiddotGoulder (1985) and Auerbach-Kotlikof (1983 1984 1987) consumers behavio is
embedded in an iterQEneraional setting Each of them has S5 age cohorts simulataneously alive
Som-e measure of jntergener1ional altruism is considered in the form of bequest motives In Ballard
(1983) and Ballard-Gouldr (1985) consumers derive utility directly from bequthing parl of
their wealth Therefore bequests assume 1he form of a scrap value of terminal wealth of a
generation In AuerbachmiddotKotlikof (1983 1984 1987) each generation empathyz bullbull with Ihe
utility of future generations ovar bequethd wealth In addition Ballard (1ge3) includes detailed
demographic poiections wfill the intent of incorporating in the policy analysis the effects of the
bull Post World War tI baby-boom
In the real world household decisions also include the optimat allo~atiOn of saving among
shyalternative physical or financial assets Theories of household portfolio behavior are relatively
Iwen~establshed However they involve uncertainty as a crucial element Accordingly they have
middotnot been incorpora~ed in the context of the determinIstic dynamic models we sutyey here We will 1 Come backt-O this Issue below
32 Producer behavior
The efforts to build dynamic features into the economic behavior of producers are more recellt
and less widely adopted The first attempts are due to Bovnberg (1984 1985) and Summers
(1965) Dynamic behavior has been more fully incorporated In the recent models of Andersson
(1987) Auerbach-Kotlikoff (19B7) GouldermiddotSummers (1987) and Pereira (1986b 1987d)
Part of the reason why production side dinamlcs has been more slowty adopted is the weak
supply of accepted theories referring to the dynamic behavior of the firms In the models referred
I 17
to above gynamlc producion and investment b~havior are middotinduced by the existence of capital
adjustment costs and linked to Tobins q theory Adjuslment costs are designed to capture both the
incomplete mobility of capital aoross industries and installation costs ~e the costs of adjusting
capital towards its cpHmai level)
Adjllstment costs can be conceived as internal to the firm and measured in terms of foregone
outpul along the lines of Lucas (1967) Alternatively adjustment costs can be viewed as actual
external cests incurred together with Ihe purchase costs along the Unes of Gould (1969) With the
exception of Pereira (198Gb 1937d) all of the CGE authors follow the former approach
The firms in the economy maximize their market value as the present discountad value of the
future stream of dividends In Andersson (1981) and Pereira (198Gb 1987d) firms are seen as
maximizing the present discounted value of net cash flow Maximization is constrained by Ihe
adjustmenl cosl lechnology and ~n equation of motion describing Ihe evolution of the capital stock
It should be noted that undereogenous divldendrelention rules the two problems are equivalent II
should also be nOled Ihat a salisfectory economic rational for th bullbullistoo of dMdonds with the
present tax code is missing in the profession (Shoven (1986))
The models with static formulation of producers behavior are characterized by passive
investment behavior Investment merely accomodates to saving in the eccnomy With a dynamic
formulation induced b adjustment costs real investmen decisions are forward looking Investmenl
Is endogenously and optimally determined by the firms A fundamental difference between the
shonmiddotrun in which capital stock is given and longmiddotrun in which Ihe level of capital is allowed to be
optimally determined is emphasized
This extra richness of produc1ion dynamics is not without costs A careful look at the summary
tables will clearly show an inverse relation between the adoptlon of dynamic features in production
and the level of disagreggation of the production side of the economy In fact with production
bull
j
1 6
dynamics $e dimension of the problems is immensly increased Let IS be more specific
In a static framework with constant returns to scale production technology the output level ismiddot
indeterminate The optimal aliocation of inputs can be obtained by cost minimization with any
feasible output level generatillg zero profits Such zero profit conditions are used to solve for the
output prices in terms of the factor prices and hencemiddot to reduce the dlmensionality of the problem
from the -number of commodities and factors to the number of factors Thus even if the model deals
with 30 production sectors the computation of an economic equilibrium can take place using only
the dimensionality of the primary inputs in the economy
Now under certain regularity conditions on the production and adjustment costs technologies
leading to enough concavity of the optimality objective the intertemporal output path for the firm is
endogenously optimally and uniquely determined even with constant relums to scale technologies
(see Pereira (1985 1987a) Accordin~y with adjustment costs in the model optimal profits
will in general be non-zero This result has crucial implications for the computatiQn of equilibrium
in the model~ with production dynamics With adjustment costs no reduction of dimensionality is
possible The curse of dimensionality returns as a binding constraint
The introduction of adjustment costs adds another significant complication to the model With
capital being less that perfectly mobile in the economy different rates of return on capital will
exist in different sectors This is a difficult problem to tackle conceptually in the absence of
uncertainty Also in the real world producer maximization choices include also its choice of
financial ratios (debVequity) and payout rates (dividendretained earnings) These financial
subjects are difficult and much Clf this behavior is still taken as exogenous by the modellers We
will come back to these issues below
middot i
19
bull 33 government behavior
middot
Two issues dominate the modelling of government behavIor First govemment behavior has
Ibeen typically seen in Ihe CGE liIerature for lax poliCY evaluation as constrained by yearly balanced
I budsets (see for example 8allard-Fullrton-S~ovn-Whafley 098Sraquo The analysis of
govemment defidlS and public debl in a CelE context requires a dynamic selling Second the level of
government expenditures is either exegenouty given as in Auerbach-KoUikofi (1984 1987)
6ovenbrg (1984 1985) Feltenstein (1984 1986) and Jorgenson-Yun (1984) or
endogenously (but no optimally) determined by the balanced budget conditions as in
6allard-Fullerton-Shoven-Whalley (1985) Andersson (1987) Erllch-Ginshurgh-Heyden
~
(1987) Gouder (19aS) and Gouder-Summers (1987) In the second case the composition of
bull public expenditures is often optimally determined However the leve~ of government expendi1ures
can only 00 endogenously and optimally determined if the government is seen as an optimizing agent
and is allowed to run deficits
The first attempts to deal wiU the government defiCits in CGE tax models are due to
Auerbach-Kollilltoff (1984) and Feensain (1984) In Auerbach-Kotlikof (1984) government
expenditures afe exogenous They grow at the rate of growth of population However given 1he tax
structure and tax revenues yearly deficits and surpluses are atTowed subject to an in1ertempond
constraint that the present value of future tax revenues equals the present value of future
expenditures
In Fellenstein (1984 1986) government expenditures are also exogenously given He also
allows the government to run deficits to finance expenditures in excess to 1ax revenues However
surpluses are returned to CQnsumars in the form cf transfers Accordingly government ts not
subject to any constraint regarding the future repayment of public debt
In GouJder (19851 the 9Vernmenl maximizes a static social welfare ~nctio subject to a-
bull
20 I i
I
I I I
I bull
balanced budget conSirtlL This follows the optimal alocatkm of government expenditures alpng
the lines of Baliard-FHeiton-Shoven-Whalley (1985) Ttle model Is generallzea to allow for
exogenous changes in HIe time path of government expenditures Two financing alternatives are
consjdeft~d and contr8s1ed EldditionaJ tax revenues and bond issuance
Pereira (198Gb 1137d) auempts to address both the incorporation of deficits and the
determination of govement expenditures The path of government expenditures and the path of
delicilSlsurpluses (ano therefore the path lor debt) are endogenously and optimally determined The
government is seen as mltJximizing n intertemporal social welfare function given the 13x structure
Optimization is subject to a sequence of recursive equations of motion reflecting the evolution of the
ptlbllc debt allowing fDr government budget imbalances It should be stressathat this specification
01 the government cOMtcalnt is equivalent to the specificaUon in Auerbaeh-Kotlikoff (1983 1987)
when no liquidity constraints affect -gove~nment behavior
The extra richness in the treatment of government behavior allows tfte examination of
government debt poices in a truly dynamic setting Also financial crowding out effects induced by
government deficits can be analyzed (see Auerbach-Kottlikoff (1987) Feltenstein (1986) and
Perer (198Gb 1987e))
34 Foreign sector
Most of the open economy models follow the assumptions of balanced trade with import and
export net demands characterized by constant elasticities along the line of Ballardmiddot
-Fullerton-Shaven-Whalley (1965) Such is the case of Salfard (1963) Ballard-Goulder
(19S5) and Goulder-Sllmmers (1987)_ In Feltenstin (1985) the rest of the world is treated as
21
an addjlion~ consumer group
Bovcoberg (1986) develops a model In which two economies lIr considered each following
intertempoml perlect foresight paths These economies meet in the international forum Their
trade relaroMhips are characterized by yearly balanced trade accounts
None of the new generation of dynamic cae tax models has yet IncorPorated the lnternational
capital flows as done in aoulder-Shoven-Whalley (1983) for the earlier Ballarrshy
-Fullertonmiddot Shaven-Whalley (1985) model This is unfortunate as there is an important
Intertmporal aspect to international lending The first al1empts along Ihese lines are due to
Andersson (1987) and Erlich-Glnsburgh-Heyden (1987) Andersson (1987) in his model of the
Swedish economy adopts a closed economy appraoch in which rates of return in the domesJic
economy are largely determined by the international capital markets Fidinterest rate induce
Intlnatlonal capital flows which determine and finance the International lrade imbalance In
lurn In ErlichGinsburghHeydenmiddot (1987) foreign trade Is generaled according to an
intertemporal Irade welfare lunction with constant import and export elasticities In the shortmiddotrun
they allow inlernational trade inbalances which generate capilal ftows to the domestic households In
the long run however trade balance is assumed
35 The need for financial markets
A dynamic economic Wucture not only provides the ideal environment 10 model many features of
economic behavior it also permits one to Incorporate the financial side of the economy If the
government is allowed to run deficits the question 01 deficll financing automatically follows If
investment is optimally dewmined and returns to capital are different aClOSS sectors problems of
investment financing arise If there are seeral final1cfal assets In the economymiddot government
4-
I i
ibull
I I I I
I
j
22
bonds privpte bonds and equity (or simply physical capital installed inmiddot different sectors) the
problem of allocation of saving among assets with potentially different returns arises
The papers surveyed var greatly with respect to l~e extent of their attention to the financial
side of the economy At Clne extreme are the models in ~anardmiddotFuliertonmiddotShoven-Whalley (1985)
Andersson (1987) Ballltld (1983) Ballard-Goulder (1985) Bovenberg (1985 1986) and
Erlich-Gjnsburgh-Heyden C19B which are devoted exclusively to the real side of the economy
In turn Auerbach-KotliKoff (1984 1987) Feltenstein (1984 1986) and Goulde (1985)
allow for government debt In these models saving finances changes in government debt and physical
capital Private and publlc 2ssets are perceived by the households as perfect substitutes The
allocation of saving merely adjusts to the relative demands for funds
Feltenstein (1984 1986) is the only model surveyed which introduces money Government deficits are financed by issuing money and bonds according to an exogenously given rule Money is
demanded by consumers for transaction motives and an exogenously given fraction as a store of value
On the other hand government bonds and physical capital are the vehicles for the intertemporal
transfer of wealth
Summers (1985) and GOlldermiddotSummers (1987) introduce a whole menu of financial assets shy
firm specific equity capital Different assets earn different rates of retum However such rates
are equal up to constant and exogenous sector speCific risk premia Therefore the introduction of
constant exogenous risk premia as helpful as it may be in the context of calibration does not solve
the main issue of the non-optimality of the allocation of saving Also talking about risk premia in a
deterministic context is somewhat unsatisfactory
Pereira (1986b 1987d) also introduces a whole menu of assets private and public bonds and
firm specific equity However all the assets are expected to yield the same rate of return and
therefore perc~jved as perfect substitutes The different asset types allow consideration of
23
exogenOU$~8tUequjty triO dividendlrc~enlion rufes and therefore seve-ral sources of investment
financing bond equity and retained earnings
The nonmiddotcptimalHy of the allocation of saving and the absence or exogenelty of corporate
financial rules is _a reftectlon of the limitations of the determInistic approach Either in a context of
perfect anticipation oi the future prices Of in general within the t$atm of point price
expectations all the papers follow a deterministic approach Under such circumstances consumers
either expect diHerent rales of return (inclusive of risk premium) across assets in which case
they wilt buy only one asset (that with highest rate) Of they expect equal rat~s of relurn in which
case Ihey are indifferent about the asset composition of Iheir portfoliO There is no way of
Iradingmiddotoff rales of return and risks to obtain an optimal interiof solution to the problem of the
allocation of saving
The most advanced contribution in the modelling of saving allocation in a CGE setting is due to
Slemrod (1980 1983) in the context of a static one-period model In his model consumers act
according to a two stage separable decision process They Hrsl decide on how much to save Then
they decide on the allocation of saving according to an Indirect uttlity function dependent on the rates
of ratum and variances offered by the different assets in the eoonomy The sourCe of riskiness in the
ampConory comes from an uncertain marginal product of capital On the other hand aside from
portfolio decisions the res of the economy is insulated from uncertainty
The most complete contribution in terms of the treatment of the corporate financial rules is
FulienonmiddotGordon (1983) In a variant 01 Ihe BailardmiddotFulienon-ShovenmiddotWhelley (1985) model
Fullerton and Gordon have capital intensity and optimal financial decisions Jointly deter~ined
through a two siage process The cosl of financing capital is minimized by trading off the tax
advantages of debt against the epected real bankrupcy coSls inherenl with high debVequity ratios
Given the optimal debtequity ratio the level of inveSlment Is chosen such that at Ihe margin the
24 l
bull
rerurn pn ~uity equas the return on bonds plus an exogenous risk premium
36 Market assumptions equilibrium and expectations
Virtually all of the papers arc characterized by Walrasian market clearing assumptions All
markets are perlect1y competi1ive Atomistic competition among agents 1s assumed even thoughmiddot only
a flnile number of agents is considered Virtually no mark~t disequilibria or price stickiness afe
considered The only exception is ErlichmiddotGinsburghmiddotHeyden (1987) bull In thei model of the Belgium
economy I the _wage rate is fixed in the short-run Therefore in the shOrlwrun desequilibrium in
the labor market will generate endogenous unemployment However in lh$ long-run all prices
Including the wage rata are Ilexible and accordingly ali markets ciear
Given the dynamic nature of behavlor in the economy markelmiddotcfearing prices in each period
depend on expectations of fture prices and on tax variables In the economy There are essentially
two ways at interpreting the economic equilibrium in such a dynamic conte~bull If future prices are
perfectly anticipated (io expectations are self-fulfilling) a perfect foresight equilibrium
prevaHs Then future actions are merely the implementation of current decisions for future
periods However jf ptica expecla110ns are not perfect (Le agents make mistakes w1th re$pect 10
future prices) then a temporary or shortrun equilibrium prevails Markets ciear and clearing
prices depend on fu1ure price expectations Current plans about the future are typically not
precisely implemented They will be revised as more or better information becomes available to the
economic agents See Grandmant (1982) for a comprehensive survey 01 the temporary equilibrium
literature
Wilh the exception of Gould (1985) and Pereira (19S6b 19S7d) all the models surveyed
adopt th~ concept of a perfect foresight equilibrium In turn BaliardmiddotGouldr (1965) considers a
J
2S
flexible am9unt of foresight In terms of the number of years over which price movements are ~
foreseen Pereiras model (1986b 1987b)) is flexible in a somewhat different way in that it can
indude any range of foresight from myopia to perfect foresight The choice between the perfct foresight and lemporary equilibrium is ultimately to be made on
philosophic grounds It can be argued that less than perfect expectations Imply that agents are
irralional in some way (see Auerbach-Kctiikoff (1987 p 10) However the reverse argument can
be made One can question wiether agents are reaHy rationar and perfectly knowledgable about
future prices
Recent evidene of Balird (1987) Ballard-Goulder (19851 Goulder (19851 and Pereira
(1986b 1987d) suggests that the choice in modemng expectations is an important one They show
that the degree of foresight inl0 thc future (ranging from perfect foresight to-myopic expectations)
may have dramatic impacts on the pollcy conclusions of Ihe model Accordingly the best research
strategy may b to design models which are flexible enough 10 allow for different rules regarding the
formation of expectations With the exception of the articles just mentioned sensitivity analysis
bullhave previously not been performed along this dimension
In terms of implementation the two concepts of equilibrium - perfect foresight and temporary
equilibrium have different implications The dimensionality of the equ1ibrium soll1ion algorIthm
is involved Suppose we have a model with ten markets to be run for a period of 50 years Aside
from normalization a perfect foresight model implies computing prices In 500 dimensions while a
-t~mporary equilibrium model requires sorving SO equilibria each in ten dimensions Given that
computational speed often varies with the cube of the number of dimensions the lemporary
equilibrium formulation is potentially strikingly more feasible However Ballard (1987) and
Goulder-Ballard (1985) have devoloped techniques to greatly speed the computation of a perfect
foresight equilibrium
26
j
I
4 Jrnplemenlalion and Issues on policy evaluallon
41 Calibration and equilibrium comparlslon
CGE models at typically parameterized by the use of a calibration procedure Some parameters
are exogenously given However some crucial parameters are determined in such a way that the
model replicates the data for a given base year See MansurmiddotWhalleY 11984) for an eXlensive
discussion of 1~i$ issue
CalibratIon in a dynamic context is generaliy interpreted as requiring two properties First
replication of base year data is required Second the model is parameterized to simulate an
intertemporal balanced growth path when Ihe base policy is maintaineq This Is the approach
followed by Ballard (1983) BallardmiddotGoulder (1985) Goulder (1985) and SummersmiddotGoulde
(1986) It follows the practice of BallardmiddotFuliertonmiddotShovenmiddotWhalley (1985) with their
sequential cquilibrium dynamics
Other authors Auerbach-Kotlikof (1983 1984 1987) Boyenberg (1984 1985 1986)
and Pereira (198Gb 1987d) follow whal we cali a qualitalive calibration The slructural
palametars are exogenously chosen so that lh economy follows a reasonable path into Ihe fulure
ThaI has 10 do wllh the fact thaI given the recursive nature 01 the dynamic economy and aside tom
such structural parameters only iniUa stock values are needed to run these models Given initial
conditions on the stocks of say private wealth capltal and government debt agents will optimize
and Ihereby generate a lirst round of net demands and equilibrium conditions In lurn the
equilibrium prices will determine the evolution of the stock variables into the next perIod
There are several potential problems with calibration in the conlext of dynamic models The
assumption of a steadymiddotstate growth ~ath in the base case can be questioned First while
27 I Ibullbull
steadymiddotstaU is a possibity it certainly is not the only meaningfull solution to dynamic models
Even in the case of a perfect foresight equilibrium the model implies an equilibrium path which i may or may not involve colanced growth The model not the modeller should dictate the nature 01 I themiddotbase case path Second in the context of a temporarY equilibrium path a steady state solutlcn is Inol a likely model QutcOn1tl In fact unlessmiddot expectations are stattc short run behavior consistent
with a steady-state evolution will in general not be generated On the other hand if static I expectations are self~fufjili[1g we have in fact a perfect foresight model Third even the base year
replication requirement may cause problems In Ihe cOntexl of temporary equilibrium Any
calibration parameter would be conditional on expectation rules which is probably an undesirable
feature
The nalur of the two-requirement calibration strategy is very muchin the spirit of the bull
traditional design of the comparision of alternative equilibria comparis_ion between a steady-state
base case on one hand and alternative paths including a transition period and a finat sleadyyenstate on
the other hand Pereira (198Gb 1987d) compare different (not necessarily steadystate)
equilibrium paths The arguments against such a procedure are based on the idea that the impact of
policy changes can be observed most easily since all departures from the steadymiddotstate can be
attributed to the altemalive polioy On the other hand the base case so dEfined as a steadymiddotstate is
consist~nl with previous work in a less dynamtc setting and Iherefore allows a common standard
for comparing model results
42 Computation algorithms
We are still al a stage in which basically each author uses a different computational technique
Th development of dynamic models - in parlicular wilh adjustment costs andior perfect foresight
bull
28
has corresponded with the decline in the use of fixed point algorithms In -fact given the relative
arge dimensions inevitably involved such algorithms lend to be very inefficient at the best and
chen prohibitively stow See Stone (1985) and Preckel (1965) for a comparative assessment of
different computation techniques Among the models surveyed only Ball1rdFulle(ton~
-ShavenmiddotWhalley (1985) and Feltensteln 1985 use Merrills variant of the fixed point
algorithm 1echnique
AuerbachmiddotKoHikoff (1963 1984 1967) follow a three stage procedure They first compute a
base case steady-state then a revised case steady state and finally transition path for the economy
berveen these two steady-states tn all stages a Gauss-Seidel iterative procedure is used
Ballard (1982) Ballard-Goulder (1965) Goulder (HiSS) and Summers-Goulder (1986)
~
use a method developed by Ballard and Goutder which is similar in many awects to the FairmiddotTaylQr bull
(1983) algorl1hm Short-run equilibria are calculated (using Merrills algorithm) parametric on
nrice expeC~lions The model is then iterated to generate self-fulfilling intertemporaf expectations
and the corresponding perfect foresight equilibrium In a relatively similar approach Andersson
1987 uses a simulation program SIMNON developed In the University of Lund This program
can handle two-paint boundary problems In a fashion consistent with the multiple shooting
algcrlthm (see Lipton-Poterba-Sachs-Eummers (1982))
Boyenbergs computational approach (1985 1985) differs from the other models surveyed in
that he relies heavily on analytical techniques Computations are done by using a dynamic version of
Johansons linearization method Being essentially determined by the continuous time nature of the
I I
f
model thiS linearization model has the disadvantage of confining the analysls to infinltesinal changes
around the base case equilibrium (see Bovenberg (1985) p 53) Pereira (1986b 1987d) uses an optimization algorithm NPSOL - developped by
rGillMurray-Saunders-Wrlght (1986)) This algorithm is used 10 compute the sequence of
I
29
Shortmiddotrun temporary equilibrium which makes up the intertemporal equilibrium path The
equilibrium conditions are seen as nonlinear equality constraints in the minimization of an artlcial
objective function The prices are normalized to the unit simplex by an additional linear equality I I 1 bull constraint
II ErlichmiddotGinsburgh-Heyden (1987) follow a unique approach in that they use a variant of the
optimization technique introduced by Negishl (1960) The economic equilibrium can be generated i
as a solution of a mathemallcal program Ihe objective function of which is a weighted sum of the
utility functions of the various agents while tho constraints set consists of the market dearing 1
conditions Ginsburgh-Heyden (1985) have extended Negishis resulllO tho case of downward price
lrigidities I
Finally the paper by Jorgonson-Yun (1984) is also unique in tIlat it is the only I
econometrically estimated model Different blocks for the consumption and production Side of the
-
model are separately estimated to provide the necessary structural parameters i
I The diversity of computation techniques is yet another indicator of the exploratory nature of the
body of literature surveyed in this article
43 Equal yield comparisions
The link between a base case and counteriactual simulations Is usually provided by the concept of
equal yield Shaven-Whaley (1977) discuss the meaning 01 equal yield in a general equilibrium
context when government is not allowed to run defICits Equal yield is interpreted to moan constant
public utility Government base case utility is maintained in the counferfactual experiments With
balanced budgets the concepl of equal yield is unambiguous The new equUibrium prices and the
balanced budge condition will del ermine the minimal expendilure and taxes needed to maintain base
30
case publi~ utility Accordingly in general equal yield is inconsistent with equal nominal tax
revenue Some change in tax revenue Is necessary Different tax replacement schemes ate
considered to assure that enough tax revenue is collected This is the approach essentially followed
by most 01 the papers surveyed Only Feltenstein (1985) Goulder (198) and Jorgenson-Yun
(1984) chose not to follow an equal yield strategy
The question is of how to Interprete lila concept of equal yield when the government is alowed to
fUn deflcLts The optimal leve of expenditure for base ease public trliUty can now be taAinanced
bond financed or financed by a mix of bonds and taxation We have several versions of equal yield In
particular equal yield may now be consistent with equal tax revenUE Furthermore some meaSure
ltgtf financial crowding out effects of government deficns can be inferred from the comparislon of the
several equal yield alternatives These issues are extensively discussed ~Pereira (1987b)
44 Price expectations and the dynamic generalization of compensation
Indicators
The sum of equivalent and compensation variations over households is the most widely used
aggregated measure of efficiency gains or losses In the context of $teady~state eomparisions andor
when complete future markets exist andlor perfect foresighl is assumed there are no difficulties
associated with the use of the standard Hicksian indicators In fact correct future prices are known
in these cases
If expectations are not selfmiddot fulfilling andlor future markels are not available a dynamic
generalization is necessary 8allardmiddotGoulder (1985) provide some steps in that direction by
defining an indicator that accomodates periods far into the future when households do not have
perfect foresight but a steady~state prevails On he other hand this issue is more fundamental in
i
[
I
i I I I I [
I I
bullI
31
the contex~ of a Gner1 lemporary equilibrium ftamawor~ In such Circumstances Pereira
(1986a) dev~lops a cnamlc generalization which is obtained as the present discounted value of a i i
sequence of short-fUn optimal expenditure functions consistent with a base case expected future I stream of IlWities I
32
5 Empirical evidenee from selected policy Issues
Dynamic tax models have generated several important results which escaped Sialic modellers
The following is a seleCpoundd set of issuos whIch stress the marginal benefits of dynam~ modelHng of
economic behavior in innovltlttive areas of analysis The discussion of a consumption tax emphasizes
the benefits of dynamic tusehold behavior and intergenerational aspects The study of the
flliminatlon or the re~intToduction of the investsnemt tax credits is made meaningful by the dynamic
modelling of production betlavor Modelling governmenl in a dynamic conlexts allows the modeller
to study the irrrpact of government deflcits Finally a dynamic framework lets us appreCiate the I
relevance of consumer expectations in terms of the evaluation of policy alternatives
i
i
51 Consumption Tax
t
By Its very nature a consumption tax can not be adequately investigated with a static model bull
Fullerion-Shoven-Whaicy (1983) use Ihe model of the US economy s described in
Ballard-Fullerton-Shoven-Whalley (1935) to evaluate the movement from the currenl US tax Isys1em to a progressive consumption tax Since their modeJ Incorporates a Jaborleisure choice
where leisure is an untaxed commodlty their results reflect the fact that both the consumption tax
and Ihe present system are dslortionary
Concentrating on the rntertemporaf distortions they show that sheltering more saving from lhe
current US income tax could improve econornic efficiency even if marginal tax rate increases are
necessary in order to maintain government fevenue~ At first you have a revenue shortfall
However the economy moves to a higher sustatned growth path and uitimatefy lower tax rates
cangenerate the same revenue path Also wages increase in the long run with this policy The
bull
bullbull ~
33 shyt
i
I
I
present value of welfare gi[lS for a potiey of complete savin9 deduction wilh marginal tate
adjustments (consumption lax) is simulated to be around $500 billion to $600 billion 01 1973
dollars
They also investiga1e th 1ngth of ttme it takes th~ economy to adjust to these policy changes
Roughly they estimate the long run to b thirty years although this figure Is very sensitive to the
specification of the savings elasticity - a crucial parameler in this model
52 Investment T~x Credit
Goulder-$ummers (1 S87) address the impact of eliminating the lnvestmenl Tax Cradit (ITC) on
intersectoral capital formation and on economic growth Their model isYa1ticularly adequate to
addies$ this issu~ in that is postulates a forward looking investment m9we with adjustment costs
They show that the eliminating the ITC causes a reduction in the rale of investment Inveslment is
estimated to fall by about 7 in the short-run and by about 12 in Ihe new long run steady state
In lurn a previous policy anncuncement lowers Ihe overall al1ractiveness of Investment and leads 10
a downward shift of the Investrrent profile On the other hand the combined effect of a revenue
neutral simultaneous efimiraron of the ITC and reduction of the corporate tax rates is a long run
reduction ef the capital Sieck by 35 This pattern suggests that a revenue neutral increament in
both Ihe corporale tax rates and the inveSlment lax credits would be preferable 10 the formula
adopted in Ihe US Tax Reform Act of 1986
Pereira (1987d) addresses the impact of the ITe from Ihe oland point of Ihe curTenllax syslem
under the Tax Reform Act of 1986 Given Ihe concern over the potential depressiVe effects upon
savings and investment of the new tax system in conjunction with deficit reduC1ion mechanisms of
the GrammmiddotRudman type a pertinent question is should Ih lTC be e-introduced In the context
bull
34
-of his mod~1 of the US economy Pereira focuses on the trade-off between the distortion in the
intersectoral allocation of capital induced by the lTC the lower tax revenues it generates and
potential financial crowding-out effects on one hand and the positive effects of lowering the relative
price of new capital goods on the other hand Preliminary results confirm the qualitative results in
Goulder-Summers (1987) suggestting a welfare gain of about 2 of the present value of GNP in
the best scenario of absence of replacement taxes Furthermore preliminary results show an
important time pattern to welfare gains with the average benefits increasing the further you look
into the future This pattern is due to the presence of constraints to intersectoral mobility of capital
and inertia in the adjustment towards the optimal capital levels as reflected by the presence of
adjustment costs
I
I
bullAuerbach-Katlikaff (1gB) in the context of their intergenerational model of the US economy
consider the impact on savings capital formation and interest rates of deficit policies and balanced
budget increases in government consumption They conclude that deficit finance and government
consumption can significa-ntly crowd out capital formation and lower the welfare of future
generations However crowding out from deficit finance is a very slow process because it results
from increased government spending over potentially long horizons Also deficit policies may
substantially influence the longmiddotrun interest rates while leaving the short-run rates essentially
unaffected Finally and in opposition to the central role adjustment costs seem to play in both
53 Financial crowding outeff~cts of government deficits
Goulder-Summers and Pereiras models Auerbach-Kotlikoff suggest that the time path of interest
rates induced by a policy of deficit financing seems to be insensitive to the presence of adjustment
costs
bull
35
54 The role 01 consumer expectatlons
The expectations analysis has uncovered one of the benefits of dynamic modelling
Ballard-Goulder (1985) find that the appeal of adopting a consumption tax depends on the level of
foresight possessed by the consumers Furthermore additional foresight may be welfare worsening
This is a second best type of result This tends 10 occur under policies that lead to capital deepening
and declining rate of return to capital over time To the extent that consumers have more foresight
they will be beller equippod to anticipate the fall in the rental price of capital Consequently if a
saving incentive is enacted people save more with myopia than with perfect foresight Given the
exi~tence of taxes on capita and the discrepancy between the private an~--sccial returns to capital
the greater savings with myopia is beller socially Ballard-Goulder find that the welfare gains
from a consumption tax are reduced by about 10 when we move from myopia to farleet foresight
Therefore the attractivenness of adopting a consumption tax seems fairly [obust across these
foresight specifications
I
bull
36
6 Tbe power and weaknesses 01 dynamic modelling the example 01 corporale
tax Integrallon
The Issue of corporate tax integration allows us to sh~w the stre-ngths and weaknesses or the
dynamic approach Empirical evidence using CGE techniques indicates that depending on the precise
scheme of integration and the 1ax replacement methods integratiOn may have substantial effects In
Ihe work of FuliertonmiddotKingmiddotShoven-Whalley (1980 1981) total integration was found to yield an bull
annual static eHielency gain of $4 to $8 billion in 1973 dollars Simulated dynamic gains may be as
large as $695 billion or abcu 14 of the present value of tulUre consumption and leisure in the
US economy These gains result primarily from interindustry reallocalions of Investment and an
improved inter~emporal allocation of consumption
Mora recent work emphasizes hOw consumers asset portfolio decisions and firms finanCial
decisions affect the efficiency gains from integration Slemrod (1980) focusing on consumers
asset ponfoio deCisions finds static efficiency gains which are about twice -as Iarge as those
repOrted iiy Fulierton-King-ShovenWhalley (1961) FuUrlon-Gurdon (1983) focus on the
firms financial decisions They roporl efficiency gains of 6 of GNP from the elimination of the
tax distor1ions favoring debt However when they eliminate the corporate tax and repiace it with
increaSed persona income taxes addlHonal dls10t1ions are created in the optimal labot~leisure
decisions These distortions tend to dominate the analysis slJ(h that the overall effects of complete
integration are very modest Galpr-LuckmiddotTodr (1986) address simultaneously consumers
asset pOrlfollo decisions and the firms financial decisions They also find very modest efficiency
gains from integralion
The above results are important but may be severely biased There are severa aspects of
economic behavior and modelling crucial for the study of Income tax Integration whiCh have no been
i 37
captured innny of the above papers One aspect is the absence of government deficitsmiddot a balanced
budget is assumed It is true Ihal resource crowding OUI is caplured In these mOdels but financial
crowding out induced by government spending is nOI Second investment is not derived from
optimtztlcn behavior Investment behavior passively accomodates endogenous saving decisions As
a consequence the differential impacts of policies in the incentives to save and to invest are not
captured Aso full capital mobility across secors is assumed with instantaneous capital
adjustmenlS towards optimal levels This assumption rules qut different costs of capital across
seclors and therefore differentiated reactions to tax policies changes Finallyhe modelling of both
gov9rnmerlt deficits and of endogenous real and financial investment decisions necessitates the
~nsideraiion of a dynamilt framework and the introduction of financiaJ assets govemment bonds and
iprivate financial assets A dynamic framework also highlights the efficiency effects of Integration on r
ithe ptimal intertemporal decisions The above models are either sIalic (Slemrods and
GalpermiddotltJckemiddotTode(sj or a dynamic sequ~nce of otherwise Slatic model bullbull I Pereir a (1986b) develops a dynamic applied general equilibrium model bull to study the tbull
efficiency effects of integration on the growth and allocation of investmenl across sectors Special
Iattention is paid to the structural affeelS of resource and financial crowding oul induced by
gove mect spending and deficits and to the real and financial Investment reactions across industries
to both tax integration and flnancial crowding out His mode departs from previous work on income
lax Integraticn and for that matler from most 01 the CGIE literalure for tax policy evaluation in
several direcllons It encompasses an endogenous sequential equilibrium struelure founded on
dynamic behavior with flexible expectations Government deficits are optlmaUy determined
Investment decisions are tward looking and the resull of optimizing behavior Several financial
assetsmiddot public and privatemiddot are considered
Simulation resulls Indicale that the welfare gains from Integration under large deficits are at
32
best lJlodest when measured in terms of the GNP The elimination of the corporate tax and its I replacemenl by increased income lax rates yields long run benefits which are never larger than
2 of the presenl yalue ct telr consumption and leisure onder the previous tax regime and 1
under the current tax regime However the short run w~lare effects of full integration tend 10 be
very low and in in some scenarios even negative This is a flew intertemporar pattern of efficiency Ieffects which reflects an adjJsiment lag in the interindustry investment decisions due to the
existence of costs of adjustment On the of her hand partial integration achieved by excluding I dividends from the corporale lax base syslematically yields negalive efficlencyeffeclS In Ihe long Inm these can be as high as 3 of present value of the future value of consumption and leisure This
is a new second best effect suggesllng Ihat less than complete integration may have perverse I efficiency effects
The dynamic slructure of PerclrBs model snowing for an improved freatment of real investment
declsions and government deficts provides a potential setting for a more accurate measure of the
costs and benefits of inlegration The simulation resullS suggest lower benefits and an intertempora Ibull
pattern of growing benefits
Simulation results also clearly suggest robust negative effects from partial integration How
reliable is this result If dJvidends were deductible from the corporate lax base (as are interest
payments) the preferenlial Ireatmant of debl over equily would be ellminaled Corporallons should
be expecled 10 react by decreasing the optimal deWeqully rallo However corporale financial
gecisions are exogenously set In Pereiras model as in aft of the other models surveyed that go as far
as dealing with the issue
Also withdrawing dividends from Ihe corporate tax base is a way of eliminating Ihe double
taxation of dividends at belh Ihe corporate and personnal Income levels How will the optimal
aliocellon of household saving accomodate that change in Ihe relalive return to corporale equity
39 I
t However hOusehold portfolio decisions are exogenouly set in Pereiras model as in all of the other
models surveyeltj go as far as dealing with the issue
IIt is sate to say that the evaluation 01 the benefits 01 partial Integration as defined above are
sevrey underestima1eO Meanwhile the distortionSoenerated by the increase in the marginal
personal tax rates designed to ralse the revenue foregone by the dividend exclusion are fully
aCCltlunted for A good measure of the costs of integration together With a poor evaluation of the
benefils may very well be the reason why partial integration is Simulated 10 yield neg alive
eHikiency gains
On a different vein as most of Ihe model surveyed here Pereiras is a closed economy model It
is legitimate to wonder how the resullS would change If international capital flows were anowed
IThis almost certainly would affect financial crowding out since a good d~al of the capital inflows
could be used to finance public debt
For these various reasons the results should be treated as preliminary but they strongly I suggosl that the dynamic structure provides valuable new insights into the corporate tax integration SSU2
40
7 Concluding remarks I I I i
What has been acccmplished Ihis far The success of Ihe CGE research in laelding the challenge
of dynamic modelling can not be denied
I iGreat progress in the modelling of household behavior has been achieved Promising
developm~nts in the modeUing of production and government behavior have also been accomplished
Interesting innovations in the concept of equilibrium and computation techniques were discussed
The policy analysis was greatly enriched by considering transitional effects together with longer
run steadymiddotstate equilibrium paths generalized equal yield strategies accomodating different
financial crowding out impacts and generalized dynamic policy evaluation indicators
An important set of issues have been addresed by the models surveyed Traditional issues
focusing on the effects of capital taxation and consumption taxes have been pursued In terms of
capital taxation for example the intertemporal nature of the issue was enhanced by allowing an
optimal evolution of capital stock in the economy and forward looking optimal investment decisions
In Andersson (1987) GouldermiddotSummers (1987) and Pereira (1986b 1987d) this is coupled
with a improved treatment of several tax provisions like investment tax credits and depreciation
allowances In terms of the consumption taxes developments in the specification of Intertemporal
household behavior the introduction of overlapping generations and the modelling of
intergenerational links via bequest motives as well as a closer attention to demographic evolution
provided a much improved economic setting for the understanding of the several aspects of the
problem
Furthermore dynamic modelling has permitted the addressing of a set of new issues Policy
Issues ranging from dynamic tax policy analysis to the evaluation of exogenous changes in
government expenditures to the impact of different methods of financing government expenditure to
4 1
the importance of tinanciai croHrii1g out to 1he relevance of consumorS expectations in tne policy
results have been addressed
Despile all of the progress and in part due to such progress several avenues are wide open for
much neded additional research The following seem to be the most Interesting and promising
areasmiddot First the specincation of liquidity constraints to household behavior may help to obtain a
better understanding of policy chenges that affect directly the interest rates in the economy Second
major attention needs to be paid to the incorporafion of a roraign sector (with open capital markets)
Third some efforts are needed in terms of finding adequate computation techniques now that
dynamic modelling eally makes the curse of dimensionality worse than eVOrbull
The modelling of financial markets is the single most unsatisfactory speet of the dynamic
models surveyed here~ The problems associated with provlding a mo~ adequate treatment of
financial markets and in general endogenous and optimal financial deci$ions bull both at the corporate
and at the personal levelsmiddot have 10 be addressed That necessarily requires Introducing uncertainty
into the models To b adequate this must go well beyond the mere assumption of point expeetalions
with the whole array of modelling implementalion and evaluation problems thai generates
bull I i
42
REFERENCES
Abel A and O Blanchard 1983 An Intertemporal Model of Saving and Interest Econometrica
51 pp 675-92
~ Andersson K 1987 Tar-ation of Capital in Sweden - A General Equilibrium Model Ph D
Dissenation UnIversity of Lund Sweder
Auerbach A L KotlikoH 1983 National Savings Economic Welfare and the Structure of
Taxation in M Feldstein ed Behavioral Simulation Methods in Tax Policy- Analysis Chicago
University of Chicago Press
Auerbach A l Kotlikoff 1984 Social Security and the Economics of Demographic-
Transition in H Aaron and G Burtless eds Retirement and Economic Behavior Washington DC
Brookings Institution
Auerbach A L Kotlikoff 1987 Dynamic Fiscal Policy Cambridge Cambridge University
Press
Auerbach A L Kotlikof and J Skinner 1983 The Efficiency Gains from Dynamic Tax
Reform Internatiooaj Economic Reyiew 24 pp 81middot100
Ballard C 1983 Evaluation of the Consumption Tax with Dynamic General Equilibrium
Models Ph D Dissertation Stanford University
Ballard C 1987 Tax Policy and Consumer Foresight A General Equilibrium Simulation
Study Economic Inquiry 25 pp 267-284
Ballard C and L Goulder 1985 Consumption Taxes ForeSight and Welfare a Computatonal
General Equilibrium Analysis in J Piggott and J Whalley eds New Deyelooments in Apolied
General Equilibrium Analysis Cambridge Cambridge University Press
Ballard C D Fullerton J Shoven and J Whalley 1985 A General Equilibrium Model for
t 43
Tax Eoliey Eyalua~iQO NBER University or Chicago Press
Borges A (1987) bull A Survey of Compula1ional General Equilibnum Models CECD Economic
Slud~ 8
Bovenbig L 19B3 imperfect Capital Mobility in a Perfect Foresight Herberger Model
Mimea UC Br~eley
Bovenberg L 1984 Capitel Accumulation and Capital Immobility Qmiddottheory In a Dynamic General Equilibrium Framework PhD dissertation University of California at Berkeley t
Bovenberg L 19850 Dynamic General Equilibrium Tax Models with Adjustment Costs in A iManne ed bull 1985 EtQnomrc ECUjilbrium Model Formulation and Solution Mathematical
lProgramming Siudy 23 Arnsterdam NorthmiddotHoliand
Bovenberg L t 1986 Capital Income Taxation in Growing Open Eeondmies Journa of Pubtic
Eamp(omlcs 3t pp 347middot376 IBovenberg L ald W Keller 1981 Dynamics and Nonlinearlties in Applied General
Equilibrium Models Internal Report Mlmiddot81middot208 Department for Statistical Methods CBS bull
Vooiburg The Netherlands
Charnley C 1981 The Welfare Cost of C~pltal income Taxation In a Growing Economy
IJournal of Political ECQnQm~ 89 3 pp 468middot496 I
Charnley C 19810 Efficient Tax Reform in a Dynamic Model of General Equilibrium I Mimeo Yale University I Charnley C 1982 On the Optimal Taxation In Stylized Models 01 Inlartemporal General
Equilibrium Mimeo Yale University I Decaluw B and A Manens 1985 Pays en Oeveloppement at Modeles Calculables OEquilibre
General - Une Revue de al litterature empirique Monographie Sene 9 Volume 3 Centre de
Recherche et Developpemont Economlque Unrslt de Montreal
44
bull
I I
I
I
1
Erlich $ V Ginsburgh and L van der Heyden 1987 Wher~ do Real Wage Policies Lead
Belgium bull A General Equilibrium Analysis European Economic Reyiew (forthcoming)
air R and J Taylor 1983 Solution and Maximum Likelihood Estimation of Dynamic
Nonlinear Ratioant Expeclations Models Ecooometrica 51 pp 1169middot1186
Feltenstein A 1984 Money and Bonds In a Disaggregated Ec~nonYin Searl 1-1 J Shoven
Eds Aoolied ~eoeral EqJiliDCiulD tioalYsis Cambridge University Press Cambridge
Feltenstein A 19B6 A Dynamic General Equllibrlu Analysis of Financial Crowding Out
Theory with an Application to Australia Journal of Public Economics 31 No1
Fullerlon D R Gorden 1983 A Reexamination of Tax Distorllons in General Equilibrium
Models in M Feldstein Ed Behayora Simulation Methods in Tax pOlicy Analysis Chicago
University of Chipgo Press
Fvllerlon D J Shaven and J Whalley 1983 Replacing the US Income Tax with a
Progressive Consumption Tax JoulOal of Public ECQnomics 20 pp3middot23
Fullerton D Y Henderson J Shaven 1984 A COmparision of Methodologies in Empirical
General Equilibrium Models of Taxation in Searl H J Shoven Eds ~pQlied Geoaral Jguilibrium
lIoalyss Cambridge University Press Cambridge
Fullerton 0 A King J Shaven and J Whaney 1981 Corporate Tax Integration in the United
Statesmiddot a General Equiiibdufi1 Approach American Economic Review 71 pp~ 677~691
Galper H R Lucke E Tader 1985 Taxation portlor Choice and the Allocation of Capital A
General Equilibrium Approach Mimeo
Gill p W Murray M Saunders M Wright 1986 Users Guide for NPSOL (Version 40) A
Fortran Package for Nonlinear Programming T echnicsl Report SOL 86middot2 Deparlment of
Operations Research Stanford University
Glnsbur~h V and L van der Heyden 1985 General Equilibrium with Wage Rigidities An
45
AppJication to Belgium in A Manne ed t Economic EQumprit1m~ Made Formulation and $oJioo
Mathematical Programming Study 23 Amsterdam NorthmiddotHolland ~
Goulet J 1968 1 Adjustrent CO$1S in the Theory o(the Firm Reyiew of EconomiC Studies 35 bull
pp47middot55
Goulder L 1985 Tax-Financed versus BondmiddotFinancedGhanges in Governent Spending
bull Mimeo Harva(d UniversilYmiddot
Goulder l J Shaven and J Whalley 1983 Domestic Tax Policy and the Foreign Sector The
importance of altemative foreign policy formulations to results from a general equilibrium tax
analysis model In ~n methods in Ibe antico of income from caoia1 ed Martin
Feldstein Chicago University of Chicago riSS
Goulder l L Summers 1987 Tax Policy Asset Prices and 3rowtly ~
A General Equilibrium Analysis NBER Working Pper No 212B
Grandmont J 1982 Temporary General Equilibrium Theory Ch 20 in Handbook 01
Mathemalical Eccnornics Vol II NorthmiddotHoliand
Harberger A 1959 The Corporate Income Tax An empirical Appraisal In lax Revision
Comoendjum Voll House Cornmitee on Ways and Means Washington DC Govemment Printing
Office
Harberger A 1962 The Incidence of the Corporate Income Tax JOULn1 of Political
Economy 70 pp 215middot240
Harberger A 1964 Taxation Resource Aliocalrlon and Welfare fn lb Role of Dir bullbulll and
Indjresl Taxes io tbe Fede[al Reserve s~I~Ill Princeton Princeton Unlveresity Press
James J 1985 New Developments in the Applications of General Equilibrium models to
Economic History in Piggott aand Whalley Eds
Jorgenson D and K Yun 1984 Tax Policy and Capital Allocation Harvard Inslitute of
bull
I
I f
Economic Research WP 1107
Judd K 1965 ShonmiddotRun Analysis of Fiscal Policy in a Simple Perfect Foresight Model
Journal of poJiHcal 5CCJiW1l pp 298middot319 r
Upton D J Potbc J Sachsa nd L Summers 1982 Multiple Shooting in Rational
Expectations Models fuoometrlc 50 1329middot1333
Manne A ed 1985 fconQrnjc Eoullibrium Model FOrIDujaljQo and Solution Mathematical
Pogrammlng Study 23 Amsterdam NorthmiddotHolland
Manne A 1985 Or the Formulation and Solution 01 Economic Equilibrium Models In A
Manne ed ECQOQlli-Jr~ibmiddotum ~tQdel EQlJulailon and Solution Malhematical Programming
Study 23 Amsterdam NorthmiddotHolland
Mansur A and J Whalley 1984 Numerical Specification of Applied General Equilibrium
Models in Scarf H J Shoven Eds Aoplied General Eouillbrium AnaiysectIsect Cambridge Universily
Press Cambridge
Merrill 0 1972 Applications and Extensions of an Algorithm thaI Computes FixedmiddotPoints of
Cenain Upper Semimiddotcominuou Point to Set Mappings PhD Disserlation Unlvel$ity 01 Michigan
Negishi T 1960 Welfar Economics and Exislence I an Equnlbrlum for a Competitive
Economy MelroeCOOQ11ka 12 92middot97
Pereira A 1985 On the Existence and Uniqueness of Optimal Ou1put Path lor CRTS Firms in
Adjustment Costs Technologies Mimeo Stanford University
Pereira A 1985a Consumer Expectations and Ihe COSI 01 UllUty In an Intertemporat
Framework Mlmeo Stanford University
Pereira A 1985b Corporale Tax Integrallon and Government Deficits A Dynamic General
Equilibrium Anaiysls Mlmeo Stanford University
Pereir A 1987 On Ihe Computation of the Users Costs of Stocks Ecooomi (forthcoming)
47
Pereira A bull 1987b Equal Yield Alternatives and Government Oeficlts Mimeof Stanford
University
Pereira A 1987c A Dynamic Applied General ~uHlbrium Model for Tax Policy Evaluation
Comlletamp Docurnent2tion Mimeo University of California San Diego (in progress)
Pereira A 1987d Should the investment Tax Credit be Rei(ltroduced Mlmeo University
of California San Diego (ir progreSs)
Piggott J and J Whalley eds 1985 tJew QeyeQpments in Applied General EQuilibrium
bullenalysjs Cambridge Cambridge University Press
Preckel Pbull 1985 Alternative algorithms for computing economic equilibria In A Manne Ed
Radnor R 19amp3 Equllibrium under Uncertainty In K Arrow M Intrilligaor ods
I ibull
Robinson S 1 g86 Mullisectorai Models of Oeveloping CountrieS a Survey Oivision of
Agricullure and Resources UC Borkeley WP 401
Scarf H 1967 The approximation of fixed points of a continuous mapping ~MM Jouwal of
epplid Mathornallcs 15 pp 1328middot43
Scarf H with T Hansen 1973 00 the Computation of EconomiC Eguilibria New Haven Yale
University Press
Scarf H J Shoven Eds 1984 61l1llied Genera Elujlibrium 6nalyssect Cambridge University
~ Press Cambridge
Shoven J 1976 The Incidence and Efficiency Effects of Taxe on Income from Capital
Jurnal of ramie1 Economy 84 pp 1261-83
Shoven J 1983 Applied General Equilibrium Tax Modemng IMF Slaff PaoerS 30
Shoven J and LB Simon 1986 Share Repurchases and Acquisitions An Analysis of which
Firms Participate NBER Working Paper No 2243
bull
48
IShaven J and J Whalley 1972 ft General Equmbrium Calculation of lhe Effects of
Differential Taxation of Incoma from Capital in the US Jauroe of public Econgmics 1
pp281middot321 (
Shoven J bull and J Whalley 1973 General Equilibrium with Tax$s a computation procedure
and an existence pr00f ReviEhi of Economic Studies 40 pp 475middot490
Shoven J J Whalley 1977 Equal Yield Tax Alternatives A General Equilibrium
Computa~ui0nal Tfchnique J~HJrnal of PUblic EconomIcs 8 pp 211~224 bull
Shaven J J Whalley 1984 Applied GeneralmiddotEquilibrlum MOdels of Taxalion and
International Tr8de An Introduction and Survey Journal of Economic literature 23t pp
1007middot1051
Slemrod J 1980 A General Equilibrium Model of Capital Income Taxation Ph D bullDissertation Harvard Univnrsity
Slemrod J 1983 A General Equilibrium Model of Taxation with Endogenous Financial
Behavior in M Feldstein edbull Behavioral Simulation Methods In Tax Policy Analysis Chicago
UniVerSity of Chicago Press
Stone J 1985 Sequential Optimization and Complementarily Techniques for computing
Economic equilibrium in A Manne Ed
Summers L 1981 Taxation and Corporate Investment a q-Theory Approach BroQkinas
Summers L 1985 Taxation and the Size and ComposHion 01 the Capital Stock An Asset Price
Approach Mlmeo Harvard University
Whalley J 1975 A Geneal Equiloibrium Assessment of the 1973 United Kingdom tax
reform Economica 42 pp 139middot61
1JLTIllOS JIQll~)ltG PAPERS PUBLICADOS bull
I
I I
j
112 70 - LUCENA iJiCIIJo ~ IIEnvironrilent Honitoring and Orqanization Structure T (Setellibro 1987)
nI 71 - L~ENA Diogo A Note on the Representation of Information Stxueture (SeteJObro 1987)
nQ 72 - LUCENA D1CgO Envuorment 111)nitoring and Organization
Structure II (Setelilhro 1987)
ril 73 - AlmmE3 Antonio Pais e CAsPAR Yitor bull Tributacao Inoentivos e InvesUlllentos Analise Qualitativa noYellibro 19S7)
n2 74 - liAlltOSA rose Pedro Opt1lllal lIage R1gidity A SUggestedtlthodology to Teet tb Theory and an Application (Outubro 1987)
nI 75 - CABRAL L 11 B Three lIotes on SYJIIlletric Gemes witt itsYJIIlletric Equilibri (NoveJObro 1987)
n2 76 - LUCENlt D10gOTo Search or Not to Search (Fevereiro1967)
nQ 17 - SA Jorge asconcelos euroI A Theory of SynergyB (Fevereiro1987)
n2 76 - VILARES tanuei Jose Os Be IntenaMios IlIIportados ColIIO Factor de Produao (Julho 1987)
nQ 79 - SA J~rge VasCQncel03 e HoY to Copete and Co~unicate in tature Industrial Products_ (aio 1988)
n2 80 - ROIl Rafael Learning and Capacity Expansion in a New larket Under Uncertainty (Fevereiro 1988)
n2 81 PEREIRA Alfredo llarvilo Survey of Dynamic Computational General EquilibriUlll Hodel for Tax Policy Evaluation (Outubro 1987)
Qualquer in ormacao sabre o 1IIorlltlng Papers ja publlcados sera prestada pelo Secretariado de Apoio aos Docentes podendo os ~e3OSf
ser adquirido5 ne Secyo de Vlndas da Faculdade de EconollLia UNL na Travessa Esteviio Pinto Cltpolide - 1000 LISllOA
3
numerical W~ieral equilibrium modelling designed 10 addresS tax polic issues we abstract from
several ether related areas and approaches First we abstract from analytle models along the lines
of Abel-Blanchard (1983) and Judd (1985) Second we abslract from centralized growth-type
megels of laxation like Charnley (1981 1982) Third we abstract from eentralielti models
I
associated with World Bank researchers See for example Adelman-Robinson (1978) and
Dervis-de Mele-Robinson (1981) Finally we abstract from other CIl fields See for example
ShovenVlcalley (1984) for a sUlVey ollotematlonal trade applications DecaluwmiddotMartens
(1985) ane Robinson (1986) for applications In the area of development James (1985) for a bull I I
survey of economic history applications and A Manne (1986) and BOrges (1988) far all I encompassing surveys
The models included in our comparfsion are I
I 1 8allardFuliertan-ShavenmiddotWhalley (198S) l 2 Andersson (1987)
3 AuerbachmiddotKotlikoff (1963 1987)
4 Bailarcl (1983) and BallardmiddotGould (19aS)
5 Bovenbeg (1985 1985)
6 ErlichmiddotGinsburgh-Heyden (1987)
7 FeltenSleln (1984 1985)
S Goulde (1985)
9 GouldermiddotSummers (1987)
10 Jorgenson Yun (19B4)
11 Pereira (1986b 19870)
bull 4 I
The kei features cf these models afe summarized in Table 1 which provides a structure for the
discussion Most of toe models Ihal we examine could be said to be In the Yale tradition That Is
they wer developped oy sudnts oi Herbert Scarf or students of the students of Herber1 Scar This
shyincludes BalladmiddotFuerM-Shoven-Whalley Andersson Ballard-Goulder ~)Venberg FellenSlein I
ibull Go-cider GouldermiddotSJmmers ard Pereira The exceptions from this heritage are Auerbachshybull
Kotlikoff Erllch-GinsburghHeyden and JorgensonmiddotYun I This survey is organized ltis follows rne second section provides a brief overview of I
Inon-dynamic CGE models The third seolion offers a discussion of modelling economic behavior and
equilibrium The fourth section focuses on model implemenlalion and policy ev~uatlon The fifth
Isection summarrzes empirical evidence en selected policy issues The sixth section illustrates the
power and wea~nesses of dynamic modelling with the issue of corporate laxintegration Finafly the last section summarizes the state~of-the-art in terms of dynamic modelling for lax policy
evalutlon and sugg-ess areas of future research
_
--
--- -~ ----- __-shy
_ _ - _- - Con~um9rs_ 8amphavlor
Bajla rd -Fulflonmiddot -Shaven-Whalley
1985 Anderaaon
1087
Auerba ch~ Ko1Uko 1983 1987
BaHElfdmiddotGouldor 1983 1905
Sovanborg 1985 198e
ErHchmiddotGlnaburghmiddot middotHeydeR
1987 Feltensteln 1984 1966
Gouldar 1985
GouldorSummor 1987
Jotgeneon-Yun 1984
Porelr 1966 1991
Count dynamlc dl9agg~ objecth lastes constraint ondog endog OvorlapBnqvMI of datil
year _seltlng labor no savings gonorst mollves
inlertsmporal 12 Income otility eel IntertempOfll1 yes yo no noLS 1973 Ymiddotmiddot
_ groups max ---_
0 Intertnmporal rllcufsivs
Swaden 1984 yes aggrogate utility CES eq of molion 00 yos na no consumer max for weaHh
tntortorrporal Isoeiastlc~ 00-1983
ill yo 55 poundtUS u~illy middot19831 In tcemporal yesmiddot 11El7 yes yo 00
cohorts rn~x CES-19C7 plauslblo
bYPothlC1 inlsrtamporal
19n yes 55 ago uliJity Isoelastlc Intertemporal ybullbull ye ye ysLS BFSWexpanOOd _ cohorts max
0118 inlertemporal recursivo LS plausible ye aggregate ollllly lsoolastic eq ot motion no yes no no
hYPOlhotic consumer max for weallh jnterlemporlll Boigium 1960 1983 ro aggrogate utility St()ne~Gearyjner1emporal no yo no no
consumer max Inlerl(lmporal two periodmiddot two yes
Austrs1i 1991middot1982 yes 2 income ulllly CD yeluly ybullbull tn the flrsl no no 2-perlod) groups max constraint porlod
Intertamporal LS 1973 ybullbull 55 aga utility eel intarhtmporal ye ybullbull ybullbull y
BFSW bullbullponded cohorts max
intert9mpor~1 W 1973 ves aggregate ullily 180085110 intertemporal no ybullbull no no
BFSW bullbulllMded consumer max (I~gt IntertempQ(a~
W 1955-80 ybullbull aggregate ulilily trenslog intertemporal ybullbull yes no no consumer max
Inlerlemporal recursive 1973LS yes a income ulllily CmiddotD eq of motion yos ybullbull 00 no
BFSWxpanded groups max for wealth
middot5middot
bull-- _---- -- _-----shy~ -~
I AlJlt 1 - uynafllc (lIt MOOero fur all runey tVlIIVllllQlI
dynamic dluggI utllng
I Ba~ll-a rd-fullerton
-Shoven-Vlhalley I no 19 sectors 19S5
Andarsson 1987 ye 2 sectors
~ Auerbach~KotHkof no~1gB3
1983 1967 yesmiddot1g87 1 sector
Ballard~Go uld ar 1983 1985 no 19 sectors
Bovenborg 1965 1986 yes 2 sectors
Erllch~GIn5burg hshy-Heyden no 24 seclors
1957 1__ Feltenataln 1964 196e no 30 sectors
GoUldsr 1985 no 3 sectors
~ GoutdarmiddotSumrnamprs
1987 yes 5 sectors
_ Jorgen9ClIO-Yun ona sector
1984 no produces two goods
Poralr 19S~ 1987 yos 4 sectors
~bJecllve
Produc~rs
technology behavior InpIJ1 1-0
matrix adl costs opt Imal
capital stock
sialic cost min
CES in valu$ added
110
capital labor yes 00
np Invest adjusts
to sFvnns
max 01
interlempora flot rash flow
cnsl min-lOS) Imax of valul)
of the firm~1987
ill
CD-ISBa CES-1987
capital labor
captel labor
no
no
YA quadratic in
I nO-191)3J
qutldratic in Ilk-1gB
ye
YAS
static cost min
CES in vafl1oadded
10
capita labor yes no
no invest edjuss
10 savifQs max ot capita yas ~
value ot the CD labor 00 quadratic In yo Ilrm Ilk
static CD in several no global ltlost mln vaiue added
110 ltsp1al
Jabor Ves no Inves adjusts
to MVrHIS
statle oost min
____ slatio
cost min
max of value of the
firm equivalent 10 two~stage
~ost min me) of
CO in value ad~ed
110 CES In
value added 110
CESin shyvalue added
110 dual
trBRslog price function Cobb-Douglas
capUal labor
cepHal labor
capital labor
capital labor
capital
yes
yes
yes
no
no invost adjusts no to S8vinos net of
El-xog 9011 deflcitsl no
no invest adjusts to savings r
yes quadratic In Y
Ilk
no yo I
yes Intartamporal net cash flow
In V Added 110
labof Y quadratic In I
y bullbull
- 6 -
~ --- ~ ~- -- -~------ --- --- - --~--~ ---~ shy
_
- ---
--
--------
- - - --------
H n v I taxes
olllng allowed debt Ballardfullamprton~
dynamic objective constrfnt tupendlttire (IlleUs endogenous
aU major balanced trade yoarly ybullbull rly endogenous -f ~Shovon WtlitUy US taxes with C(in~Hlt00 pubHc uility balancod cQmposilon no no debt
eiasticllv junel l 1985 rnagt eCP-REV~d1 aU fn3jO[AndtH3Son vearly balanco small opan Gcon J Swodish inturn fa to cr raturn1987 no - inaia dlJn axogonolJs 00 Ini ITnininod poundgtPfUV
-I L)~ in~h~ --~ ~~ ~pound~~---1- ----shy
runrac h~KQllfkofi ai gulD
1963 1937 US m1l3syo - intertnmponl tixogancuamp yo ys cos~d ampConoroy bandmiddotfinanced flon-oppoundIla
B iii lIa fe-Gould 9r nil llltl1gtr balanced 1rndayearly yearly crcoge)Jus
US t)X6S wiih conslanl I
1963 1965 no public utility balanced cOlYlposWon~ no nD debt max budnt ExP~RCV alastlcHy lunct
Bovenberg simp~iriad 1985middotclosod econybullbull rl~ no - balanced exogenous no no debt US laxes 19a6~two counlry1985 1986 bodgel model
Ert leh~G tn aburghmiddot yeally simplified Inlert walt funet Ishorl~run ttads Heyden no balanced exogenous 00 no Belgium
taxes _ desloullbtium flltUensleln
1967 budaet EXP=REV 6-1mprifia-dmlm costs of yoarly yo rast of tho world
1984 1985 producing a ExTax Rev exogenous financed with yo Australian ~s a Ihlrd conSUmer public good money and bonds non-optimal taxes group
Goulder yaarly annual and endogenous aU maJor
laquo1()Sed economy1985 publlc utilily In1ertemporal composillon bull yes yes U~ taxesY rna versions IlEFElQgtfpoundV bond-financed (non-opllmal balanced trademiddotGoulder-Summer yeary yoarly endogonous bull all major
1987 00 public utility balMc composition no no US taxes with constant max budGel EXPREV alasticilY fund Jl)rgenaon~Yun yoarly aU major
no balanced exogenous no1964 US taxes closed economy budGet inteftemporal recurslve endogenousPereira all major
yo public utility eq of molion lovel and yo yos1995 1987 US laxes closed economy rna for debt composition bondmiddot financed oplloll
- 7 shy
bull - -__----- _ - ------ -- ------- _-----
---- ---
------
------
ll-ltiLt 1 uynamlc Lut MOOSS fOf I ax YOtlcy tv8luallon
Inanell ~ets and maf~9ts ojumLrtm- ----_ ----c_~
privata do btl GlvldtHlI pL~lc aliocalQflo rls shy
1rlc ypa expGct assets equity retention dobl 01 savlnq$ promlums path
Sal ra rd fu Barton phys cap1 used 10 buy soquence lt
Shavon-Wha1ley no financial exogenous oxogenous no physicBl 00 bull 01 sIalic 1E myopic 1985 assets capital flqrlil -
Andersson phys cap used 10 buy 1987 no financial no no yo physical no dynamic ~ perfect
assets capital Auerbach~KolUko If phys cap usod to buy
1983 1987 neo financial no no physical cap no dyoamlc ~ perfectYmiddotmiddotnSSf)ts and Qovt debt SailBrd-Goutder phys cap
1983 1985 no financ[al assets-shy -------
Bovenberg phys cap 1985 1986 no iinenclaf
asso1s Ertlch-Glnsburgh~ phys cap1
-Heyden no financlaJ 1987 assefs
FeltoRsteln phys cap 1984 1986 Invostment is
bondmiddotfinanced
used to buy
exogenous exogenous nObull physl~al no dynamic ~ perfllcl capital
used to buy
no no no physical capital
no dynamic ~onlinuos lima
PI perfect
used 10 buy no no no physical no dynamIc PF perfect
capital jllNt) porlod) bull used 10 buy
no no ye physical cap no dynamic PI p4)rfect and govt debt (Iwo Iods) bull
Goulder phys cap used to buy 1995 no financial exogenous exogenous yenbullbull physical cap no PFlfE bull IdynamIc perfoct~assels and am dabt slaUc
- -~ - shyGoulder-Summers phy$ cap
1987 bonds and no e)(oganOU$ no used 10 buy ye dynamic PF perfect _ bullbullultv bonds MullY (exoaenous
Jorgensonyun phys cap 1984 no financial
essels -~ - ------
Pereira phys cap 1986 1967 bonds end
equHy
________
no
8xogenoos
no
exogenous
no
ye
used to buy two type$ of physical
copltal used 10 buy
go dobl bonds equity
no
no
dynamIc
-
dynamlc
~
lE
perfect
fIbullbulllble (nonperlecB
middot6middot
~-- _middot~ ___~w_ - - ___ bullbullbull___ -------- shy~
-----
------
- ---
-
IlflpnUItll tit lun pVII-y PIlUAlIV
q6H1brla oqual fax polley ovnluaU-on efUclency derlvatton Ioorllhm
calibration param~hr computation COmelfed llold tndfcaHH eHecls
Bnl jard~FuU6rtQ n base year back soll1lion steady-sta1e yo inlsrgroup inerseel or Iii
VS transilion lump-sum with Interfemporai 1965
~Shovn Whallmiddotv replication wllh exog Morrills nmamo1ers alQorilhm _~ond~middotsta1$ ears Inc CVor EV
i~noerucn sleady-sfala no agg(ogaled lnlalsectora 1967
base year back sobHon SiMNON package replication wHn oxog valiant of VS Iransilion no Indicamptor lnlertamporal
paramelsrs multiple shoo ling 6~~~l~slat8
AU8rt)ach~Kotllkot steady-slato yo Inlerguna-r a lions
1893 1llH
quatitaUve by replication a$Sium~lkml Gauss-Soidel liS Uln1WOn any rule -lilh 1111ermpOrB
sonsililJ1y I _ ~poundI~Jjnnelill $t(lfdvmiddot~luiA J3djlt15S ~~ nnrilW ~--------
a ail r nimiddotGould or bas9 9tH bac~ solution Ballard~ staadymiddotslalo )05 kltHSclo(a~
1983 1985 VS Iransllinn lump~sum dynalTlk CV inlerlemporal steadY slafe parameters alaorlthm replicaUon+ with oxog Gourder
+stead~stato (ers loc inlareel10rat BCvonborg qunmaiive by dynamic version stoadY-$Igtto yes 1985 1985 YS transition some govt eN Inlortempora
S9rHltivilv OnearizaJian method ropllcaUon assumptionl of Johansons
+steady-sfata paoo palh Erllch-Glnsbu rgh~ actual vorsus no aootagalod inlersectoral2 bftse years back solu1ion variant of Negishis
Heyden replication wIth alltogbull optlmizaUon countorfaclUal no ~odjcator inlottempora 1987 paramlJters aocroach
faltonstaln baso yeat back solution actual versus no aggrampgalfld intersElclornl replication wilh exog Merrills counlerfaclutll no indlclltor middotlntamprlempOull1984 19B6
parameters al90rithm 1981middot1982
Gouldar staady~slala Inlar5ecioralbase year back souUon Variant or 1S lronsilion no dynamic CV inlortempora
stoady state parameters Newlon 1985 feplicalion+ with iUog Fair~Taylorl
+slaildy-sla1e Intargeoorat Gtlulder~Summers base year back solution varIant of stlJadymiddotstate fixed no aggregated fntersectoral
vs trarlsition government Indicator Intertempo1al steady state oatameters Newton
repticaUon+ with exog Fair-T aylQf 1967 +ste-adv-stata bullbullbullndlno financial
Jorgenon-Yun econometric steady-stale Money melfic In1ersecloral
- - 8slimalion vs traosUian no social weUare functton intertampotal1984 +~teadlmiddotstatl)
~ shy
qtJalitaUv8 by NPSCt optimal path ye generalized inlersectorai 1986 1987 Perelr
replication 8S$umptlonl optimizatIon V$ lumpmiddotsum dynamic lntertemporat se~~ttilitf ____Jllgorilhm opllmal pah pen inc CVoclV financial
- 9 shy
bull ~ ___ - _______ - ____ bull___ OM bullbull __ bullbull __~____ ~_
--
- - - --------
--------- - - - --------
---
1nltJ1 bull I ~ 1J11ltI1 v I ~
-
Sa Ita rd-Futlertonmiddot intargration consumption lax ~StHlven~Whall~y VAT capila~ or labor tax
Ch~mqflS1985 Andersson
consumpion tax neutral corporals lax system 1987
it ue rb a c h~ i(at II Sltotf COIlSUft1ptian lax (poundpird ItS hibor Mx~on ufocto of taxation on capital fOfrnJilti1
dvnamio fiscal policv effects af OQvernmanl defioit (3tlard-Goulder
1883 1987
allocts 01 the adoption of a 1ge3 1985 pure consumpHon lax
im~rtmce of cons fCreSClht -Bovenborg analytically orianiad approach
1985 1986 effects ot pure conumption tax capllal income taxation in Opo economies
Erllch-Glnburghshy-Heyden reat wage policios in Belgium
1987 Felter1steln 1984 - theorotical
1985 ~ linnncial crcwdlng out1984 1986
Goulder effects of tax- vs bond-financed 1985 chanQes In pubtlc expondllufe
~~--
Gouidr~SUmmampfI torpora1$ lax cvts reduced 1987 ITO increase 10 gasolioe taxes
Jorgenson-Yun
effects ot several programs of tax reform on the allocation of ~~ptai
1984
-- Purelra
corporate la)( Integration effects cf romiddotinlroducinQ imreslamnl tax credits
1986 1967
10 bull
bull -~- ~- --- _ - ---____--_ - -----_-_ --~ ---- ~
1 1
2 Brief overview of non-dynamic CGE models
The origin of most or the models we survey goes back to the work of Scarf (1967 1973) whlgt
first developed a reliablo algorithm to compute equilibrium price for a ArrowmiddotDebreu economy
HiS algorithm used simpicial Bubdivision techniques and can -be shown to be the computational analog
of the fixed point theorems previously used to prove the existence of equilibrium His technique
could soW a model with an arbitrary number of consumers and commodities as long as all agents
were price takers consumers were subject to budget constraints demands were continuous andmiddot
production did not display increasing returns to scale The alQorithm while guaranteed to converge I
I I I
was relatively slow for prcb1ems involving more than S1If 20 dimensions A maior improvement in
computational speed Woo offered by Merrills algorithm (1972) which used tpe same fundamental
ideagt as Scarfs procedure
Scarfs model (as is true for the standard Arrow-Debreu model) does not include a government
sector ~ neither taxes nor public goocs At one of the mosi promising appHciUions of the new
computational techniqu w~s in Iha area of tax policy evaluation Shoven and Whalley (1973)
exlended the general equilibrium model and computational approach to include a wide rry of taxes
and a governl11ampnt spending plan The original ShavenmiddotWhalley model was static (although the
different commodi~ies CQuid be considered simitar goods available at different datest as In
Arrow-Debrau) and government was aumad to run a balanced budget The data are arranged as a
social accounling matrix The models spacification and calibration is checked by solving it in the
presence of the base set of laxes The result should ba exactly tha initial social accounting matrix
After having pasd this replication check the model is solved for a counterlactua equilibrium In
the presence of a new tux design The result is once again a social accounting matrix The two
equilibria are compared in order to tSS($S the impact of the new tax plan The first uses ot this
bull
1 Z
model for lax policy evauajon were ShovenWhalley (1972) Whalley (1975) and Shoven
(1916)
Completely static models as Shaven-Whalleys are unsatisfactory for many tax reform ISsues
These include corporate lax Integration effects of investment tax credits effects of accelerated
depreciations consumer Or e)penditure tzxes~ importance of saving subsidies like IRAs etc These
are essentialy dynamiC ksues They involve not only the allocation of capital across sectQfsbut
perhaps more Important the capital intensity of the economy But the capital accumulation and
capital reallocation take time and may involve adjustment costs Because of these issues Ballardmiddot
Fulierton~Shoven~Whaloy took the first steps towards developing a dynamic model
In this conlexl the ShovenWhalley model has been extended and implemented for Ihe US
-economy In BuIIBrdmiddotFulienonmiddotShovenmiddotWhalley (1985) While Ihis book completely documents the
mo~el it was used in several publicalions beginning in 1978 Their mode) consists of nineteem
production sectors twelvo households and fifteen consumer goods It includes a very detailed set of
taxes including the federal and state personallncome taxes federal and state corporate income taxes
SOCial Security taxes pro1J0rty taxes unemployment Insurance excise taxes sales taxes etc It
has been calibl att-d to reproduce the 1973 US economy Recentty a version corresponding to the
1983 JS economy has been developed
wThe 8alard-Fullenon-Shoven Whalley model is dynamic In the sense that consumers face a
choice between current consumption and leisure versus future consumption (which can be
purchased via s wlngs) The consumer classes act as if they were maximizing a nested CES utility
function over the domain of a CES aggregate of contempo-raneous consumer goods and leisure and
future consumption subject to their income constraint The parameters of those funcHons
determine the shares of income devo1ed to each commodity to saving and to the purchase of leisure
They also determine the two key elasticities in the models ~ the elasticity of labor supply with
bull
13
respecl to tgte rbullbull1 aoer lax wage rae and the elasticity of saving with respecl 10 the real after lax- rate of retum to capital
In the model consumerS have myopic expectations regarding lulure prices and in particular
regarding the future rale of return 10 capilal Ballard (1983) and BallardmiddotGoulder (1985)
inco~ora1e both perfect foresight and limited foresight into this model Future consumption is
acquired by buying a fixed composition portfolio of real investments that offer an infinite annuity
of returns
The production side of Ihe model is compielely static The mod1 Incorporates a constantmiddot
olasicily 01 substitution betweeen prlmary inputs in production (capital and labor) and fixed I coefficients for intermediate inputs The model distinguIshes between Industrial outputs and
eonsumer goods for the simple reason Ihat tile data are classified differenUy Industrial sectol$ I I involve such categories as forestry and fisherle metal mino9 and publishing and printing while
consumers purchas-3 fUfHi1ure 2ulomoblles and books This fact is recognized in the model by Ibull
incorporatirg a secord stage Of production which converts industrial outputs intO consumer goods
This technology is uStally mod9-led as a flxed~coefficient conversion matrix
The BallardmiddotFINton-ShovenWhalley model assumeS that the private sector finance
marginai ioveslment with the same composition of debt and equity as currently exists in each sector
This is the same arsumrtion that Herberger originally used Investors all hold debt and equity in tile
same proportion Thampreiore this ownership can be aggregated into simply capital ownership For
tax purposes however the separate treatment of debt and equity is taken into account at both the
corporate and persofa Io=vel SimllarlYJ the dividend poIicies on corporale equitY are established
exogenously There are no government bonds in the model since there are no government deficits
The modol is solved for a sequence (as many as 100) of temporary equilibria wltll consumers
allocating income between present and future conumption at each point in time The palh for 1he
14
economy ~ a sel of connected equilibria The connectIon is pra~ided by capital accumulation
Capital accumulation is endogenous and determined by saving The model stans with a sociagt
accOunting mi1rix In the bzse case the economy is assumed to be in a steaqy state growth -path (along
which all relative prices ~fe constant) The madel solves for both the naw steady state growth path
and the transltion 10 it zf~er a policy intervention rhe authors have frequently addressed the
question-of howong rt tKes to effectively settle un10 a new steady S1ate growth path
The dynamics of the model are limited however in that future consumption is collapsed into a
composite commodity Aso the absence of government deficits and the lack ofproductlon dynamics
limits the realism of the BallardmiddotFuliertonmiddotShoven-Whatley model for the analysis of dynamic
policy issues (such as the adoption 01 a consumption tax or Ihe elimination of the investment tax
credit) The models we survey try to Improve upon Ihe Ballard-FlIiertonmiddotShaven-Whalley
dynamic modelling The models Jn~orporate more forward lookTng behavior They improve the
dynamics on the conampumrtkm side of the economy and some of them introduce true 1ynamlc behavior
on the part oI the producers We turn now to the issues inY9lv_ed in developing such dynamic models
15 I I ~3 ~ssues in the modelling Of dynamic behavior
We will 100k firt at the speclficalion of economic behavior of consume(s~ producers
goyemment and Ihe r~sl of Ihe world The modemog of corporale and household financial decisions
I is then addr5sed Finally Ihe eencept of equilibrium is discussed
I
i
I 1
31 Consumers behavior
Early efforts 10 build dynamic features into the economic behavior 01 consumers are due to
Ballard (1963) and Auerbach-Kotllkoff (1983 1984) Their work has been closely followed by
several subsequent authors In faet dynamic household behavior Is the single most pervasive aspect
amohg the models surveyed
All the models Inc-O1orale some loro of lila-cycle behavior Household behavior is dewmloed
by the maximization of an addWVely separable time invariant intertemporal utillty function The
utility lune1ion is defined oyer the domain of the consumption goods In the economy In most of the
models leisure is also an argument in utility so that labor supply is optimally determined
Utility maximi4ation is subject to a lifetime intertemporal budget constraint which equalizes
the presen1 value of consumers income and expenditure More reeentiy in Andersson (1987) bull
Bovenberg (1985) and Pereira (1986b 1987d) the constraint i$ defined as a sequence 01
recursive equations of motion on wealth This has the potential advantage 01 accomodating liquidity
constraints However It should be recognized that in the absence 01 liquidity constraints (ie when
consumers are free to completely borrow against future income) the two specificallons of the
househofd ccnstraint are essentiay equivatent Furthermore tn both versions saving is optimally
determined as a way of translaring wealth intertemporally
1 6
Some lodels now have a sophisticated dynamic specification of hoosehokl In Ballard (1983) -
BallardmiddotGoulder (1985) and Auerbach-Kotlikof (1983 1984 1987) consumers behavio is
embedded in an iterQEneraional setting Each of them has S5 age cohorts simulataneously alive
Som-e measure of jntergener1ional altruism is considered in the form of bequest motives In Ballard
(1983) and Ballard-Gouldr (1985) consumers derive utility directly from bequthing parl of
their wealth Therefore bequests assume 1he form of a scrap value of terminal wealth of a
generation In AuerbachmiddotKotlikof (1983 1984 1987) each generation empathyz bullbull with Ihe
utility of future generations ovar bequethd wealth In addition Ballard (1ge3) includes detailed
demographic poiections wfill the intent of incorporating in the policy analysis the effects of the
bull Post World War tI baby-boom
In the real world household decisions also include the optimat allo~atiOn of saving among
shyalternative physical or financial assets Theories of household portfolio behavior are relatively
Iwen~establshed However they involve uncertainty as a crucial element Accordingly they have
middotnot been incorpora~ed in the context of the determinIstic dynamic models we sutyey here We will 1 Come backt-O this Issue below
32 Producer behavior
The efforts to build dynamic features into the economic behavior of producers are more recellt
and less widely adopted The first attempts are due to Bovnberg (1984 1985) and Summers
(1965) Dynamic behavior has been more fully incorporated In the recent models of Andersson
(1987) Auerbach-Kotlikoff (19B7) GouldermiddotSummers (1987) and Pereira (1986b 1987d)
Part of the reason why production side dinamlcs has been more slowty adopted is the weak
supply of accepted theories referring to the dynamic behavior of the firms In the models referred
I 17
to above gynamlc producion and investment b~havior are middotinduced by the existence of capital
adjustment costs and linked to Tobins q theory Adjuslment costs are designed to capture both the
incomplete mobility of capital aoross industries and installation costs ~e the costs of adjusting
capital towards its cpHmai level)
Adjllstment costs can be conceived as internal to the firm and measured in terms of foregone
outpul along the lines of Lucas (1967) Alternatively adjustment costs can be viewed as actual
external cests incurred together with Ihe purchase costs along the Unes of Gould (1969) With the
exception of Pereira (198Gb 1937d) all of the CGE authors follow the former approach
The firms in the economy maximize their market value as the present discountad value of the
future stream of dividends In Andersson (1981) and Pereira (198Gb 1987d) firms are seen as
maximizing the present discounted value of net cash flow Maximization is constrained by Ihe
adjustmenl cosl lechnology and ~n equation of motion describing Ihe evolution of the capital stock
It should be noted that undereogenous divldendrelention rules the two problems are equivalent II
should also be nOled Ihat a salisfectory economic rational for th bullbullistoo of dMdonds with the
present tax code is missing in the profession (Shoven (1986))
The models with static formulation of producers behavior are characterized by passive
investment behavior Investment merely accomodates to saving in the eccnomy With a dynamic
formulation induced b adjustment costs real investmen decisions are forward looking Investmenl
Is endogenously and optimally determined by the firms A fundamental difference between the
shonmiddotrun in which capital stock is given and longmiddotrun in which Ihe level of capital is allowed to be
optimally determined is emphasized
This extra richness of produc1ion dynamics is not without costs A careful look at the summary
tables will clearly show an inverse relation between the adoptlon of dynamic features in production
and the level of disagreggation of the production side of the economy In fact with production
bull
j
1 6
dynamics $e dimension of the problems is immensly increased Let IS be more specific
In a static framework with constant returns to scale production technology the output level ismiddot
indeterminate The optimal aliocation of inputs can be obtained by cost minimization with any
feasible output level generatillg zero profits Such zero profit conditions are used to solve for the
output prices in terms of the factor prices and hencemiddot to reduce the dlmensionality of the problem
from the -number of commodities and factors to the number of factors Thus even if the model deals
with 30 production sectors the computation of an economic equilibrium can take place using only
the dimensionality of the primary inputs in the economy
Now under certain regularity conditions on the production and adjustment costs technologies
leading to enough concavity of the optimality objective the intertemporal output path for the firm is
endogenously optimally and uniquely determined even with constant relums to scale technologies
(see Pereira (1985 1987a) Accordin~y with adjustment costs in the model optimal profits
will in general be non-zero This result has crucial implications for the computatiQn of equilibrium
in the model~ with production dynamics With adjustment costs no reduction of dimensionality is
possible The curse of dimensionality returns as a binding constraint
The introduction of adjustment costs adds another significant complication to the model With
capital being less that perfectly mobile in the economy different rates of return on capital will
exist in different sectors This is a difficult problem to tackle conceptually in the absence of
uncertainty Also in the real world producer maximization choices include also its choice of
financial ratios (debVequity) and payout rates (dividendretained earnings) These financial
subjects are difficult and much Clf this behavior is still taken as exogenous by the modellers We
will come back to these issues below
middot i
19
bull 33 government behavior
middot
Two issues dominate the modelling of government behavIor First govemment behavior has
Ibeen typically seen in Ihe CGE liIerature for lax poliCY evaluation as constrained by yearly balanced
I budsets (see for example 8allard-Fullrton-S~ovn-Whafley 098Sraquo The analysis of
govemment defidlS and public debl in a CelE context requires a dynamic selling Second the level of
government expenditures is either exegenouty given as in Auerbach-KoUikofi (1984 1987)
6ovenbrg (1984 1985) Feltenstein (1984 1986) and Jorgenson-Yun (1984) or
endogenously (but no optimally) determined by the balanced budget conditions as in
6allard-Fullerton-Shoven-Whalley (1985) Andersson (1987) Erllch-Ginshurgh-Heyden
~
(1987) Gouder (19aS) and Gouder-Summers (1987) In the second case the composition of
bull public expenditures is often optimally determined However the leve~ of government expendi1ures
can only 00 endogenously and optimally determined if the government is seen as an optimizing agent
and is allowed to run deficits
The first attempts to deal wiU the government defiCits in CGE tax models are due to
Auerbach-Kollilltoff (1984) and Feensain (1984) In Auerbach-Kotlikof (1984) government
expenditures afe exogenous They grow at the rate of growth of population However given 1he tax
structure and tax revenues yearly deficits and surpluses are atTowed subject to an in1ertempond
constraint that the present value of future tax revenues equals the present value of future
expenditures
In Fellenstein (1984 1986) government expenditures are also exogenously given He also
allows the government to run deficits to finance expenditures in excess to 1ax revenues However
surpluses are returned to CQnsumars in the form cf transfers Accordingly government ts not
subject to any constraint regarding the future repayment of public debt
In GouJder (19851 the 9Vernmenl maximizes a static social welfare ~nctio subject to a-
bull
20 I i
I
I I I
I bull
balanced budget conSirtlL This follows the optimal alocatkm of government expenditures alpng
the lines of Baliard-FHeiton-Shoven-Whalley (1985) Ttle model Is generallzea to allow for
exogenous changes in HIe time path of government expenditures Two financing alternatives are
consjdeft~d and contr8s1ed EldditionaJ tax revenues and bond issuance
Pereira (198Gb 1137d) auempts to address both the incorporation of deficits and the
determination of govement expenditures The path of government expenditures and the path of
delicilSlsurpluses (ano therefore the path lor debt) are endogenously and optimally determined The
government is seen as mltJximizing n intertemporal social welfare function given the 13x structure
Optimization is subject to a sequence of recursive equations of motion reflecting the evolution of the
ptlbllc debt allowing fDr government budget imbalances It should be stressathat this specification
01 the government cOMtcalnt is equivalent to the specificaUon in Auerbaeh-Kotlikoff (1983 1987)
when no liquidity constraints affect -gove~nment behavior
The extra richness in the treatment of government behavior allows tfte examination of
government debt poices in a truly dynamic setting Also financial crowding out effects induced by
government deficits can be analyzed (see Auerbach-Kottlikoff (1987) Feltenstein (1986) and
Perer (198Gb 1987e))
34 Foreign sector
Most of the open economy models follow the assumptions of balanced trade with import and
export net demands characterized by constant elasticities along the line of Ballardmiddot
-Fullerton-Shaven-Whalley (1965) Such is the case of Salfard (1963) Ballard-Goulder
(19S5) and Goulder-Sllmmers (1987)_ In Feltenstin (1985) the rest of the world is treated as
21
an addjlion~ consumer group
Bovcoberg (1986) develops a model In which two economies lIr considered each following
intertempoml perlect foresight paths These economies meet in the international forum Their
trade relaroMhips are characterized by yearly balanced trade accounts
None of the new generation of dynamic cae tax models has yet IncorPorated the lnternational
capital flows as done in aoulder-Shoven-Whalley (1983) for the earlier Ballarrshy
-Fullertonmiddot Shaven-Whalley (1985) model This is unfortunate as there is an important
Intertmporal aspect to international lending The first al1empts along Ihese lines are due to
Andersson (1987) and Erlich-Glnsburgh-Heyden (1987) Andersson (1987) in his model of the
Swedish economy adopts a closed economy appraoch in which rates of return in the domesJic
economy are largely determined by the international capital markets Fidinterest rate induce
Intlnatlonal capital flows which determine and finance the International lrade imbalance In
lurn In ErlichGinsburghHeydenmiddot (1987) foreign trade Is generaled according to an
intertemporal Irade welfare lunction with constant import and export elasticities In the shortmiddotrun
they allow inlernational trade inbalances which generate capilal ftows to the domestic households In
the long run however trade balance is assumed
35 The need for financial markets
A dynamic economic Wucture not only provides the ideal environment 10 model many features of
economic behavior it also permits one to Incorporate the financial side of the economy If the
government is allowed to run deficits the question 01 deficll financing automatically follows If
investment is optimally dewmined and returns to capital are different aClOSS sectors problems of
investment financing arise If there are seeral final1cfal assets In the economymiddot government
4-
I i
ibull
I I I I
I
j
22
bonds privpte bonds and equity (or simply physical capital installed inmiddot different sectors) the
problem of allocation of saving among assets with potentially different returns arises
The papers surveyed var greatly with respect to l~e extent of their attention to the financial
side of the economy At Clne extreme are the models in ~anardmiddotFuliertonmiddotShoven-Whalley (1985)
Andersson (1987) Ballltld (1983) Ballard-Goulder (1985) Bovenberg (1985 1986) and
Erlich-Gjnsburgh-Heyden C19B which are devoted exclusively to the real side of the economy
In turn Auerbach-KotliKoff (1984 1987) Feltenstein (1984 1986) and Goulde (1985)
allow for government debt In these models saving finances changes in government debt and physical
capital Private and publlc 2ssets are perceived by the households as perfect substitutes The
allocation of saving merely adjusts to the relative demands for funds
Feltenstein (1984 1986) is the only model surveyed which introduces money Government deficits are financed by issuing money and bonds according to an exogenously given rule Money is
demanded by consumers for transaction motives and an exogenously given fraction as a store of value
On the other hand government bonds and physical capital are the vehicles for the intertemporal
transfer of wealth
Summers (1985) and GOlldermiddotSummers (1987) introduce a whole menu of financial assets shy
firm specific equity capital Different assets earn different rates of retum However such rates
are equal up to constant and exogenous sector speCific risk premia Therefore the introduction of
constant exogenous risk premia as helpful as it may be in the context of calibration does not solve
the main issue of the non-optimality of the allocation of saving Also talking about risk premia in a
deterministic context is somewhat unsatisfactory
Pereira (1986b 1987d) also introduces a whole menu of assets private and public bonds and
firm specific equity However all the assets are expected to yield the same rate of return and
therefore perc~jved as perfect substitutes The different asset types allow consideration of
23
exogenOU$~8tUequjty triO dividendlrc~enlion rufes and therefore seve-ral sources of investment
financing bond equity and retained earnings
The nonmiddotcptimalHy of the allocation of saving and the absence or exogenelty of corporate
financial rules is _a reftectlon of the limitations of the determInistic approach Either in a context of
perfect anticipation oi the future prices Of in general within the t$atm of point price
expectations all the papers follow a deterministic approach Under such circumstances consumers
either expect diHerent rales of return (inclusive of risk premium) across assets in which case
they wilt buy only one asset (that with highest rate) Of they expect equal rat~s of relurn in which
case Ihey are indifferent about the asset composition of Iheir portfoliO There is no way of
Iradingmiddotoff rales of return and risks to obtain an optimal interiof solution to the problem of the
allocation of saving
The most advanced contribution in the modelling of saving allocation in a CGE setting is due to
Slemrod (1980 1983) in the context of a static one-period model In his model consumers act
according to a two stage separable decision process They Hrsl decide on how much to save Then
they decide on the allocation of saving according to an Indirect uttlity function dependent on the rates
of ratum and variances offered by the different assets in the eoonomy The sourCe of riskiness in the
ampConory comes from an uncertain marginal product of capital On the other hand aside from
portfolio decisions the res of the economy is insulated from uncertainty
The most complete contribution in terms of the treatment of the corporate financial rules is
FulienonmiddotGordon (1983) In a variant 01 Ihe BailardmiddotFulienon-ShovenmiddotWhelley (1985) model
Fullerton and Gordon have capital intensity and optimal financial decisions Jointly deter~ined
through a two siage process The cosl of financing capital is minimized by trading off the tax
advantages of debt against the epected real bankrupcy coSls inherenl with high debVequity ratios
Given the optimal debtequity ratio the level of inveSlment Is chosen such that at Ihe margin the
24 l
bull
rerurn pn ~uity equas the return on bonds plus an exogenous risk premium
36 Market assumptions equilibrium and expectations
Virtually all of the papers arc characterized by Walrasian market clearing assumptions All
markets are perlect1y competi1ive Atomistic competition among agents 1s assumed even thoughmiddot only
a flnile number of agents is considered Virtually no mark~t disequilibria or price stickiness afe
considered The only exception is ErlichmiddotGinsburghmiddotHeyden (1987) bull In thei model of the Belgium
economy I the _wage rate is fixed in the short-run Therefore in the shOrlwrun desequilibrium in
the labor market will generate endogenous unemployment However in lh$ long-run all prices
Including the wage rata are Ilexible and accordingly ali markets ciear
Given the dynamic nature of behavlor in the economy markelmiddotcfearing prices in each period
depend on expectations of fture prices and on tax variables In the economy There are essentially
two ways at interpreting the economic equilibrium in such a dynamic conte~bull If future prices are
perfectly anticipated (io expectations are self-fulfilling) a perfect foresight equilibrium
prevaHs Then future actions are merely the implementation of current decisions for future
periods However jf ptica expecla110ns are not perfect (Le agents make mistakes w1th re$pect 10
future prices) then a temporary or shortrun equilibrium prevails Markets ciear and clearing
prices depend on fu1ure price expectations Current plans about the future are typically not
precisely implemented They will be revised as more or better information becomes available to the
economic agents See Grandmant (1982) for a comprehensive survey 01 the temporary equilibrium
literature
Wilh the exception of Gould (1985) and Pereira (19S6b 19S7d) all the models surveyed
adopt th~ concept of a perfect foresight equilibrium In turn BaliardmiddotGouldr (1965) considers a
J
2S
flexible am9unt of foresight In terms of the number of years over which price movements are ~
foreseen Pereiras model (1986b 1987b)) is flexible in a somewhat different way in that it can
indude any range of foresight from myopia to perfect foresight The choice between the perfct foresight and lemporary equilibrium is ultimately to be made on
philosophic grounds It can be argued that less than perfect expectations Imply that agents are
irralional in some way (see Auerbach-Kctiikoff (1987 p 10) However the reverse argument can
be made One can question wiether agents are reaHy rationar and perfectly knowledgable about
future prices
Recent evidene of Balird (1987) Ballard-Goulder (19851 Goulder (19851 and Pereira
(1986b 1987d) suggests that the choice in modemng expectations is an important one They show
that the degree of foresight inl0 thc future (ranging from perfect foresight to-myopic expectations)
may have dramatic impacts on the pollcy conclusions of Ihe model Accordingly the best research
strategy may b to design models which are flexible enough 10 allow for different rules regarding the
formation of expectations With the exception of the articles just mentioned sensitivity analysis
bullhave previously not been performed along this dimension
In terms of implementation the two concepts of equilibrium - perfect foresight and temporary
equilibrium have different implications The dimensionality of the equ1ibrium soll1ion algorIthm
is involved Suppose we have a model with ten markets to be run for a period of 50 years Aside
from normalization a perfect foresight model implies computing prices In 500 dimensions while a
-t~mporary equilibrium model requires sorving SO equilibria each in ten dimensions Given that
computational speed often varies with the cube of the number of dimensions the lemporary
equilibrium formulation is potentially strikingly more feasible However Ballard (1987) and
Goulder-Ballard (1985) have devoloped techniques to greatly speed the computation of a perfect
foresight equilibrium
26
j
I
4 Jrnplemenlalion and Issues on policy evaluallon
41 Calibration and equilibrium comparlslon
CGE models at typically parameterized by the use of a calibration procedure Some parameters
are exogenously given However some crucial parameters are determined in such a way that the
model replicates the data for a given base year See MansurmiddotWhalleY 11984) for an eXlensive
discussion of 1~i$ issue
CalibratIon in a dynamic context is generaliy interpreted as requiring two properties First
replication of base year data is required Second the model is parameterized to simulate an
intertemporal balanced growth path when Ihe base policy is maintaineq This Is the approach
followed by Ballard (1983) BallardmiddotGoulder (1985) Goulder (1985) and SummersmiddotGoulde
(1986) It follows the practice of BallardmiddotFuliertonmiddotShovenmiddotWhalley (1985) with their
sequential cquilibrium dynamics
Other authors Auerbach-Kotlikof (1983 1984 1987) Boyenberg (1984 1985 1986)
and Pereira (198Gb 1987d) follow whal we cali a qualitalive calibration The slructural
palametars are exogenously chosen so that lh economy follows a reasonable path into Ihe fulure
ThaI has 10 do wllh the fact thaI given the recursive nature 01 the dynamic economy and aside tom
such structural parameters only iniUa stock values are needed to run these models Given initial
conditions on the stocks of say private wealth capltal and government debt agents will optimize
and Ihereby generate a lirst round of net demands and equilibrium conditions In lurn the
equilibrium prices will determine the evolution of the stock variables into the next perIod
There are several potential problems with calibration in the conlext of dynamic models The
assumption of a steadymiddotstate growth ~ath in the base case can be questioned First while
27 I Ibullbull
steadymiddotstaU is a possibity it certainly is not the only meaningfull solution to dynamic models
Even in the case of a perfect foresight equilibrium the model implies an equilibrium path which i may or may not involve colanced growth The model not the modeller should dictate the nature 01 I themiddotbase case path Second in the context of a temporarY equilibrium path a steady state solutlcn is Inol a likely model QutcOn1tl In fact unlessmiddot expectations are stattc short run behavior consistent
with a steady-state evolution will in general not be generated On the other hand if static I expectations are self~fufjili[1g we have in fact a perfect foresight model Third even the base year
replication requirement may cause problems In Ihe cOntexl of temporary equilibrium Any
calibration parameter would be conditional on expectation rules which is probably an undesirable
feature
The nalur of the two-requirement calibration strategy is very muchin the spirit of the bull
traditional design of the comparision of alternative equilibria comparis_ion between a steady-state
base case on one hand and alternative paths including a transition period and a finat sleadyyenstate on
the other hand Pereira (198Gb 1987d) compare different (not necessarily steadystate)
equilibrium paths The arguments against such a procedure are based on the idea that the impact of
policy changes can be observed most easily since all departures from the steadymiddotstate can be
attributed to the altemalive polioy On the other hand the base case so dEfined as a steadymiddotstate is
consist~nl with previous work in a less dynamtc setting and Iherefore allows a common standard
for comparing model results
42 Computation algorithms
We are still al a stage in which basically each author uses a different computational technique
Th development of dynamic models - in parlicular wilh adjustment costs andior perfect foresight
bull
28
has corresponded with the decline in the use of fixed point algorithms In -fact given the relative
arge dimensions inevitably involved such algorithms lend to be very inefficient at the best and
chen prohibitively stow See Stone (1985) and Preckel (1965) for a comparative assessment of
different computation techniques Among the models surveyed only Ball1rdFulle(ton~
-ShavenmiddotWhalley (1985) and Feltensteln 1985 use Merrills variant of the fixed point
algorithm 1echnique
AuerbachmiddotKoHikoff (1963 1984 1967) follow a three stage procedure They first compute a
base case steady-state then a revised case steady state and finally transition path for the economy
berveen these two steady-states tn all stages a Gauss-Seidel iterative procedure is used
Ballard (1982) Ballard-Goulder (1965) Goulder (HiSS) and Summers-Goulder (1986)
~
use a method developed by Ballard and Goutder which is similar in many awects to the FairmiddotTaylQr bull
(1983) algorl1hm Short-run equilibria are calculated (using Merrills algorithm) parametric on
nrice expeC~lions The model is then iterated to generate self-fulfilling intertemporaf expectations
and the corresponding perfect foresight equilibrium In a relatively similar approach Andersson
1987 uses a simulation program SIMNON developed In the University of Lund This program
can handle two-paint boundary problems In a fashion consistent with the multiple shooting
algcrlthm (see Lipton-Poterba-Sachs-Eummers (1982))
Boyenbergs computational approach (1985 1985) differs from the other models surveyed in
that he relies heavily on analytical techniques Computations are done by using a dynamic version of
Johansons linearization method Being essentially determined by the continuous time nature of the
I I
f
model thiS linearization model has the disadvantage of confining the analysls to infinltesinal changes
around the base case equilibrium (see Bovenberg (1985) p 53) Pereira (1986b 1987d) uses an optimization algorithm NPSOL - developped by
rGillMurray-Saunders-Wrlght (1986)) This algorithm is used 10 compute the sequence of
I
29
Shortmiddotrun temporary equilibrium which makes up the intertemporal equilibrium path The
equilibrium conditions are seen as nonlinear equality constraints in the minimization of an artlcial
objective function The prices are normalized to the unit simplex by an additional linear equality I I 1 bull constraint
II ErlichmiddotGinsburgh-Heyden (1987) follow a unique approach in that they use a variant of the
optimization technique introduced by Negishl (1960) The economic equilibrium can be generated i
as a solution of a mathemallcal program Ihe objective function of which is a weighted sum of the
utility functions of the various agents while tho constraints set consists of the market dearing 1
conditions Ginsburgh-Heyden (1985) have extended Negishis resulllO tho case of downward price
lrigidities I
Finally the paper by Jorgonson-Yun (1984) is also unique in tIlat it is the only I
econometrically estimated model Different blocks for the consumption and production Side of the
-
model are separately estimated to provide the necessary structural parameters i
I The diversity of computation techniques is yet another indicator of the exploratory nature of the
body of literature surveyed in this article
43 Equal yield comparisions
The link between a base case and counteriactual simulations Is usually provided by the concept of
equal yield Shaven-Whaley (1977) discuss the meaning 01 equal yield in a general equilibrium
context when government is not allowed to run defICits Equal yield is interpreted to moan constant
public utility Government base case utility is maintained in the counferfactual experiments With
balanced budgets the concepl of equal yield is unambiguous The new equUibrium prices and the
balanced budge condition will del ermine the minimal expendilure and taxes needed to maintain base
30
case publi~ utility Accordingly in general equal yield is inconsistent with equal nominal tax
revenue Some change in tax revenue Is necessary Different tax replacement schemes ate
considered to assure that enough tax revenue is collected This is the approach essentially followed
by most 01 the papers surveyed Only Feltenstein (1985) Goulder (198) and Jorgenson-Yun
(1984) chose not to follow an equal yield strategy
The question is of how to Interprete lila concept of equal yield when the government is alowed to
fUn deflcLts The optimal leve of expenditure for base ease public trliUty can now be taAinanced
bond financed or financed by a mix of bonds and taxation We have several versions of equal yield In
particular equal yield may now be consistent with equal tax revenUE Furthermore some meaSure
ltgtf financial crowding out effects of government deficns can be inferred from the comparislon of the
several equal yield alternatives These issues are extensively discussed ~Pereira (1987b)
44 Price expectations and the dynamic generalization of compensation
Indicators
The sum of equivalent and compensation variations over households is the most widely used
aggregated measure of efficiency gains or losses In the context of $teady~state eomparisions andor
when complete future markets exist andlor perfect foresighl is assumed there are no difficulties
associated with the use of the standard Hicksian indicators In fact correct future prices are known
in these cases
If expectations are not selfmiddot fulfilling andlor future markels are not available a dynamic
generalization is necessary 8allardmiddotGoulder (1985) provide some steps in that direction by
defining an indicator that accomodates periods far into the future when households do not have
perfect foresight but a steady~state prevails On he other hand this issue is more fundamental in
i
[
I
i I I I I [
I I
bullI
31
the contex~ of a Gner1 lemporary equilibrium ftamawor~ In such Circumstances Pereira
(1986a) dev~lops a cnamlc generalization which is obtained as the present discounted value of a i i
sequence of short-fUn optimal expenditure functions consistent with a base case expected future I stream of IlWities I
32
5 Empirical evidenee from selected policy Issues
Dynamic tax models have generated several important results which escaped Sialic modellers
The following is a seleCpoundd set of issuos whIch stress the marginal benefits of dynam~ modelHng of
economic behavior in innovltlttive areas of analysis The discussion of a consumption tax emphasizes
the benefits of dynamic tusehold behavior and intergenerational aspects The study of the
flliminatlon or the re~intToduction of the investsnemt tax credits is made meaningful by the dynamic
modelling of production betlavor Modelling governmenl in a dynamic conlexts allows the modeller
to study the irrrpact of government deflcits Finally a dynamic framework lets us appreCiate the I
relevance of consumer expectations in terms of the evaluation of policy alternatives
i
i
51 Consumption Tax
t
By Its very nature a consumption tax can not be adequately investigated with a static model bull
Fullerion-Shoven-Whaicy (1983) use Ihe model of the US economy s described in
Ballard-Fullerton-Shoven-Whalley (1935) to evaluate the movement from the currenl US tax Isys1em to a progressive consumption tax Since their modeJ Incorporates a Jaborleisure choice
where leisure is an untaxed commodlty their results reflect the fact that both the consumption tax
and Ihe present system are dslortionary
Concentrating on the rntertemporaf distortions they show that sheltering more saving from lhe
current US income tax could improve econornic efficiency even if marginal tax rate increases are
necessary in order to maintain government fevenue~ At first you have a revenue shortfall
However the economy moves to a higher sustatned growth path and uitimatefy lower tax rates
cangenerate the same revenue path Also wages increase in the long run with this policy The
bull
bullbull ~
33 shyt
i
I
I
present value of welfare gi[lS for a potiey of complete savin9 deduction wilh marginal tate
adjustments (consumption lax) is simulated to be around $500 billion to $600 billion 01 1973
dollars
They also investiga1e th 1ngth of ttme it takes th~ economy to adjust to these policy changes
Roughly they estimate the long run to b thirty years although this figure Is very sensitive to the
specification of the savings elasticity - a crucial parameler in this model
52 Investment T~x Credit
Goulder-$ummers (1 S87) address the impact of eliminating the lnvestmenl Tax Cradit (ITC) on
intersectoral capital formation and on economic growth Their model isYa1ticularly adequate to
addies$ this issu~ in that is postulates a forward looking investment m9we with adjustment costs
They show that the eliminating the ITC causes a reduction in the rale of investment Inveslment is
estimated to fall by about 7 in the short-run and by about 12 in Ihe new long run steady state
In lurn a previous policy anncuncement lowers Ihe overall al1ractiveness of Investment and leads 10
a downward shift of the Investrrent profile On the other hand the combined effect of a revenue
neutral simultaneous efimiraron of the ITC and reduction of the corporate tax rates is a long run
reduction ef the capital Sieck by 35 This pattern suggests that a revenue neutral increament in
both Ihe corporale tax rates and the inveSlment lax credits would be preferable 10 the formula
adopted in Ihe US Tax Reform Act of 1986
Pereira (1987d) addresses the impact of the ITe from Ihe oland point of Ihe curTenllax syslem
under the Tax Reform Act of 1986 Given Ihe concern over the potential depressiVe effects upon
savings and investment of the new tax system in conjunction with deficit reduC1ion mechanisms of
the GrammmiddotRudman type a pertinent question is should Ih lTC be e-introduced In the context
bull
34
-of his mod~1 of the US economy Pereira focuses on the trade-off between the distortion in the
intersectoral allocation of capital induced by the lTC the lower tax revenues it generates and
potential financial crowding-out effects on one hand and the positive effects of lowering the relative
price of new capital goods on the other hand Preliminary results confirm the qualitative results in
Goulder-Summers (1987) suggestting a welfare gain of about 2 of the present value of GNP in
the best scenario of absence of replacement taxes Furthermore preliminary results show an
important time pattern to welfare gains with the average benefits increasing the further you look
into the future This pattern is due to the presence of constraints to intersectoral mobility of capital
and inertia in the adjustment towards the optimal capital levels as reflected by the presence of
adjustment costs
I
I
bullAuerbach-Katlikaff (1gB) in the context of their intergenerational model of the US economy
consider the impact on savings capital formation and interest rates of deficit policies and balanced
budget increases in government consumption They conclude that deficit finance and government
consumption can significa-ntly crowd out capital formation and lower the welfare of future
generations However crowding out from deficit finance is a very slow process because it results
from increased government spending over potentially long horizons Also deficit policies may
substantially influence the longmiddotrun interest rates while leaving the short-run rates essentially
unaffected Finally and in opposition to the central role adjustment costs seem to play in both
53 Financial crowding outeff~cts of government deficits
Goulder-Summers and Pereiras models Auerbach-Kotlikoff suggest that the time path of interest
rates induced by a policy of deficit financing seems to be insensitive to the presence of adjustment
costs
bull
35
54 The role 01 consumer expectatlons
The expectations analysis has uncovered one of the benefits of dynamic modelling
Ballard-Goulder (1985) find that the appeal of adopting a consumption tax depends on the level of
foresight possessed by the consumers Furthermore additional foresight may be welfare worsening
This is a second best type of result This tends 10 occur under policies that lead to capital deepening
and declining rate of return to capital over time To the extent that consumers have more foresight
they will be beller equippod to anticipate the fall in the rental price of capital Consequently if a
saving incentive is enacted people save more with myopia than with perfect foresight Given the
exi~tence of taxes on capita and the discrepancy between the private an~--sccial returns to capital
the greater savings with myopia is beller socially Ballard-Goulder find that the welfare gains
from a consumption tax are reduced by about 10 when we move from myopia to farleet foresight
Therefore the attractivenness of adopting a consumption tax seems fairly [obust across these
foresight specifications
I
bull
36
6 Tbe power and weaknesses 01 dynamic modelling the example 01 corporale
tax Integrallon
The Issue of corporate tax integration allows us to sh~w the stre-ngths and weaknesses or the
dynamic approach Empirical evidence using CGE techniques indicates that depending on the precise
scheme of integration and the 1ax replacement methods integratiOn may have substantial effects In
Ihe work of FuliertonmiddotKingmiddotShoven-Whalley (1980 1981) total integration was found to yield an bull
annual static eHielency gain of $4 to $8 billion in 1973 dollars Simulated dynamic gains may be as
large as $695 billion or abcu 14 of the present value of tulUre consumption and leisure in the
US economy These gains result primarily from interindustry reallocalions of Investment and an
improved inter~emporal allocation of consumption
Mora recent work emphasizes hOw consumers asset portfolio decisions and firms finanCial
decisions affect the efficiency gains from integration Slemrod (1980) focusing on consumers
asset ponfoio deCisions finds static efficiency gains which are about twice -as Iarge as those
repOrted iiy Fulierton-King-ShovenWhalley (1961) FuUrlon-Gurdon (1983) focus on the
firms financial decisions They roporl efficiency gains of 6 of GNP from the elimination of the
tax distor1ions favoring debt However when they eliminate the corporate tax and repiace it with
increaSed persona income taxes addlHonal dls10t1ions are created in the optimal labot~leisure
decisions These distortions tend to dominate the analysis slJ(h that the overall effects of complete
integration are very modest Galpr-LuckmiddotTodr (1986) address simultaneously consumers
asset pOrlfollo decisions and the firms financial decisions They also find very modest efficiency
gains from integralion
The above results are important but may be severely biased There are severa aspects of
economic behavior and modelling crucial for the study of Income tax Integration whiCh have no been
i 37
captured innny of the above papers One aspect is the absence of government deficitsmiddot a balanced
budget is assumed It is true Ihal resource crowding OUI is caplured In these mOdels but financial
crowding out induced by government spending is nOI Second investment is not derived from
optimtztlcn behavior Investment behavior passively accomodates endogenous saving decisions As
a consequence the differential impacts of policies in the incentives to save and to invest are not
captured Aso full capital mobility across secors is assumed with instantaneous capital
adjustmenlS towards optimal levels This assumption rules qut different costs of capital across
seclors and therefore differentiated reactions to tax policies changes Finallyhe modelling of both
gov9rnmerlt deficits and of endogenous real and financial investment decisions necessitates the
~nsideraiion of a dynamilt framework and the introduction of financiaJ assets govemment bonds and
iprivate financial assets A dynamic framework also highlights the efficiency effects of Integration on r
ithe ptimal intertemporal decisions The above models are either sIalic (Slemrods and
GalpermiddotltJckemiddotTode(sj or a dynamic sequ~nce of otherwise Slatic model bullbull I Pereir a (1986b) develops a dynamic applied general equilibrium model bull to study the tbull
efficiency effects of integration on the growth and allocation of investmenl across sectors Special
Iattention is paid to the structural affeelS of resource and financial crowding oul induced by
gove mect spending and deficits and to the real and financial Investment reactions across industries
to both tax integration and flnancial crowding out His mode departs from previous work on income
lax Integraticn and for that matler from most 01 the CGIE literalure for tax policy evaluation in
several direcllons It encompasses an endogenous sequential equilibrium struelure founded on
dynamic behavior with flexible expectations Government deficits are optlmaUy determined
Investment decisions are tward looking and the resull of optimizing behavior Several financial
assetsmiddot public and privatemiddot are considered
Simulation resulls Indicale that the welfare gains from Integration under large deficits are at
32
best lJlodest when measured in terms of the GNP The elimination of the corporate tax and its I replacemenl by increased income lax rates yields long run benefits which are never larger than
2 of the presenl yalue ct telr consumption and leisure onder the previous tax regime and 1
under the current tax regime However the short run w~lare effects of full integration tend 10 be
very low and in in some scenarios even negative This is a flew intertemporar pattern of efficiency Ieffects which reflects an adjJsiment lag in the interindustry investment decisions due to the
existence of costs of adjustment On the of her hand partial integration achieved by excluding I dividends from the corporale lax base syslematically yields negalive efficlencyeffeclS In Ihe long Inm these can be as high as 3 of present value of the future value of consumption and leisure This
is a new second best effect suggesllng Ihat less than complete integration may have perverse I efficiency effects
The dynamic slructure of PerclrBs model snowing for an improved freatment of real investment
declsions and government deficts provides a potential setting for a more accurate measure of the
costs and benefits of inlegration The simulation resullS suggest lower benefits and an intertempora Ibull
pattern of growing benefits
Simulation results also clearly suggest robust negative effects from partial integration How
reliable is this result If dJvidends were deductible from the corporate lax base (as are interest
payments) the preferenlial Ireatmant of debl over equily would be ellminaled Corporallons should
be expecled 10 react by decreasing the optimal deWeqully rallo However corporale financial
gecisions are exogenously set In Pereiras model as in aft of the other models surveyed that go as far
as dealing with the issue
Also withdrawing dividends from Ihe corporate tax base is a way of eliminating Ihe double
taxation of dividends at belh Ihe corporate and personnal Income levels How will the optimal
aliocellon of household saving accomodate that change in Ihe relalive return to corporale equity
39 I
t However hOusehold portfolio decisions are exogenouly set in Pereiras model as in all of the other
models surveyeltj go as far as dealing with the issue
IIt is sate to say that the evaluation 01 the benefits 01 partial Integration as defined above are
sevrey underestima1eO Meanwhile the distortionSoenerated by the increase in the marginal
personal tax rates designed to ralse the revenue foregone by the dividend exclusion are fully
aCCltlunted for A good measure of the costs of integration together With a poor evaluation of the
benefils may very well be the reason why partial integration is Simulated 10 yield neg alive
eHikiency gains
On a different vein as most of Ihe model surveyed here Pereiras is a closed economy model It
is legitimate to wonder how the resullS would change If international capital flows were anowed
IThis almost certainly would affect financial crowding out since a good d~al of the capital inflows
could be used to finance public debt
For these various reasons the results should be treated as preliminary but they strongly I suggosl that the dynamic structure provides valuable new insights into the corporate tax integration SSU2
40
7 Concluding remarks I I I i
What has been acccmplished Ihis far The success of Ihe CGE research in laelding the challenge
of dynamic modelling can not be denied
I iGreat progress in the modelling of household behavior has been achieved Promising
developm~nts in the modeUing of production and government behavior have also been accomplished
Interesting innovations in the concept of equilibrium and computation techniques were discussed
The policy analysis was greatly enriched by considering transitional effects together with longer
run steadymiddotstate equilibrium paths generalized equal yield strategies accomodating different
financial crowding out impacts and generalized dynamic policy evaluation indicators
An important set of issues have been addresed by the models surveyed Traditional issues
focusing on the effects of capital taxation and consumption taxes have been pursued In terms of
capital taxation for example the intertemporal nature of the issue was enhanced by allowing an
optimal evolution of capital stock in the economy and forward looking optimal investment decisions
In Andersson (1987) GouldermiddotSummers (1987) and Pereira (1986b 1987d) this is coupled
with a improved treatment of several tax provisions like investment tax credits and depreciation
allowances In terms of the consumption taxes developments in the specification of Intertemporal
household behavior the introduction of overlapping generations and the modelling of
intergenerational links via bequest motives as well as a closer attention to demographic evolution
provided a much improved economic setting for the understanding of the several aspects of the
problem
Furthermore dynamic modelling has permitted the addressing of a set of new issues Policy
Issues ranging from dynamic tax policy analysis to the evaluation of exogenous changes in
government expenditures to the impact of different methods of financing government expenditure to
4 1
the importance of tinanciai croHrii1g out to 1he relevance of consumorS expectations in tne policy
results have been addressed
Despile all of the progress and in part due to such progress several avenues are wide open for
much neded additional research The following seem to be the most Interesting and promising
areasmiddot First the specincation of liquidity constraints to household behavior may help to obtain a
better understanding of policy chenges that affect directly the interest rates in the economy Second
major attention needs to be paid to the incorporafion of a roraign sector (with open capital markets)
Third some efforts are needed in terms of finding adequate computation techniques now that
dynamic modelling eally makes the curse of dimensionality worse than eVOrbull
The modelling of financial markets is the single most unsatisfactory speet of the dynamic
models surveyed here~ The problems associated with provlding a mo~ adequate treatment of
financial markets and in general endogenous and optimal financial deci$ions bull both at the corporate
and at the personal levelsmiddot have 10 be addressed That necessarily requires Introducing uncertainty
into the models To b adequate this must go well beyond the mere assumption of point expeetalions
with the whole array of modelling implementalion and evaluation problems thai generates
bull I i
42
REFERENCES
Abel A and O Blanchard 1983 An Intertemporal Model of Saving and Interest Econometrica
51 pp 675-92
~ Andersson K 1987 Tar-ation of Capital in Sweden - A General Equilibrium Model Ph D
Dissenation UnIversity of Lund Sweder
Auerbach A L KotlikoH 1983 National Savings Economic Welfare and the Structure of
Taxation in M Feldstein ed Behavioral Simulation Methods in Tax Policy- Analysis Chicago
University of Chicago Press
Auerbach A l Kotlikoff 1984 Social Security and the Economics of Demographic-
Transition in H Aaron and G Burtless eds Retirement and Economic Behavior Washington DC
Brookings Institution
Auerbach A L Kotlikoff 1987 Dynamic Fiscal Policy Cambridge Cambridge University
Press
Auerbach A L Kotlikof and J Skinner 1983 The Efficiency Gains from Dynamic Tax
Reform Internatiooaj Economic Reyiew 24 pp 81middot100
Ballard C 1983 Evaluation of the Consumption Tax with Dynamic General Equilibrium
Models Ph D Dissertation Stanford University
Ballard C 1987 Tax Policy and Consumer Foresight A General Equilibrium Simulation
Study Economic Inquiry 25 pp 267-284
Ballard C and L Goulder 1985 Consumption Taxes ForeSight and Welfare a Computatonal
General Equilibrium Analysis in J Piggott and J Whalley eds New Deyelooments in Apolied
General Equilibrium Analysis Cambridge Cambridge University Press
Ballard C D Fullerton J Shoven and J Whalley 1985 A General Equilibrium Model for
t 43
Tax Eoliey Eyalua~iQO NBER University or Chicago Press
Borges A (1987) bull A Survey of Compula1ional General Equilibnum Models CECD Economic
Slud~ 8
Bovenbig L 19B3 imperfect Capital Mobility in a Perfect Foresight Herberger Model
Mimea UC Br~eley
Bovenberg L 1984 Capitel Accumulation and Capital Immobility Qmiddottheory In a Dynamic General Equilibrium Framework PhD dissertation University of California at Berkeley t
Bovenberg L 19850 Dynamic General Equilibrium Tax Models with Adjustment Costs in A iManne ed bull 1985 EtQnomrc ECUjilbrium Model Formulation and Solution Mathematical
lProgramming Siudy 23 Arnsterdam NorthmiddotHoliand
Bovenberg L t 1986 Capital Income Taxation in Growing Open Eeondmies Journa of Pubtic
Eamp(omlcs 3t pp 347middot376 IBovenberg L ald W Keller 1981 Dynamics and Nonlinearlties in Applied General
Equilibrium Models Internal Report Mlmiddot81middot208 Department for Statistical Methods CBS bull
Vooiburg The Netherlands
Charnley C 1981 The Welfare Cost of C~pltal income Taxation In a Growing Economy
IJournal of Political ECQnQm~ 89 3 pp 468middot496 I
Charnley C 19810 Efficient Tax Reform in a Dynamic Model of General Equilibrium I Mimeo Yale University I Charnley C 1982 On the Optimal Taxation In Stylized Models 01 Inlartemporal General
Equilibrium Mimeo Yale University I Decaluw B and A Manens 1985 Pays en Oeveloppement at Modeles Calculables OEquilibre
General - Une Revue de al litterature empirique Monographie Sene 9 Volume 3 Centre de
Recherche et Developpemont Economlque Unrslt de Montreal
44
bull
I I
I
I
1
Erlich $ V Ginsburgh and L van der Heyden 1987 Wher~ do Real Wage Policies Lead
Belgium bull A General Equilibrium Analysis European Economic Reyiew (forthcoming)
air R and J Taylor 1983 Solution and Maximum Likelihood Estimation of Dynamic
Nonlinear Ratioant Expeclations Models Ecooometrica 51 pp 1169middot1186
Feltenstein A 1984 Money and Bonds In a Disaggregated Ec~nonYin Searl 1-1 J Shoven
Eds Aoolied ~eoeral EqJiliDCiulD tioalYsis Cambridge University Press Cambridge
Feltenstein A 19B6 A Dynamic General Equllibrlu Analysis of Financial Crowding Out
Theory with an Application to Australia Journal of Public Economics 31 No1
Fullerlon D R Gorden 1983 A Reexamination of Tax Distorllons in General Equilibrium
Models in M Feldstein Ed Behayora Simulation Methods in Tax pOlicy Analysis Chicago
University of Chipgo Press
Fvllerlon D J Shaven and J Whalley 1983 Replacing the US Income Tax with a
Progressive Consumption Tax JoulOal of Public ECQnomics 20 pp3middot23
Fullerton D Y Henderson J Shaven 1984 A COmparision of Methodologies in Empirical
General Equilibrium Models of Taxation in Searl H J Shoven Eds ~pQlied Geoaral Jguilibrium
lIoalyss Cambridge University Press Cambridge
Fullerton 0 A King J Shaven and J Whaney 1981 Corporate Tax Integration in the United
Statesmiddot a General Equiiibdufi1 Approach American Economic Review 71 pp~ 677~691
Galper H R Lucke E Tader 1985 Taxation portlor Choice and the Allocation of Capital A
General Equilibrium Approach Mimeo
Gill p W Murray M Saunders M Wright 1986 Users Guide for NPSOL (Version 40) A
Fortran Package for Nonlinear Programming T echnicsl Report SOL 86middot2 Deparlment of
Operations Research Stanford University
Glnsbur~h V and L van der Heyden 1985 General Equilibrium with Wage Rigidities An
45
AppJication to Belgium in A Manne ed t Economic EQumprit1m~ Made Formulation and $oJioo
Mathematical Programming Study 23 Amsterdam NorthmiddotHolland ~
Goulet J 1968 1 Adjustrent CO$1S in the Theory o(the Firm Reyiew of EconomiC Studies 35 bull
pp47middot55
Goulder L 1985 Tax-Financed versus BondmiddotFinancedGhanges in Governent Spending
bull Mimeo Harva(d UniversilYmiddot
Goulder l J Shaven and J Whalley 1983 Domestic Tax Policy and the Foreign Sector The
importance of altemative foreign policy formulations to results from a general equilibrium tax
analysis model In ~n methods in Ibe antico of income from caoia1 ed Martin
Feldstein Chicago University of Chicago riSS
Goulder l L Summers 1987 Tax Policy Asset Prices and 3rowtly ~
A General Equilibrium Analysis NBER Working Pper No 212B
Grandmont J 1982 Temporary General Equilibrium Theory Ch 20 in Handbook 01
Mathemalical Eccnornics Vol II NorthmiddotHoliand
Harberger A 1959 The Corporate Income Tax An empirical Appraisal In lax Revision
Comoendjum Voll House Cornmitee on Ways and Means Washington DC Govemment Printing
Office
Harberger A 1962 The Incidence of the Corporate Income Tax JOULn1 of Political
Economy 70 pp 215middot240
Harberger A 1964 Taxation Resource Aliocalrlon and Welfare fn lb Role of Dir bullbulll and
Indjresl Taxes io tbe Fede[al Reserve s~I~Ill Princeton Princeton Unlveresity Press
James J 1985 New Developments in the Applications of General Equilibrium models to
Economic History in Piggott aand Whalley Eds
Jorgenson D and K Yun 1984 Tax Policy and Capital Allocation Harvard Inslitute of
bull
I
I f
Economic Research WP 1107
Judd K 1965 ShonmiddotRun Analysis of Fiscal Policy in a Simple Perfect Foresight Model
Journal of poJiHcal 5CCJiW1l pp 298middot319 r
Upton D J Potbc J Sachsa nd L Summers 1982 Multiple Shooting in Rational
Expectations Models fuoometrlc 50 1329middot1333
Manne A ed 1985 fconQrnjc Eoullibrium Model FOrIDujaljQo and Solution Mathematical
Pogrammlng Study 23 Amsterdam NorthmiddotHolland
Manne A 1985 Or the Formulation and Solution 01 Economic Equilibrium Models In A
Manne ed ECQOQlli-Jr~ibmiddotum ~tQdel EQlJulailon and Solution Malhematical Programming
Study 23 Amsterdam NorthmiddotHolland
Mansur A and J Whalley 1984 Numerical Specification of Applied General Equilibrium
Models in Scarf H J Shoven Eds Aoplied General Eouillbrium AnaiysectIsect Cambridge Universily
Press Cambridge
Merrill 0 1972 Applications and Extensions of an Algorithm thaI Computes FixedmiddotPoints of
Cenain Upper Semimiddotcominuou Point to Set Mappings PhD Disserlation Unlvel$ity 01 Michigan
Negishi T 1960 Welfar Economics and Exislence I an Equnlbrlum for a Competitive
Economy MelroeCOOQ11ka 12 92middot97
Pereira A 1985 On the Existence and Uniqueness of Optimal Ou1put Path lor CRTS Firms in
Adjustment Costs Technologies Mimeo Stanford University
Pereira A 1985a Consumer Expectations and Ihe COSI 01 UllUty In an Intertemporat
Framework Mlmeo Stanford University
Pereira A 1985b Corporale Tax Integrallon and Government Deficits A Dynamic General
Equilibrium Anaiysls Mlmeo Stanford University
Pereir A 1987 On Ihe Computation of the Users Costs of Stocks Ecooomi (forthcoming)
47
Pereira A bull 1987b Equal Yield Alternatives and Government Oeficlts Mimeof Stanford
University
Pereira A 1987c A Dynamic Applied General ~uHlbrium Model for Tax Policy Evaluation
Comlletamp Docurnent2tion Mimeo University of California San Diego (in progress)
Pereira A 1987d Should the investment Tax Credit be Rei(ltroduced Mlmeo University
of California San Diego (ir progreSs)
Piggott J and J Whalley eds 1985 tJew QeyeQpments in Applied General EQuilibrium
bullenalysjs Cambridge Cambridge University Press
Preckel Pbull 1985 Alternative algorithms for computing economic equilibria In A Manne Ed
Radnor R 19amp3 Equllibrium under Uncertainty In K Arrow M Intrilligaor ods
I ibull
Robinson S 1 g86 Mullisectorai Models of Oeveloping CountrieS a Survey Oivision of
Agricullure and Resources UC Borkeley WP 401
Scarf H 1967 The approximation of fixed points of a continuous mapping ~MM Jouwal of
epplid Mathornallcs 15 pp 1328middot43
Scarf H with T Hansen 1973 00 the Computation of EconomiC Eguilibria New Haven Yale
University Press
Scarf H J Shoven Eds 1984 61l1llied Genera Elujlibrium 6nalyssect Cambridge University
~ Press Cambridge
Shoven J 1976 The Incidence and Efficiency Effects of Taxe on Income from Capital
Jurnal of ramie1 Economy 84 pp 1261-83
Shoven J 1983 Applied General Equilibrium Tax Modemng IMF Slaff PaoerS 30
Shoven J and LB Simon 1986 Share Repurchases and Acquisitions An Analysis of which
Firms Participate NBER Working Paper No 2243
bull
48
IShaven J and J Whalley 1972 ft General Equmbrium Calculation of lhe Effects of
Differential Taxation of Incoma from Capital in the US Jauroe of public Econgmics 1
pp281middot321 (
Shoven J bull and J Whalley 1973 General Equilibrium with Tax$s a computation procedure
and an existence pr00f ReviEhi of Economic Studies 40 pp 475middot490
Shoven J J Whalley 1977 Equal Yield Tax Alternatives A General Equilibrium
Computa~ui0nal Tfchnique J~HJrnal of PUblic EconomIcs 8 pp 211~224 bull
Shaven J J Whalley 1984 Applied GeneralmiddotEquilibrlum MOdels of Taxalion and
International Tr8de An Introduction and Survey Journal of Economic literature 23t pp
1007middot1051
Slemrod J 1980 A General Equilibrium Model of Capital Income Taxation Ph D bullDissertation Harvard Univnrsity
Slemrod J 1983 A General Equilibrium Model of Taxation with Endogenous Financial
Behavior in M Feldstein edbull Behavioral Simulation Methods In Tax Policy Analysis Chicago
UniVerSity of Chicago Press
Stone J 1985 Sequential Optimization and Complementarily Techniques for computing
Economic equilibrium in A Manne Ed
Summers L 1981 Taxation and Corporate Investment a q-Theory Approach BroQkinas
Summers L 1985 Taxation and the Size and ComposHion 01 the Capital Stock An Asset Price
Approach Mlmeo Harvard University
Whalley J 1975 A Geneal Equiloibrium Assessment of the 1973 United Kingdom tax
reform Economica 42 pp 139middot61
1JLTIllOS JIQll~)ltG PAPERS PUBLICADOS bull
I
I I
j
112 70 - LUCENA iJiCIIJo ~ IIEnvironrilent Honitoring and Orqanization Structure T (Setellibro 1987)
nI 71 - L~ENA Diogo A Note on the Representation of Information Stxueture (SeteJObro 1987)
nQ 72 - LUCENA D1CgO Envuorment 111)nitoring and Organization
Structure II (Setelilhro 1987)
ril 73 - AlmmE3 Antonio Pais e CAsPAR Yitor bull Tributacao Inoentivos e InvesUlllentos Analise Qualitativa noYellibro 19S7)
n2 74 - liAlltOSA rose Pedro Opt1lllal lIage R1gidity A SUggestedtlthodology to Teet tb Theory and an Application (Outubro 1987)
nI 75 - CABRAL L 11 B Three lIotes on SYJIIlletric Gemes witt itsYJIIlletric Equilibri (NoveJObro 1987)
n2 76 - LUCENlt D10gOTo Search or Not to Search (Fevereiro1967)
nQ 17 - SA Jorge asconcelos euroI A Theory of SynergyB (Fevereiro1987)
n2 76 - VILARES tanuei Jose Os Be IntenaMios IlIIportados ColIIO Factor de Produao (Julho 1987)
nQ 79 - SA J~rge VasCQncel03 e HoY to Copete and Co~unicate in tature Industrial Products_ (aio 1988)
n2 80 - ROIl Rafael Learning and Capacity Expansion in a New larket Under Uncertainty (Fevereiro 1988)
n2 81 PEREIRA Alfredo llarvilo Survey of Dynamic Computational General EquilibriUlll Hodel for Tax Policy Evaluation (Outubro 1987)
Qualquer in ormacao sabre o 1IIorlltlng Papers ja publlcados sera prestada pelo Secretariado de Apoio aos Docentes podendo os ~e3OSf
ser adquirido5 ne Secyo de Vlndas da Faculdade de EconollLia UNL na Travessa Esteviio Pinto Cltpolide - 1000 LISllOA
bull 4 I
The kei features cf these models afe summarized in Table 1 which provides a structure for the
discussion Most of toe models Ihal we examine could be said to be In the Yale tradition That Is
they wer developped oy sudnts oi Herbert Scarf or students of the students of Herber1 Scar This
shyincludes BalladmiddotFuerM-Shoven-Whalley Andersson Ballard-Goulder ~)Venberg FellenSlein I
ibull Go-cider GouldermiddotSJmmers ard Pereira The exceptions from this heritage are Auerbachshybull
Kotlikoff Erllch-GinsburghHeyden and JorgensonmiddotYun I This survey is organized ltis follows rne second section provides a brief overview of I
Inon-dynamic CGE models The third seolion offers a discussion of modelling economic behavior and
equilibrium The fourth section focuses on model implemenlalion and policy ev~uatlon The fifth
Isection summarrzes empirical evidence en selected policy issues The sixth section illustrates the
power and wea~nesses of dynamic modelling with the issue of corporate laxintegration Finafly the last section summarizes the state~of-the-art in terms of dynamic modelling for lax policy
evalutlon and sugg-ess areas of future research
_
--
--- -~ ----- __-shy
_ _ - _- - Con~um9rs_ 8amphavlor
Bajla rd -Fulflonmiddot -Shaven-Whalley
1985 Anderaaon
1087
Auerba ch~ Ko1Uko 1983 1987
BaHElfdmiddotGouldor 1983 1905
Sovanborg 1985 198e
ErHchmiddotGlnaburghmiddot middotHeydeR
1987 Feltensteln 1984 1966
Gouldar 1985
GouldorSummor 1987
Jotgeneon-Yun 1984
Porelr 1966 1991
Count dynamlc dl9agg~ objecth lastes constraint ondog endog OvorlapBnqvMI of datil
year _seltlng labor no savings gonorst mollves
inlertsmporal 12 Income otility eel IntertempOfll1 yes yo no noLS 1973 Ymiddotmiddot
_ groups max ---_
0 Intertnmporal rllcufsivs
Swaden 1984 yes aggrogate utility CES eq of molion 00 yos na no consumer max for weaHh
tntortorrporal Isoeiastlc~ 00-1983
ill yo 55 poundtUS u~illy middot19831 In tcemporal yesmiddot 11El7 yes yo 00
cohorts rn~x CES-19C7 plauslblo
bYPothlC1 inlsrtamporal
19n yes 55 ago uliJity Isoelastlc Intertemporal ybullbull ye ye ysLS BFSWexpanOOd _ cohorts max
0118 inlertemporal recursivo LS plausible ye aggregate ollllly lsoolastic eq ot motion no yes no no
hYPOlhotic consumer max for weallh jnterlemporlll Boigium 1960 1983 ro aggrogate utility St()ne~Gearyjner1emporal no yo no no
consumer max Inlerl(lmporal two periodmiddot two yes
Austrs1i 1991middot1982 yes 2 income ulllly CD yeluly ybullbull tn the flrsl no no 2-perlod) groups max constraint porlod
Intertamporal LS 1973 ybullbull 55 aga utility eel intarhtmporal ye ybullbull ybullbull y
BFSW bullbullponded cohorts max
intert9mpor~1 W 1973 ves aggregate ullily 180085110 intertemporal no ybullbull no no
BFSW bullbulllMded consumer max (I~gt IntertempQ(a~
W 1955-80 ybullbull aggregate ulilily trenslog intertemporal ybullbull yes no no consumer max
Inlerlemporal recursive 1973LS yes a income ulllily CmiddotD eq of motion yos ybullbull 00 no
BFSWxpanded groups max for wealth
middot5middot
bull-- _---- -- _-----shy~ -~
I AlJlt 1 - uynafllc (lIt MOOero fur all runey tVlIIVllllQlI
dynamic dluggI utllng
I Ba~ll-a rd-fullerton
-Shoven-Vlhalley I no 19 sectors 19S5
Andarsson 1987 ye 2 sectors
~ Auerbach~KotHkof no~1gB3
1983 1967 yesmiddot1g87 1 sector
Ballard~Go uld ar 1983 1985 no 19 sectors
Bovenborg 1965 1986 yes 2 sectors
Erllch~GIn5burg hshy-Heyden no 24 seclors
1957 1__ Feltenataln 1964 196e no 30 sectors
GoUldsr 1985 no 3 sectors
~ GoutdarmiddotSumrnamprs
1987 yes 5 sectors
_ Jorgen9ClIO-Yun ona sector
1984 no produces two goods
Poralr 19S~ 1987 yos 4 sectors
~bJecllve
Produc~rs
technology behavior InpIJ1 1-0
matrix adl costs opt Imal
capital stock
sialic cost min
CES in valu$ added
110
capital labor yes 00
np Invest adjusts
to sFvnns
max 01
interlempora flot rash flow
cnsl min-lOS) Imax of valul)
of the firm~1987
ill
CD-ISBa CES-1987
capital labor
captel labor
no
no
YA quadratic in
I nO-191)3J
qutldratic in Ilk-1gB
ye
YAS
static cost min
CES in vafl1oadded
10
capita labor yes no
no invest edjuss
10 savifQs max ot capita yas ~
value ot the CD labor 00 quadratic In yo Ilrm Ilk
static CD in several no global ltlost mln vaiue added
110 ltsp1al
Jabor Ves no Inves adjusts
to MVrHIS
statle oost min
____ slatio
cost min
max of value of the
firm equivalent 10 two~stage
~ost min me) of
CO in value ad~ed
110 CES In
value added 110
CESin shyvalue added
110 dual
trBRslog price function Cobb-Douglas
capUal labor
cepHal labor
capital labor
capital labor
capital
yes
yes
yes
no
no invost adjusts no to S8vinos net of
El-xog 9011 deflcitsl no
no invest adjusts to savings r
yes quadratic In Y
Ilk
no yo I
yes Intartamporal net cash flow
In V Added 110
labof Y quadratic In I
y bullbull
- 6 -
~ --- ~ ~- -- -~------ --- --- - --~--~ ---~ shy
_
- ---
--
--------
- - - --------
H n v I taxes
olllng allowed debt Ballardfullamprton~
dynamic objective constrfnt tupendlttire (IlleUs endogenous
aU major balanced trade yoarly ybullbull rly endogenous -f ~Shovon WtlitUy US taxes with C(in~Hlt00 pubHc uility balancod cQmposilon no no debt
eiasticllv junel l 1985 rnagt eCP-REV~d1 aU fn3jO[AndtH3Son vearly balanco small opan Gcon J Swodish inturn fa to cr raturn1987 no - inaia dlJn axogonolJs 00 Ini ITnininod poundgtPfUV
-I L)~ in~h~ --~ ~~ ~pound~~---1- ----shy
runrac h~KQllfkofi ai gulD
1963 1937 US m1l3syo - intertnmponl tixogancuamp yo ys cos~d ampConoroy bandmiddotfinanced flon-oppoundIla
B iii lIa fe-Gould 9r nil llltl1gtr balanced 1rndayearly yearly crcoge)Jus
US t)X6S wiih conslanl I
1963 1965 no public utility balanced cOlYlposWon~ no nD debt max budnt ExP~RCV alastlcHy lunct
Bovenberg simp~iriad 1985middotclosod econybullbull rl~ no - balanced exogenous no no debt US laxes 19a6~two counlry1985 1986 bodgel model
Ert leh~G tn aburghmiddot yeally simplified Inlert walt funet Ishorl~run ttads Heyden no balanced exogenous 00 no Belgium
taxes _ desloullbtium flltUensleln
1967 budaet EXP=REV 6-1mprifia-dmlm costs of yoarly yo rast of tho world
1984 1985 producing a ExTax Rev exogenous financed with yo Australian ~s a Ihlrd conSUmer public good money and bonds non-optimal taxes group
Goulder yaarly annual and endogenous aU maJor
laquo1()Sed economy1985 publlc utilily In1ertemporal composillon bull yes yes U~ taxesY rna versions IlEFElQgtfpoundV bond-financed (non-opllmal balanced trademiddotGoulder-Summer yeary yoarly endogonous bull all major
1987 00 public utility balMc composition no no US taxes with constant max budGel EXPREV alasticilY fund Jl)rgenaon~Yun yoarly aU major
no balanced exogenous no1964 US taxes closed economy budGet inteftemporal recurslve endogenousPereira all major
yo public utility eq of molion lovel and yo yos1995 1987 US laxes closed economy rna for debt composition bondmiddot financed oplloll
- 7 shy
bull - -__----- _ - ------ -- ------- _-----
---- ---
------
------
ll-ltiLt 1 uynamlc Lut MOOSS fOf I ax YOtlcy tv8luallon
Inanell ~ets and maf~9ts ojumLrtm- ----_ ----c_~
privata do btl GlvldtHlI pL~lc aliocalQflo rls shy
1rlc ypa expGct assets equity retention dobl 01 savlnq$ promlums path
Sal ra rd fu Barton phys cap1 used 10 buy soquence lt
Shavon-Wha1ley no financial exogenous oxogenous no physicBl 00 bull 01 sIalic 1E myopic 1985 assets capital flqrlil -
Andersson phys cap used 10 buy 1987 no financial no no yo physical no dynamic ~ perfect
assets capital Auerbach~KolUko If phys cap usod to buy
1983 1987 neo financial no no physical cap no dyoamlc ~ perfectYmiddotmiddotnSSf)ts and Qovt debt SailBrd-Goutder phys cap
1983 1985 no financ[al assets-shy -------
Bovenberg phys cap 1985 1986 no iinenclaf
asso1s Ertlch-Glnsburgh~ phys cap1
-Heyden no financlaJ 1987 assefs
FeltoRsteln phys cap 1984 1986 Invostment is
bondmiddotfinanced
used to buy
exogenous exogenous nObull physl~al no dynamic ~ perfllcl capital
used to buy
no no no physical capital
no dynamic ~onlinuos lima
PI perfect
used 10 buy no no no physical no dynamIc PF perfect
capital jllNt) porlod) bull used 10 buy
no no ye physical cap no dynamic PI p4)rfect and govt debt (Iwo Iods) bull
Goulder phys cap used to buy 1995 no financial exogenous exogenous yenbullbull physical cap no PFlfE bull IdynamIc perfoct~assels and am dabt slaUc
- -~ - shyGoulder-Summers phy$ cap
1987 bonds and no e)(oganOU$ no used 10 buy ye dynamic PF perfect _ bullbullultv bonds MullY (exoaenous
Jorgensonyun phys cap 1984 no financial
essels -~ - ------
Pereira phys cap 1986 1967 bonds end
equHy
________
no
8xogenoos
no
exogenous
no
ye
used to buy two type$ of physical
copltal used 10 buy
go dobl bonds equity
no
no
dynamIc
-
dynamlc
~
lE
perfect
fIbullbulllble (nonperlecB
middot6middot
~-- _middot~ ___~w_ - - ___ bullbullbull___ -------- shy~
-----
------
- ---
-
IlflpnUItll tit lun pVII-y PIlUAlIV
q6H1brla oqual fax polley ovnluaU-on efUclency derlvatton Ioorllhm
calibration param~hr computation COmelfed llold tndfcaHH eHecls
Bnl jard~FuU6rtQ n base year back soll1lion steady-sta1e yo inlsrgroup inerseel or Iii
VS transilion lump-sum with Interfemporai 1965
~Shovn Whallmiddotv replication wllh exog Morrills nmamo1ers alQorilhm _~ond~middotsta1$ ears Inc CVor EV
i~noerucn sleady-sfala no agg(ogaled lnlalsectora 1967
base year back sobHon SiMNON package replication wHn oxog valiant of VS Iransilion no Indicamptor lnlertamporal
paramelsrs multiple shoo ling 6~~~l~slat8
AU8rt)ach~Kotllkot steady-slato yo Inlerguna-r a lions
1893 1llH
quatitaUve by replication a$Sium~lkml Gauss-Soidel liS Uln1WOn any rule -lilh 1111ermpOrB
sonsililJ1y I _ ~poundI~Jjnnelill $t(lfdvmiddot~luiA J3djlt15S ~~ nnrilW ~--------
a ail r nimiddotGould or bas9 9tH bac~ solution Ballard~ staadymiddotslalo )05 kltHSclo(a~
1983 1985 VS Iransllinn lump~sum dynalTlk CV inlerlemporal steadY slafe parameters alaorlthm replicaUon+ with oxog Gourder
+stead~stato (ers loc inlareel10rat BCvonborg qunmaiive by dynamic version stoadY-$Igtto yes 1985 1985 YS transition some govt eN Inlortempora
S9rHltivilv OnearizaJian method ropllcaUon assumptionl of Johansons
+steady-sfata paoo palh Erllch-Glnsbu rgh~ actual vorsus no aootagalod inlersectoral2 bftse years back solu1ion variant of Negishis
Heyden replication wIth alltogbull optlmizaUon countorfaclUal no ~odjcator inlottempora 1987 paramlJters aocroach
faltonstaln baso yeat back solution actual versus no aggrampgalfld intersElclornl replication wilh exog Merrills counlerfaclutll no indlclltor middotlntamprlempOull1984 19B6
parameters al90rithm 1981middot1982
Gouldar staady~slala Inlar5ecioralbase year back souUon Variant or 1S lronsilion no dynamic CV inlortempora
stoady state parameters Newlon 1985 feplicalion+ with iUog Fair~Taylorl
+slaildy-sla1e Intargeoorat Gtlulder~Summers base year back solution varIant of stlJadymiddotstate fixed no aggregated fntersectoral
vs trarlsition government Indicator Intertempo1al steady state oatameters Newton
repticaUon+ with exog Fair-T aylQf 1967 +ste-adv-stata bullbullbullndlno financial
Jorgenon-Yun econometric steady-stale Money melfic In1ersecloral
- - 8slimalion vs traosUian no social weUare functton intertampotal1984 +~teadlmiddotstatl)
~ shy
qtJalitaUv8 by NPSCt optimal path ye generalized inlersectorai 1986 1987 Perelr
replication 8S$umptlonl optimizatIon V$ lumpmiddotsum dynamic lntertemporat se~~ttilitf ____Jllgorilhm opllmal pah pen inc CVoclV financial
- 9 shy
bull ~ ___ - _______ - ____ bull___ OM bullbull __ bullbull __~____ ~_
--
- - - --------
--------- - - - --------
---
1nltJ1 bull I ~ 1J11ltI1 v I ~
-
Sa Ita rd-Futlertonmiddot intargration consumption lax ~StHlven~Whall~y VAT capila~ or labor tax
Ch~mqflS1985 Andersson
consumpion tax neutral corporals lax system 1987
it ue rb a c h~ i(at II Sltotf COIlSUft1ptian lax (poundpird ItS hibor Mx~on ufocto of taxation on capital fOfrnJilti1
dvnamio fiscal policv effects af OQvernmanl defioit (3tlard-Goulder
1883 1987
allocts 01 the adoption of a 1ge3 1985 pure consumpHon lax
im~rtmce of cons fCreSClht -Bovenborg analytically orianiad approach
1985 1986 effects ot pure conumption tax capllal income taxation in Opo economies
Erllch-Glnburghshy-Heyden reat wage policios in Belgium
1987 Felter1steln 1984 - theorotical
1985 ~ linnncial crcwdlng out1984 1986
Goulder effects of tax- vs bond-financed 1985 chanQes In pubtlc expondllufe
~~--
Gouidr~SUmmampfI torpora1$ lax cvts reduced 1987 ITO increase 10 gasolioe taxes
Jorgenson-Yun
effects ot several programs of tax reform on the allocation of ~~ptai
1984
-- Purelra
corporate la)( Integration effects cf romiddotinlroducinQ imreslamnl tax credits
1986 1967
10 bull
bull -~- ~- --- _ - ---____--_ - -----_-_ --~ ---- ~
1 1
2 Brief overview of non-dynamic CGE models
The origin of most or the models we survey goes back to the work of Scarf (1967 1973) whlgt
first developed a reliablo algorithm to compute equilibrium price for a ArrowmiddotDebreu economy
HiS algorithm used simpicial Bubdivision techniques and can -be shown to be the computational analog
of the fixed point theorems previously used to prove the existence of equilibrium His technique
could soW a model with an arbitrary number of consumers and commodities as long as all agents
were price takers consumers were subject to budget constraints demands were continuous andmiddot
production did not display increasing returns to scale The alQorithm while guaranteed to converge I
I I I
was relatively slow for prcb1ems involving more than S1If 20 dimensions A maior improvement in
computational speed Woo offered by Merrills algorithm (1972) which used tpe same fundamental
ideagt as Scarfs procedure
Scarfs model (as is true for the standard Arrow-Debreu model) does not include a government
sector ~ neither taxes nor public goocs At one of the mosi promising appHciUions of the new
computational techniqu w~s in Iha area of tax policy evaluation Shoven and Whalley (1973)
exlended the general equilibrium model and computational approach to include a wide rry of taxes
and a governl11ampnt spending plan The original ShavenmiddotWhalley model was static (although the
different commodi~ies CQuid be considered simitar goods available at different datest as In
Arrow-Debrau) and government was aumad to run a balanced budget The data are arranged as a
social accounling matrix The models spacification and calibration is checked by solving it in the
presence of the base set of laxes The result should ba exactly tha initial social accounting matrix
After having pasd this replication check the model is solved for a counterlactua equilibrium In
the presence of a new tux design The result is once again a social accounting matrix The two
equilibria are compared in order to tSS($S the impact of the new tax plan The first uses ot this
bull
1 Z
model for lax policy evauajon were ShovenWhalley (1972) Whalley (1975) and Shoven
(1916)
Completely static models as Shaven-Whalleys are unsatisfactory for many tax reform ISsues
These include corporate lax Integration effects of investment tax credits effects of accelerated
depreciations consumer Or e)penditure tzxes~ importance of saving subsidies like IRAs etc These
are essentialy dynamiC ksues They involve not only the allocation of capital across sectQfsbut
perhaps more Important the capital intensity of the economy But the capital accumulation and
capital reallocation take time and may involve adjustment costs Because of these issues Ballardmiddot
Fulierton~Shoven~Whaloy took the first steps towards developing a dynamic model
In this conlexl the ShovenWhalley model has been extended and implemented for Ihe US
-economy In BuIIBrdmiddotFulienonmiddotShovenmiddotWhalley (1985) While Ihis book completely documents the
mo~el it was used in several publicalions beginning in 1978 Their mode) consists of nineteem
production sectors twelvo households and fifteen consumer goods It includes a very detailed set of
taxes including the federal and state personallncome taxes federal and state corporate income taxes
SOCial Security taxes pro1J0rty taxes unemployment Insurance excise taxes sales taxes etc It
has been calibl att-d to reproduce the 1973 US economy Recentty a version corresponding to the
1983 JS economy has been developed
wThe 8alard-Fullenon-Shoven Whalley model is dynamic In the sense that consumers face a
choice between current consumption and leisure versus future consumption (which can be
purchased via s wlngs) The consumer classes act as if they were maximizing a nested CES utility
function over the domain of a CES aggregate of contempo-raneous consumer goods and leisure and
future consumption subject to their income constraint The parameters of those funcHons
determine the shares of income devo1ed to each commodity to saving and to the purchase of leisure
They also determine the two key elasticities in the models ~ the elasticity of labor supply with
bull
13
respecl to tgte rbullbull1 aoer lax wage rae and the elasticity of saving with respecl 10 the real after lax- rate of retum to capital
In the model consumerS have myopic expectations regarding lulure prices and in particular
regarding the future rale of return 10 capilal Ballard (1983) and BallardmiddotGoulder (1985)
inco~ora1e both perfect foresight and limited foresight into this model Future consumption is
acquired by buying a fixed composition portfolio of real investments that offer an infinite annuity
of returns
The production side of Ihe model is compielely static The mod1 Incorporates a constantmiddot
olasicily 01 substitution betweeen prlmary inputs in production (capital and labor) and fixed I coefficients for intermediate inputs The model distinguIshes between Industrial outputs and
eonsumer goods for the simple reason Ihat tile data are classified differenUy Industrial sectol$ I I involve such categories as forestry and fisherle metal mino9 and publishing and printing while
consumers purchas-3 fUfHi1ure 2ulomoblles and books This fact is recognized in the model by Ibull
incorporatirg a secord stage Of production which converts industrial outputs intO consumer goods
This technology is uStally mod9-led as a flxed~coefficient conversion matrix
The BallardmiddotFINton-ShovenWhalley model assumeS that the private sector finance
marginai ioveslment with the same composition of debt and equity as currently exists in each sector
This is the same arsumrtion that Herberger originally used Investors all hold debt and equity in tile
same proportion Thampreiore this ownership can be aggregated into simply capital ownership For
tax purposes however the separate treatment of debt and equity is taken into account at both the
corporate and persofa Io=vel SimllarlYJ the dividend poIicies on corporale equitY are established
exogenously There are no government bonds in the model since there are no government deficits
The modol is solved for a sequence (as many as 100) of temporary equilibria wltll consumers
allocating income between present and future conumption at each point in time The palh for 1he
14
economy ~ a sel of connected equilibria The connectIon is pra~ided by capital accumulation
Capital accumulation is endogenous and determined by saving The model stans with a sociagt
accOunting mi1rix In the bzse case the economy is assumed to be in a steaqy state growth -path (along
which all relative prices ~fe constant) The madel solves for both the naw steady state growth path
and the transltion 10 it zf~er a policy intervention rhe authors have frequently addressed the
question-of howong rt tKes to effectively settle un10 a new steady S1ate growth path
The dynamics of the model are limited however in that future consumption is collapsed into a
composite commodity Aso the absence of government deficits and the lack ofproductlon dynamics
limits the realism of the BallardmiddotFuliertonmiddotShoven-Whatley model for the analysis of dynamic
policy issues (such as the adoption 01 a consumption tax or Ihe elimination of the investment tax
credit) The models we survey try to Improve upon Ihe Ballard-FlIiertonmiddotShaven-Whalley
dynamic modelling The models Jn~orporate more forward lookTng behavior They improve the
dynamics on the conampumrtkm side of the economy and some of them introduce true 1ynamlc behavior
on the part oI the producers We turn now to the issues inY9lv_ed in developing such dynamic models
15 I I ~3 ~ssues in the modelling Of dynamic behavior
We will 100k firt at the speclficalion of economic behavior of consume(s~ producers
goyemment and Ihe r~sl of Ihe world The modemog of corporale and household financial decisions
I is then addr5sed Finally Ihe eencept of equilibrium is discussed
I
i
I 1
31 Consumers behavior
Early efforts 10 build dynamic features into the economic behavior 01 consumers are due to
Ballard (1963) and Auerbach-Kotllkoff (1983 1984) Their work has been closely followed by
several subsequent authors In faet dynamic household behavior Is the single most pervasive aspect
amohg the models surveyed
All the models Inc-O1orale some loro of lila-cycle behavior Household behavior is dewmloed
by the maximization of an addWVely separable time invariant intertemporal utillty function The
utility lune1ion is defined oyer the domain of the consumption goods In the economy In most of the
models leisure is also an argument in utility so that labor supply is optimally determined
Utility maximi4ation is subject to a lifetime intertemporal budget constraint which equalizes
the presen1 value of consumers income and expenditure More reeentiy in Andersson (1987) bull
Bovenberg (1985) and Pereira (1986b 1987d) the constraint i$ defined as a sequence 01
recursive equations of motion on wealth This has the potential advantage 01 accomodating liquidity
constraints However It should be recognized that in the absence 01 liquidity constraints (ie when
consumers are free to completely borrow against future income) the two specificallons of the
househofd ccnstraint are essentiay equivatent Furthermore tn both versions saving is optimally
determined as a way of translaring wealth intertemporally
1 6
Some lodels now have a sophisticated dynamic specification of hoosehokl In Ballard (1983) -
BallardmiddotGoulder (1985) and Auerbach-Kotlikof (1983 1984 1987) consumers behavio is
embedded in an iterQEneraional setting Each of them has S5 age cohorts simulataneously alive
Som-e measure of jntergener1ional altruism is considered in the form of bequest motives In Ballard
(1983) and Ballard-Gouldr (1985) consumers derive utility directly from bequthing parl of
their wealth Therefore bequests assume 1he form of a scrap value of terminal wealth of a
generation In AuerbachmiddotKotlikof (1983 1984 1987) each generation empathyz bullbull with Ihe
utility of future generations ovar bequethd wealth In addition Ballard (1ge3) includes detailed
demographic poiections wfill the intent of incorporating in the policy analysis the effects of the
bull Post World War tI baby-boom
In the real world household decisions also include the optimat allo~atiOn of saving among
shyalternative physical or financial assets Theories of household portfolio behavior are relatively
Iwen~establshed However they involve uncertainty as a crucial element Accordingly they have
middotnot been incorpora~ed in the context of the determinIstic dynamic models we sutyey here We will 1 Come backt-O this Issue below
32 Producer behavior
The efforts to build dynamic features into the economic behavior of producers are more recellt
and less widely adopted The first attempts are due to Bovnberg (1984 1985) and Summers
(1965) Dynamic behavior has been more fully incorporated In the recent models of Andersson
(1987) Auerbach-Kotlikoff (19B7) GouldermiddotSummers (1987) and Pereira (1986b 1987d)
Part of the reason why production side dinamlcs has been more slowty adopted is the weak
supply of accepted theories referring to the dynamic behavior of the firms In the models referred
I 17
to above gynamlc producion and investment b~havior are middotinduced by the existence of capital
adjustment costs and linked to Tobins q theory Adjuslment costs are designed to capture both the
incomplete mobility of capital aoross industries and installation costs ~e the costs of adjusting
capital towards its cpHmai level)
Adjllstment costs can be conceived as internal to the firm and measured in terms of foregone
outpul along the lines of Lucas (1967) Alternatively adjustment costs can be viewed as actual
external cests incurred together with Ihe purchase costs along the Unes of Gould (1969) With the
exception of Pereira (198Gb 1937d) all of the CGE authors follow the former approach
The firms in the economy maximize their market value as the present discountad value of the
future stream of dividends In Andersson (1981) and Pereira (198Gb 1987d) firms are seen as
maximizing the present discounted value of net cash flow Maximization is constrained by Ihe
adjustmenl cosl lechnology and ~n equation of motion describing Ihe evolution of the capital stock
It should be noted that undereogenous divldendrelention rules the two problems are equivalent II
should also be nOled Ihat a salisfectory economic rational for th bullbullistoo of dMdonds with the
present tax code is missing in the profession (Shoven (1986))
The models with static formulation of producers behavior are characterized by passive
investment behavior Investment merely accomodates to saving in the eccnomy With a dynamic
formulation induced b adjustment costs real investmen decisions are forward looking Investmenl
Is endogenously and optimally determined by the firms A fundamental difference between the
shonmiddotrun in which capital stock is given and longmiddotrun in which Ihe level of capital is allowed to be
optimally determined is emphasized
This extra richness of produc1ion dynamics is not without costs A careful look at the summary
tables will clearly show an inverse relation between the adoptlon of dynamic features in production
and the level of disagreggation of the production side of the economy In fact with production
bull
j
1 6
dynamics $e dimension of the problems is immensly increased Let IS be more specific
In a static framework with constant returns to scale production technology the output level ismiddot
indeterminate The optimal aliocation of inputs can be obtained by cost minimization with any
feasible output level generatillg zero profits Such zero profit conditions are used to solve for the
output prices in terms of the factor prices and hencemiddot to reduce the dlmensionality of the problem
from the -number of commodities and factors to the number of factors Thus even if the model deals
with 30 production sectors the computation of an economic equilibrium can take place using only
the dimensionality of the primary inputs in the economy
Now under certain regularity conditions on the production and adjustment costs technologies
leading to enough concavity of the optimality objective the intertemporal output path for the firm is
endogenously optimally and uniquely determined even with constant relums to scale technologies
(see Pereira (1985 1987a) Accordin~y with adjustment costs in the model optimal profits
will in general be non-zero This result has crucial implications for the computatiQn of equilibrium
in the model~ with production dynamics With adjustment costs no reduction of dimensionality is
possible The curse of dimensionality returns as a binding constraint
The introduction of adjustment costs adds another significant complication to the model With
capital being less that perfectly mobile in the economy different rates of return on capital will
exist in different sectors This is a difficult problem to tackle conceptually in the absence of
uncertainty Also in the real world producer maximization choices include also its choice of
financial ratios (debVequity) and payout rates (dividendretained earnings) These financial
subjects are difficult and much Clf this behavior is still taken as exogenous by the modellers We
will come back to these issues below
middot i
19
bull 33 government behavior
middot
Two issues dominate the modelling of government behavIor First govemment behavior has
Ibeen typically seen in Ihe CGE liIerature for lax poliCY evaluation as constrained by yearly balanced
I budsets (see for example 8allard-Fullrton-S~ovn-Whafley 098Sraquo The analysis of
govemment defidlS and public debl in a CelE context requires a dynamic selling Second the level of
government expenditures is either exegenouty given as in Auerbach-KoUikofi (1984 1987)
6ovenbrg (1984 1985) Feltenstein (1984 1986) and Jorgenson-Yun (1984) or
endogenously (but no optimally) determined by the balanced budget conditions as in
6allard-Fullerton-Shoven-Whalley (1985) Andersson (1987) Erllch-Ginshurgh-Heyden
~
(1987) Gouder (19aS) and Gouder-Summers (1987) In the second case the composition of
bull public expenditures is often optimally determined However the leve~ of government expendi1ures
can only 00 endogenously and optimally determined if the government is seen as an optimizing agent
and is allowed to run deficits
The first attempts to deal wiU the government defiCits in CGE tax models are due to
Auerbach-Kollilltoff (1984) and Feensain (1984) In Auerbach-Kotlikof (1984) government
expenditures afe exogenous They grow at the rate of growth of population However given 1he tax
structure and tax revenues yearly deficits and surpluses are atTowed subject to an in1ertempond
constraint that the present value of future tax revenues equals the present value of future
expenditures
In Fellenstein (1984 1986) government expenditures are also exogenously given He also
allows the government to run deficits to finance expenditures in excess to 1ax revenues However
surpluses are returned to CQnsumars in the form cf transfers Accordingly government ts not
subject to any constraint regarding the future repayment of public debt
In GouJder (19851 the 9Vernmenl maximizes a static social welfare ~nctio subject to a-
bull
20 I i
I
I I I
I bull
balanced budget conSirtlL This follows the optimal alocatkm of government expenditures alpng
the lines of Baliard-FHeiton-Shoven-Whalley (1985) Ttle model Is generallzea to allow for
exogenous changes in HIe time path of government expenditures Two financing alternatives are
consjdeft~d and contr8s1ed EldditionaJ tax revenues and bond issuance
Pereira (198Gb 1137d) auempts to address both the incorporation of deficits and the
determination of govement expenditures The path of government expenditures and the path of
delicilSlsurpluses (ano therefore the path lor debt) are endogenously and optimally determined The
government is seen as mltJximizing n intertemporal social welfare function given the 13x structure
Optimization is subject to a sequence of recursive equations of motion reflecting the evolution of the
ptlbllc debt allowing fDr government budget imbalances It should be stressathat this specification
01 the government cOMtcalnt is equivalent to the specificaUon in Auerbaeh-Kotlikoff (1983 1987)
when no liquidity constraints affect -gove~nment behavior
The extra richness in the treatment of government behavior allows tfte examination of
government debt poices in a truly dynamic setting Also financial crowding out effects induced by
government deficits can be analyzed (see Auerbach-Kottlikoff (1987) Feltenstein (1986) and
Perer (198Gb 1987e))
34 Foreign sector
Most of the open economy models follow the assumptions of balanced trade with import and
export net demands characterized by constant elasticities along the line of Ballardmiddot
-Fullerton-Shaven-Whalley (1965) Such is the case of Salfard (1963) Ballard-Goulder
(19S5) and Goulder-Sllmmers (1987)_ In Feltenstin (1985) the rest of the world is treated as
21
an addjlion~ consumer group
Bovcoberg (1986) develops a model In which two economies lIr considered each following
intertempoml perlect foresight paths These economies meet in the international forum Their
trade relaroMhips are characterized by yearly balanced trade accounts
None of the new generation of dynamic cae tax models has yet IncorPorated the lnternational
capital flows as done in aoulder-Shoven-Whalley (1983) for the earlier Ballarrshy
-Fullertonmiddot Shaven-Whalley (1985) model This is unfortunate as there is an important
Intertmporal aspect to international lending The first al1empts along Ihese lines are due to
Andersson (1987) and Erlich-Glnsburgh-Heyden (1987) Andersson (1987) in his model of the
Swedish economy adopts a closed economy appraoch in which rates of return in the domesJic
economy are largely determined by the international capital markets Fidinterest rate induce
Intlnatlonal capital flows which determine and finance the International lrade imbalance In
lurn In ErlichGinsburghHeydenmiddot (1987) foreign trade Is generaled according to an
intertemporal Irade welfare lunction with constant import and export elasticities In the shortmiddotrun
they allow inlernational trade inbalances which generate capilal ftows to the domestic households In
the long run however trade balance is assumed
35 The need for financial markets
A dynamic economic Wucture not only provides the ideal environment 10 model many features of
economic behavior it also permits one to Incorporate the financial side of the economy If the
government is allowed to run deficits the question 01 deficll financing automatically follows If
investment is optimally dewmined and returns to capital are different aClOSS sectors problems of
investment financing arise If there are seeral final1cfal assets In the economymiddot government
4-
I i
ibull
I I I I
I
j
22
bonds privpte bonds and equity (or simply physical capital installed inmiddot different sectors) the
problem of allocation of saving among assets with potentially different returns arises
The papers surveyed var greatly with respect to l~e extent of their attention to the financial
side of the economy At Clne extreme are the models in ~anardmiddotFuliertonmiddotShoven-Whalley (1985)
Andersson (1987) Ballltld (1983) Ballard-Goulder (1985) Bovenberg (1985 1986) and
Erlich-Gjnsburgh-Heyden C19B which are devoted exclusively to the real side of the economy
In turn Auerbach-KotliKoff (1984 1987) Feltenstein (1984 1986) and Goulde (1985)
allow for government debt In these models saving finances changes in government debt and physical
capital Private and publlc 2ssets are perceived by the households as perfect substitutes The
allocation of saving merely adjusts to the relative demands for funds
Feltenstein (1984 1986) is the only model surveyed which introduces money Government deficits are financed by issuing money and bonds according to an exogenously given rule Money is
demanded by consumers for transaction motives and an exogenously given fraction as a store of value
On the other hand government bonds and physical capital are the vehicles for the intertemporal
transfer of wealth
Summers (1985) and GOlldermiddotSummers (1987) introduce a whole menu of financial assets shy
firm specific equity capital Different assets earn different rates of retum However such rates
are equal up to constant and exogenous sector speCific risk premia Therefore the introduction of
constant exogenous risk premia as helpful as it may be in the context of calibration does not solve
the main issue of the non-optimality of the allocation of saving Also talking about risk premia in a
deterministic context is somewhat unsatisfactory
Pereira (1986b 1987d) also introduces a whole menu of assets private and public bonds and
firm specific equity However all the assets are expected to yield the same rate of return and
therefore perc~jved as perfect substitutes The different asset types allow consideration of
23
exogenOU$~8tUequjty triO dividendlrc~enlion rufes and therefore seve-ral sources of investment
financing bond equity and retained earnings
The nonmiddotcptimalHy of the allocation of saving and the absence or exogenelty of corporate
financial rules is _a reftectlon of the limitations of the determInistic approach Either in a context of
perfect anticipation oi the future prices Of in general within the t$atm of point price
expectations all the papers follow a deterministic approach Under such circumstances consumers
either expect diHerent rales of return (inclusive of risk premium) across assets in which case
they wilt buy only one asset (that with highest rate) Of they expect equal rat~s of relurn in which
case Ihey are indifferent about the asset composition of Iheir portfoliO There is no way of
Iradingmiddotoff rales of return and risks to obtain an optimal interiof solution to the problem of the
allocation of saving
The most advanced contribution in the modelling of saving allocation in a CGE setting is due to
Slemrod (1980 1983) in the context of a static one-period model In his model consumers act
according to a two stage separable decision process They Hrsl decide on how much to save Then
they decide on the allocation of saving according to an Indirect uttlity function dependent on the rates
of ratum and variances offered by the different assets in the eoonomy The sourCe of riskiness in the
ampConory comes from an uncertain marginal product of capital On the other hand aside from
portfolio decisions the res of the economy is insulated from uncertainty
The most complete contribution in terms of the treatment of the corporate financial rules is
FulienonmiddotGordon (1983) In a variant 01 Ihe BailardmiddotFulienon-ShovenmiddotWhelley (1985) model
Fullerton and Gordon have capital intensity and optimal financial decisions Jointly deter~ined
through a two siage process The cosl of financing capital is minimized by trading off the tax
advantages of debt against the epected real bankrupcy coSls inherenl with high debVequity ratios
Given the optimal debtequity ratio the level of inveSlment Is chosen such that at Ihe margin the
24 l
bull
rerurn pn ~uity equas the return on bonds plus an exogenous risk premium
36 Market assumptions equilibrium and expectations
Virtually all of the papers arc characterized by Walrasian market clearing assumptions All
markets are perlect1y competi1ive Atomistic competition among agents 1s assumed even thoughmiddot only
a flnile number of agents is considered Virtually no mark~t disequilibria or price stickiness afe
considered The only exception is ErlichmiddotGinsburghmiddotHeyden (1987) bull In thei model of the Belgium
economy I the _wage rate is fixed in the short-run Therefore in the shOrlwrun desequilibrium in
the labor market will generate endogenous unemployment However in lh$ long-run all prices
Including the wage rata are Ilexible and accordingly ali markets ciear
Given the dynamic nature of behavlor in the economy markelmiddotcfearing prices in each period
depend on expectations of fture prices and on tax variables In the economy There are essentially
two ways at interpreting the economic equilibrium in such a dynamic conte~bull If future prices are
perfectly anticipated (io expectations are self-fulfilling) a perfect foresight equilibrium
prevaHs Then future actions are merely the implementation of current decisions for future
periods However jf ptica expecla110ns are not perfect (Le agents make mistakes w1th re$pect 10
future prices) then a temporary or shortrun equilibrium prevails Markets ciear and clearing
prices depend on fu1ure price expectations Current plans about the future are typically not
precisely implemented They will be revised as more or better information becomes available to the
economic agents See Grandmant (1982) for a comprehensive survey 01 the temporary equilibrium
literature
Wilh the exception of Gould (1985) and Pereira (19S6b 19S7d) all the models surveyed
adopt th~ concept of a perfect foresight equilibrium In turn BaliardmiddotGouldr (1965) considers a
J
2S
flexible am9unt of foresight In terms of the number of years over which price movements are ~
foreseen Pereiras model (1986b 1987b)) is flexible in a somewhat different way in that it can
indude any range of foresight from myopia to perfect foresight The choice between the perfct foresight and lemporary equilibrium is ultimately to be made on
philosophic grounds It can be argued that less than perfect expectations Imply that agents are
irralional in some way (see Auerbach-Kctiikoff (1987 p 10) However the reverse argument can
be made One can question wiether agents are reaHy rationar and perfectly knowledgable about
future prices
Recent evidene of Balird (1987) Ballard-Goulder (19851 Goulder (19851 and Pereira
(1986b 1987d) suggests that the choice in modemng expectations is an important one They show
that the degree of foresight inl0 thc future (ranging from perfect foresight to-myopic expectations)
may have dramatic impacts on the pollcy conclusions of Ihe model Accordingly the best research
strategy may b to design models which are flexible enough 10 allow for different rules regarding the
formation of expectations With the exception of the articles just mentioned sensitivity analysis
bullhave previously not been performed along this dimension
In terms of implementation the two concepts of equilibrium - perfect foresight and temporary
equilibrium have different implications The dimensionality of the equ1ibrium soll1ion algorIthm
is involved Suppose we have a model with ten markets to be run for a period of 50 years Aside
from normalization a perfect foresight model implies computing prices In 500 dimensions while a
-t~mporary equilibrium model requires sorving SO equilibria each in ten dimensions Given that
computational speed often varies with the cube of the number of dimensions the lemporary
equilibrium formulation is potentially strikingly more feasible However Ballard (1987) and
Goulder-Ballard (1985) have devoloped techniques to greatly speed the computation of a perfect
foresight equilibrium
26
j
I
4 Jrnplemenlalion and Issues on policy evaluallon
41 Calibration and equilibrium comparlslon
CGE models at typically parameterized by the use of a calibration procedure Some parameters
are exogenously given However some crucial parameters are determined in such a way that the
model replicates the data for a given base year See MansurmiddotWhalleY 11984) for an eXlensive
discussion of 1~i$ issue
CalibratIon in a dynamic context is generaliy interpreted as requiring two properties First
replication of base year data is required Second the model is parameterized to simulate an
intertemporal balanced growth path when Ihe base policy is maintaineq This Is the approach
followed by Ballard (1983) BallardmiddotGoulder (1985) Goulder (1985) and SummersmiddotGoulde
(1986) It follows the practice of BallardmiddotFuliertonmiddotShovenmiddotWhalley (1985) with their
sequential cquilibrium dynamics
Other authors Auerbach-Kotlikof (1983 1984 1987) Boyenberg (1984 1985 1986)
and Pereira (198Gb 1987d) follow whal we cali a qualitalive calibration The slructural
palametars are exogenously chosen so that lh economy follows a reasonable path into Ihe fulure
ThaI has 10 do wllh the fact thaI given the recursive nature 01 the dynamic economy and aside tom
such structural parameters only iniUa stock values are needed to run these models Given initial
conditions on the stocks of say private wealth capltal and government debt agents will optimize
and Ihereby generate a lirst round of net demands and equilibrium conditions In lurn the
equilibrium prices will determine the evolution of the stock variables into the next perIod
There are several potential problems with calibration in the conlext of dynamic models The
assumption of a steadymiddotstate growth ~ath in the base case can be questioned First while
27 I Ibullbull
steadymiddotstaU is a possibity it certainly is not the only meaningfull solution to dynamic models
Even in the case of a perfect foresight equilibrium the model implies an equilibrium path which i may or may not involve colanced growth The model not the modeller should dictate the nature 01 I themiddotbase case path Second in the context of a temporarY equilibrium path a steady state solutlcn is Inol a likely model QutcOn1tl In fact unlessmiddot expectations are stattc short run behavior consistent
with a steady-state evolution will in general not be generated On the other hand if static I expectations are self~fufjili[1g we have in fact a perfect foresight model Third even the base year
replication requirement may cause problems In Ihe cOntexl of temporary equilibrium Any
calibration parameter would be conditional on expectation rules which is probably an undesirable
feature
The nalur of the two-requirement calibration strategy is very muchin the spirit of the bull
traditional design of the comparision of alternative equilibria comparis_ion between a steady-state
base case on one hand and alternative paths including a transition period and a finat sleadyyenstate on
the other hand Pereira (198Gb 1987d) compare different (not necessarily steadystate)
equilibrium paths The arguments against such a procedure are based on the idea that the impact of
policy changes can be observed most easily since all departures from the steadymiddotstate can be
attributed to the altemalive polioy On the other hand the base case so dEfined as a steadymiddotstate is
consist~nl with previous work in a less dynamtc setting and Iherefore allows a common standard
for comparing model results
42 Computation algorithms
We are still al a stage in which basically each author uses a different computational technique
Th development of dynamic models - in parlicular wilh adjustment costs andior perfect foresight
bull
28
has corresponded with the decline in the use of fixed point algorithms In -fact given the relative
arge dimensions inevitably involved such algorithms lend to be very inefficient at the best and
chen prohibitively stow See Stone (1985) and Preckel (1965) for a comparative assessment of
different computation techniques Among the models surveyed only Ball1rdFulle(ton~
-ShavenmiddotWhalley (1985) and Feltensteln 1985 use Merrills variant of the fixed point
algorithm 1echnique
AuerbachmiddotKoHikoff (1963 1984 1967) follow a three stage procedure They first compute a
base case steady-state then a revised case steady state and finally transition path for the economy
berveen these two steady-states tn all stages a Gauss-Seidel iterative procedure is used
Ballard (1982) Ballard-Goulder (1965) Goulder (HiSS) and Summers-Goulder (1986)
~
use a method developed by Ballard and Goutder which is similar in many awects to the FairmiddotTaylQr bull
(1983) algorl1hm Short-run equilibria are calculated (using Merrills algorithm) parametric on
nrice expeC~lions The model is then iterated to generate self-fulfilling intertemporaf expectations
and the corresponding perfect foresight equilibrium In a relatively similar approach Andersson
1987 uses a simulation program SIMNON developed In the University of Lund This program
can handle two-paint boundary problems In a fashion consistent with the multiple shooting
algcrlthm (see Lipton-Poterba-Sachs-Eummers (1982))
Boyenbergs computational approach (1985 1985) differs from the other models surveyed in
that he relies heavily on analytical techniques Computations are done by using a dynamic version of
Johansons linearization method Being essentially determined by the continuous time nature of the
I I
f
model thiS linearization model has the disadvantage of confining the analysls to infinltesinal changes
around the base case equilibrium (see Bovenberg (1985) p 53) Pereira (1986b 1987d) uses an optimization algorithm NPSOL - developped by
rGillMurray-Saunders-Wrlght (1986)) This algorithm is used 10 compute the sequence of
I
29
Shortmiddotrun temporary equilibrium which makes up the intertemporal equilibrium path The
equilibrium conditions are seen as nonlinear equality constraints in the minimization of an artlcial
objective function The prices are normalized to the unit simplex by an additional linear equality I I 1 bull constraint
II ErlichmiddotGinsburgh-Heyden (1987) follow a unique approach in that they use a variant of the
optimization technique introduced by Negishl (1960) The economic equilibrium can be generated i
as a solution of a mathemallcal program Ihe objective function of which is a weighted sum of the
utility functions of the various agents while tho constraints set consists of the market dearing 1
conditions Ginsburgh-Heyden (1985) have extended Negishis resulllO tho case of downward price
lrigidities I
Finally the paper by Jorgonson-Yun (1984) is also unique in tIlat it is the only I
econometrically estimated model Different blocks for the consumption and production Side of the
-
model are separately estimated to provide the necessary structural parameters i
I The diversity of computation techniques is yet another indicator of the exploratory nature of the
body of literature surveyed in this article
43 Equal yield comparisions
The link between a base case and counteriactual simulations Is usually provided by the concept of
equal yield Shaven-Whaley (1977) discuss the meaning 01 equal yield in a general equilibrium
context when government is not allowed to run defICits Equal yield is interpreted to moan constant
public utility Government base case utility is maintained in the counferfactual experiments With
balanced budgets the concepl of equal yield is unambiguous The new equUibrium prices and the
balanced budge condition will del ermine the minimal expendilure and taxes needed to maintain base
30
case publi~ utility Accordingly in general equal yield is inconsistent with equal nominal tax
revenue Some change in tax revenue Is necessary Different tax replacement schemes ate
considered to assure that enough tax revenue is collected This is the approach essentially followed
by most 01 the papers surveyed Only Feltenstein (1985) Goulder (198) and Jorgenson-Yun
(1984) chose not to follow an equal yield strategy
The question is of how to Interprete lila concept of equal yield when the government is alowed to
fUn deflcLts The optimal leve of expenditure for base ease public trliUty can now be taAinanced
bond financed or financed by a mix of bonds and taxation We have several versions of equal yield In
particular equal yield may now be consistent with equal tax revenUE Furthermore some meaSure
ltgtf financial crowding out effects of government deficns can be inferred from the comparislon of the
several equal yield alternatives These issues are extensively discussed ~Pereira (1987b)
44 Price expectations and the dynamic generalization of compensation
Indicators
The sum of equivalent and compensation variations over households is the most widely used
aggregated measure of efficiency gains or losses In the context of $teady~state eomparisions andor
when complete future markets exist andlor perfect foresighl is assumed there are no difficulties
associated with the use of the standard Hicksian indicators In fact correct future prices are known
in these cases
If expectations are not selfmiddot fulfilling andlor future markels are not available a dynamic
generalization is necessary 8allardmiddotGoulder (1985) provide some steps in that direction by
defining an indicator that accomodates periods far into the future when households do not have
perfect foresight but a steady~state prevails On he other hand this issue is more fundamental in
i
[
I
i I I I I [
I I
bullI
31
the contex~ of a Gner1 lemporary equilibrium ftamawor~ In such Circumstances Pereira
(1986a) dev~lops a cnamlc generalization which is obtained as the present discounted value of a i i
sequence of short-fUn optimal expenditure functions consistent with a base case expected future I stream of IlWities I
32
5 Empirical evidenee from selected policy Issues
Dynamic tax models have generated several important results which escaped Sialic modellers
The following is a seleCpoundd set of issuos whIch stress the marginal benefits of dynam~ modelHng of
economic behavior in innovltlttive areas of analysis The discussion of a consumption tax emphasizes
the benefits of dynamic tusehold behavior and intergenerational aspects The study of the
flliminatlon or the re~intToduction of the investsnemt tax credits is made meaningful by the dynamic
modelling of production betlavor Modelling governmenl in a dynamic conlexts allows the modeller
to study the irrrpact of government deflcits Finally a dynamic framework lets us appreCiate the I
relevance of consumer expectations in terms of the evaluation of policy alternatives
i
i
51 Consumption Tax
t
By Its very nature a consumption tax can not be adequately investigated with a static model bull
Fullerion-Shoven-Whaicy (1983) use Ihe model of the US economy s described in
Ballard-Fullerton-Shoven-Whalley (1935) to evaluate the movement from the currenl US tax Isys1em to a progressive consumption tax Since their modeJ Incorporates a Jaborleisure choice
where leisure is an untaxed commodlty their results reflect the fact that both the consumption tax
and Ihe present system are dslortionary
Concentrating on the rntertemporaf distortions they show that sheltering more saving from lhe
current US income tax could improve econornic efficiency even if marginal tax rate increases are
necessary in order to maintain government fevenue~ At first you have a revenue shortfall
However the economy moves to a higher sustatned growth path and uitimatefy lower tax rates
cangenerate the same revenue path Also wages increase in the long run with this policy The
bull
bullbull ~
33 shyt
i
I
I
present value of welfare gi[lS for a potiey of complete savin9 deduction wilh marginal tate
adjustments (consumption lax) is simulated to be around $500 billion to $600 billion 01 1973
dollars
They also investiga1e th 1ngth of ttme it takes th~ economy to adjust to these policy changes
Roughly they estimate the long run to b thirty years although this figure Is very sensitive to the
specification of the savings elasticity - a crucial parameler in this model
52 Investment T~x Credit
Goulder-$ummers (1 S87) address the impact of eliminating the lnvestmenl Tax Cradit (ITC) on
intersectoral capital formation and on economic growth Their model isYa1ticularly adequate to
addies$ this issu~ in that is postulates a forward looking investment m9we with adjustment costs
They show that the eliminating the ITC causes a reduction in the rale of investment Inveslment is
estimated to fall by about 7 in the short-run and by about 12 in Ihe new long run steady state
In lurn a previous policy anncuncement lowers Ihe overall al1ractiveness of Investment and leads 10
a downward shift of the Investrrent profile On the other hand the combined effect of a revenue
neutral simultaneous efimiraron of the ITC and reduction of the corporate tax rates is a long run
reduction ef the capital Sieck by 35 This pattern suggests that a revenue neutral increament in
both Ihe corporale tax rates and the inveSlment lax credits would be preferable 10 the formula
adopted in Ihe US Tax Reform Act of 1986
Pereira (1987d) addresses the impact of the ITe from Ihe oland point of Ihe curTenllax syslem
under the Tax Reform Act of 1986 Given Ihe concern over the potential depressiVe effects upon
savings and investment of the new tax system in conjunction with deficit reduC1ion mechanisms of
the GrammmiddotRudman type a pertinent question is should Ih lTC be e-introduced In the context
bull
34
-of his mod~1 of the US economy Pereira focuses on the trade-off between the distortion in the
intersectoral allocation of capital induced by the lTC the lower tax revenues it generates and
potential financial crowding-out effects on one hand and the positive effects of lowering the relative
price of new capital goods on the other hand Preliminary results confirm the qualitative results in
Goulder-Summers (1987) suggestting a welfare gain of about 2 of the present value of GNP in
the best scenario of absence of replacement taxes Furthermore preliminary results show an
important time pattern to welfare gains with the average benefits increasing the further you look
into the future This pattern is due to the presence of constraints to intersectoral mobility of capital
and inertia in the adjustment towards the optimal capital levels as reflected by the presence of
adjustment costs
I
I
bullAuerbach-Katlikaff (1gB) in the context of their intergenerational model of the US economy
consider the impact on savings capital formation and interest rates of deficit policies and balanced
budget increases in government consumption They conclude that deficit finance and government
consumption can significa-ntly crowd out capital formation and lower the welfare of future
generations However crowding out from deficit finance is a very slow process because it results
from increased government spending over potentially long horizons Also deficit policies may
substantially influence the longmiddotrun interest rates while leaving the short-run rates essentially
unaffected Finally and in opposition to the central role adjustment costs seem to play in both
53 Financial crowding outeff~cts of government deficits
Goulder-Summers and Pereiras models Auerbach-Kotlikoff suggest that the time path of interest
rates induced by a policy of deficit financing seems to be insensitive to the presence of adjustment
costs
bull
35
54 The role 01 consumer expectatlons
The expectations analysis has uncovered one of the benefits of dynamic modelling
Ballard-Goulder (1985) find that the appeal of adopting a consumption tax depends on the level of
foresight possessed by the consumers Furthermore additional foresight may be welfare worsening
This is a second best type of result This tends 10 occur under policies that lead to capital deepening
and declining rate of return to capital over time To the extent that consumers have more foresight
they will be beller equippod to anticipate the fall in the rental price of capital Consequently if a
saving incentive is enacted people save more with myopia than with perfect foresight Given the
exi~tence of taxes on capita and the discrepancy between the private an~--sccial returns to capital
the greater savings with myopia is beller socially Ballard-Goulder find that the welfare gains
from a consumption tax are reduced by about 10 when we move from myopia to farleet foresight
Therefore the attractivenness of adopting a consumption tax seems fairly [obust across these
foresight specifications
I
bull
36
6 Tbe power and weaknesses 01 dynamic modelling the example 01 corporale
tax Integrallon
The Issue of corporate tax integration allows us to sh~w the stre-ngths and weaknesses or the
dynamic approach Empirical evidence using CGE techniques indicates that depending on the precise
scheme of integration and the 1ax replacement methods integratiOn may have substantial effects In
Ihe work of FuliertonmiddotKingmiddotShoven-Whalley (1980 1981) total integration was found to yield an bull
annual static eHielency gain of $4 to $8 billion in 1973 dollars Simulated dynamic gains may be as
large as $695 billion or abcu 14 of the present value of tulUre consumption and leisure in the
US economy These gains result primarily from interindustry reallocalions of Investment and an
improved inter~emporal allocation of consumption
Mora recent work emphasizes hOw consumers asset portfolio decisions and firms finanCial
decisions affect the efficiency gains from integration Slemrod (1980) focusing on consumers
asset ponfoio deCisions finds static efficiency gains which are about twice -as Iarge as those
repOrted iiy Fulierton-King-ShovenWhalley (1961) FuUrlon-Gurdon (1983) focus on the
firms financial decisions They roporl efficiency gains of 6 of GNP from the elimination of the
tax distor1ions favoring debt However when they eliminate the corporate tax and repiace it with
increaSed persona income taxes addlHonal dls10t1ions are created in the optimal labot~leisure
decisions These distortions tend to dominate the analysis slJ(h that the overall effects of complete
integration are very modest Galpr-LuckmiddotTodr (1986) address simultaneously consumers
asset pOrlfollo decisions and the firms financial decisions They also find very modest efficiency
gains from integralion
The above results are important but may be severely biased There are severa aspects of
economic behavior and modelling crucial for the study of Income tax Integration whiCh have no been
i 37
captured innny of the above papers One aspect is the absence of government deficitsmiddot a balanced
budget is assumed It is true Ihal resource crowding OUI is caplured In these mOdels but financial
crowding out induced by government spending is nOI Second investment is not derived from
optimtztlcn behavior Investment behavior passively accomodates endogenous saving decisions As
a consequence the differential impacts of policies in the incentives to save and to invest are not
captured Aso full capital mobility across secors is assumed with instantaneous capital
adjustmenlS towards optimal levels This assumption rules qut different costs of capital across
seclors and therefore differentiated reactions to tax policies changes Finallyhe modelling of both
gov9rnmerlt deficits and of endogenous real and financial investment decisions necessitates the
~nsideraiion of a dynamilt framework and the introduction of financiaJ assets govemment bonds and
iprivate financial assets A dynamic framework also highlights the efficiency effects of Integration on r
ithe ptimal intertemporal decisions The above models are either sIalic (Slemrods and
GalpermiddotltJckemiddotTode(sj or a dynamic sequ~nce of otherwise Slatic model bullbull I Pereir a (1986b) develops a dynamic applied general equilibrium model bull to study the tbull
efficiency effects of integration on the growth and allocation of investmenl across sectors Special
Iattention is paid to the structural affeelS of resource and financial crowding oul induced by
gove mect spending and deficits and to the real and financial Investment reactions across industries
to both tax integration and flnancial crowding out His mode departs from previous work on income
lax Integraticn and for that matler from most 01 the CGIE literalure for tax policy evaluation in
several direcllons It encompasses an endogenous sequential equilibrium struelure founded on
dynamic behavior with flexible expectations Government deficits are optlmaUy determined
Investment decisions are tward looking and the resull of optimizing behavior Several financial
assetsmiddot public and privatemiddot are considered
Simulation resulls Indicale that the welfare gains from Integration under large deficits are at
32
best lJlodest when measured in terms of the GNP The elimination of the corporate tax and its I replacemenl by increased income lax rates yields long run benefits which are never larger than
2 of the presenl yalue ct telr consumption and leisure onder the previous tax regime and 1
under the current tax regime However the short run w~lare effects of full integration tend 10 be
very low and in in some scenarios even negative This is a flew intertemporar pattern of efficiency Ieffects which reflects an adjJsiment lag in the interindustry investment decisions due to the
existence of costs of adjustment On the of her hand partial integration achieved by excluding I dividends from the corporale lax base syslematically yields negalive efficlencyeffeclS In Ihe long Inm these can be as high as 3 of present value of the future value of consumption and leisure This
is a new second best effect suggesllng Ihat less than complete integration may have perverse I efficiency effects
The dynamic slructure of PerclrBs model snowing for an improved freatment of real investment
declsions and government deficts provides a potential setting for a more accurate measure of the
costs and benefits of inlegration The simulation resullS suggest lower benefits and an intertempora Ibull
pattern of growing benefits
Simulation results also clearly suggest robust negative effects from partial integration How
reliable is this result If dJvidends were deductible from the corporate lax base (as are interest
payments) the preferenlial Ireatmant of debl over equily would be ellminaled Corporallons should
be expecled 10 react by decreasing the optimal deWeqully rallo However corporale financial
gecisions are exogenously set In Pereiras model as in aft of the other models surveyed that go as far
as dealing with the issue
Also withdrawing dividends from Ihe corporate tax base is a way of eliminating Ihe double
taxation of dividends at belh Ihe corporate and personnal Income levels How will the optimal
aliocellon of household saving accomodate that change in Ihe relalive return to corporale equity
39 I
t However hOusehold portfolio decisions are exogenouly set in Pereiras model as in all of the other
models surveyeltj go as far as dealing with the issue
IIt is sate to say that the evaluation 01 the benefits 01 partial Integration as defined above are
sevrey underestima1eO Meanwhile the distortionSoenerated by the increase in the marginal
personal tax rates designed to ralse the revenue foregone by the dividend exclusion are fully
aCCltlunted for A good measure of the costs of integration together With a poor evaluation of the
benefils may very well be the reason why partial integration is Simulated 10 yield neg alive
eHikiency gains
On a different vein as most of Ihe model surveyed here Pereiras is a closed economy model It
is legitimate to wonder how the resullS would change If international capital flows were anowed
IThis almost certainly would affect financial crowding out since a good d~al of the capital inflows
could be used to finance public debt
For these various reasons the results should be treated as preliminary but they strongly I suggosl that the dynamic structure provides valuable new insights into the corporate tax integration SSU2
40
7 Concluding remarks I I I i
What has been acccmplished Ihis far The success of Ihe CGE research in laelding the challenge
of dynamic modelling can not be denied
I iGreat progress in the modelling of household behavior has been achieved Promising
developm~nts in the modeUing of production and government behavior have also been accomplished
Interesting innovations in the concept of equilibrium and computation techniques were discussed
The policy analysis was greatly enriched by considering transitional effects together with longer
run steadymiddotstate equilibrium paths generalized equal yield strategies accomodating different
financial crowding out impacts and generalized dynamic policy evaluation indicators
An important set of issues have been addresed by the models surveyed Traditional issues
focusing on the effects of capital taxation and consumption taxes have been pursued In terms of
capital taxation for example the intertemporal nature of the issue was enhanced by allowing an
optimal evolution of capital stock in the economy and forward looking optimal investment decisions
In Andersson (1987) GouldermiddotSummers (1987) and Pereira (1986b 1987d) this is coupled
with a improved treatment of several tax provisions like investment tax credits and depreciation
allowances In terms of the consumption taxes developments in the specification of Intertemporal
household behavior the introduction of overlapping generations and the modelling of
intergenerational links via bequest motives as well as a closer attention to demographic evolution
provided a much improved economic setting for the understanding of the several aspects of the
problem
Furthermore dynamic modelling has permitted the addressing of a set of new issues Policy
Issues ranging from dynamic tax policy analysis to the evaluation of exogenous changes in
government expenditures to the impact of different methods of financing government expenditure to
4 1
the importance of tinanciai croHrii1g out to 1he relevance of consumorS expectations in tne policy
results have been addressed
Despile all of the progress and in part due to such progress several avenues are wide open for
much neded additional research The following seem to be the most Interesting and promising
areasmiddot First the specincation of liquidity constraints to household behavior may help to obtain a
better understanding of policy chenges that affect directly the interest rates in the economy Second
major attention needs to be paid to the incorporafion of a roraign sector (with open capital markets)
Third some efforts are needed in terms of finding adequate computation techniques now that
dynamic modelling eally makes the curse of dimensionality worse than eVOrbull
The modelling of financial markets is the single most unsatisfactory speet of the dynamic
models surveyed here~ The problems associated with provlding a mo~ adequate treatment of
financial markets and in general endogenous and optimal financial deci$ions bull both at the corporate
and at the personal levelsmiddot have 10 be addressed That necessarily requires Introducing uncertainty
into the models To b adequate this must go well beyond the mere assumption of point expeetalions
with the whole array of modelling implementalion and evaluation problems thai generates
bull I i
42
REFERENCES
Abel A and O Blanchard 1983 An Intertemporal Model of Saving and Interest Econometrica
51 pp 675-92
~ Andersson K 1987 Tar-ation of Capital in Sweden - A General Equilibrium Model Ph D
Dissenation UnIversity of Lund Sweder
Auerbach A L KotlikoH 1983 National Savings Economic Welfare and the Structure of
Taxation in M Feldstein ed Behavioral Simulation Methods in Tax Policy- Analysis Chicago
University of Chicago Press
Auerbach A l Kotlikoff 1984 Social Security and the Economics of Demographic-
Transition in H Aaron and G Burtless eds Retirement and Economic Behavior Washington DC
Brookings Institution
Auerbach A L Kotlikoff 1987 Dynamic Fiscal Policy Cambridge Cambridge University
Press
Auerbach A L Kotlikof and J Skinner 1983 The Efficiency Gains from Dynamic Tax
Reform Internatiooaj Economic Reyiew 24 pp 81middot100
Ballard C 1983 Evaluation of the Consumption Tax with Dynamic General Equilibrium
Models Ph D Dissertation Stanford University
Ballard C 1987 Tax Policy and Consumer Foresight A General Equilibrium Simulation
Study Economic Inquiry 25 pp 267-284
Ballard C and L Goulder 1985 Consumption Taxes ForeSight and Welfare a Computatonal
General Equilibrium Analysis in J Piggott and J Whalley eds New Deyelooments in Apolied
General Equilibrium Analysis Cambridge Cambridge University Press
Ballard C D Fullerton J Shoven and J Whalley 1985 A General Equilibrium Model for
t 43
Tax Eoliey Eyalua~iQO NBER University or Chicago Press
Borges A (1987) bull A Survey of Compula1ional General Equilibnum Models CECD Economic
Slud~ 8
Bovenbig L 19B3 imperfect Capital Mobility in a Perfect Foresight Herberger Model
Mimea UC Br~eley
Bovenberg L 1984 Capitel Accumulation and Capital Immobility Qmiddottheory In a Dynamic General Equilibrium Framework PhD dissertation University of California at Berkeley t
Bovenberg L 19850 Dynamic General Equilibrium Tax Models with Adjustment Costs in A iManne ed bull 1985 EtQnomrc ECUjilbrium Model Formulation and Solution Mathematical
lProgramming Siudy 23 Arnsterdam NorthmiddotHoliand
Bovenberg L t 1986 Capital Income Taxation in Growing Open Eeondmies Journa of Pubtic
Eamp(omlcs 3t pp 347middot376 IBovenberg L ald W Keller 1981 Dynamics and Nonlinearlties in Applied General
Equilibrium Models Internal Report Mlmiddot81middot208 Department for Statistical Methods CBS bull
Vooiburg The Netherlands
Charnley C 1981 The Welfare Cost of C~pltal income Taxation In a Growing Economy
IJournal of Political ECQnQm~ 89 3 pp 468middot496 I
Charnley C 19810 Efficient Tax Reform in a Dynamic Model of General Equilibrium I Mimeo Yale University I Charnley C 1982 On the Optimal Taxation In Stylized Models 01 Inlartemporal General
Equilibrium Mimeo Yale University I Decaluw B and A Manens 1985 Pays en Oeveloppement at Modeles Calculables OEquilibre
General - Une Revue de al litterature empirique Monographie Sene 9 Volume 3 Centre de
Recherche et Developpemont Economlque Unrslt de Montreal
44
bull
I I
I
I
1
Erlich $ V Ginsburgh and L van der Heyden 1987 Wher~ do Real Wage Policies Lead
Belgium bull A General Equilibrium Analysis European Economic Reyiew (forthcoming)
air R and J Taylor 1983 Solution and Maximum Likelihood Estimation of Dynamic
Nonlinear Ratioant Expeclations Models Ecooometrica 51 pp 1169middot1186
Feltenstein A 1984 Money and Bonds In a Disaggregated Ec~nonYin Searl 1-1 J Shoven
Eds Aoolied ~eoeral EqJiliDCiulD tioalYsis Cambridge University Press Cambridge
Feltenstein A 19B6 A Dynamic General Equllibrlu Analysis of Financial Crowding Out
Theory with an Application to Australia Journal of Public Economics 31 No1
Fullerlon D R Gorden 1983 A Reexamination of Tax Distorllons in General Equilibrium
Models in M Feldstein Ed Behayora Simulation Methods in Tax pOlicy Analysis Chicago
University of Chipgo Press
Fvllerlon D J Shaven and J Whalley 1983 Replacing the US Income Tax with a
Progressive Consumption Tax JoulOal of Public ECQnomics 20 pp3middot23
Fullerton D Y Henderson J Shaven 1984 A COmparision of Methodologies in Empirical
General Equilibrium Models of Taxation in Searl H J Shoven Eds ~pQlied Geoaral Jguilibrium
lIoalyss Cambridge University Press Cambridge
Fullerton 0 A King J Shaven and J Whaney 1981 Corporate Tax Integration in the United
Statesmiddot a General Equiiibdufi1 Approach American Economic Review 71 pp~ 677~691
Galper H R Lucke E Tader 1985 Taxation portlor Choice and the Allocation of Capital A
General Equilibrium Approach Mimeo
Gill p W Murray M Saunders M Wright 1986 Users Guide for NPSOL (Version 40) A
Fortran Package for Nonlinear Programming T echnicsl Report SOL 86middot2 Deparlment of
Operations Research Stanford University
Glnsbur~h V and L van der Heyden 1985 General Equilibrium with Wage Rigidities An
45
AppJication to Belgium in A Manne ed t Economic EQumprit1m~ Made Formulation and $oJioo
Mathematical Programming Study 23 Amsterdam NorthmiddotHolland ~
Goulet J 1968 1 Adjustrent CO$1S in the Theory o(the Firm Reyiew of EconomiC Studies 35 bull
pp47middot55
Goulder L 1985 Tax-Financed versus BondmiddotFinancedGhanges in Governent Spending
bull Mimeo Harva(d UniversilYmiddot
Goulder l J Shaven and J Whalley 1983 Domestic Tax Policy and the Foreign Sector The
importance of altemative foreign policy formulations to results from a general equilibrium tax
analysis model In ~n methods in Ibe antico of income from caoia1 ed Martin
Feldstein Chicago University of Chicago riSS
Goulder l L Summers 1987 Tax Policy Asset Prices and 3rowtly ~
A General Equilibrium Analysis NBER Working Pper No 212B
Grandmont J 1982 Temporary General Equilibrium Theory Ch 20 in Handbook 01
Mathemalical Eccnornics Vol II NorthmiddotHoliand
Harberger A 1959 The Corporate Income Tax An empirical Appraisal In lax Revision
Comoendjum Voll House Cornmitee on Ways and Means Washington DC Govemment Printing
Office
Harberger A 1962 The Incidence of the Corporate Income Tax JOULn1 of Political
Economy 70 pp 215middot240
Harberger A 1964 Taxation Resource Aliocalrlon and Welfare fn lb Role of Dir bullbulll and
Indjresl Taxes io tbe Fede[al Reserve s~I~Ill Princeton Princeton Unlveresity Press
James J 1985 New Developments in the Applications of General Equilibrium models to
Economic History in Piggott aand Whalley Eds
Jorgenson D and K Yun 1984 Tax Policy and Capital Allocation Harvard Inslitute of
bull
I
I f
Economic Research WP 1107
Judd K 1965 ShonmiddotRun Analysis of Fiscal Policy in a Simple Perfect Foresight Model
Journal of poJiHcal 5CCJiW1l pp 298middot319 r
Upton D J Potbc J Sachsa nd L Summers 1982 Multiple Shooting in Rational
Expectations Models fuoometrlc 50 1329middot1333
Manne A ed 1985 fconQrnjc Eoullibrium Model FOrIDujaljQo and Solution Mathematical
Pogrammlng Study 23 Amsterdam NorthmiddotHolland
Manne A 1985 Or the Formulation and Solution 01 Economic Equilibrium Models In A
Manne ed ECQOQlli-Jr~ibmiddotum ~tQdel EQlJulailon and Solution Malhematical Programming
Study 23 Amsterdam NorthmiddotHolland
Mansur A and J Whalley 1984 Numerical Specification of Applied General Equilibrium
Models in Scarf H J Shoven Eds Aoplied General Eouillbrium AnaiysectIsect Cambridge Universily
Press Cambridge
Merrill 0 1972 Applications and Extensions of an Algorithm thaI Computes FixedmiddotPoints of
Cenain Upper Semimiddotcominuou Point to Set Mappings PhD Disserlation Unlvel$ity 01 Michigan
Negishi T 1960 Welfar Economics and Exislence I an Equnlbrlum for a Competitive
Economy MelroeCOOQ11ka 12 92middot97
Pereira A 1985 On the Existence and Uniqueness of Optimal Ou1put Path lor CRTS Firms in
Adjustment Costs Technologies Mimeo Stanford University
Pereira A 1985a Consumer Expectations and Ihe COSI 01 UllUty In an Intertemporat
Framework Mlmeo Stanford University
Pereira A 1985b Corporale Tax Integrallon and Government Deficits A Dynamic General
Equilibrium Anaiysls Mlmeo Stanford University
Pereir A 1987 On Ihe Computation of the Users Costs of Stocks Ecooomi (forthcoming)
47
Pereira A bull 1987b Equal Yield Alternatives and Government Oeficlts Mimeof Stanford
University
Pereira A 1987c A Dynamic Applied General ~uHlbrium Model for Tax Policy Evaluation
Comlletamp Docurnent2tion Mimeo University of California San Diego (in progress)
Pereira A 1987d Should the investment Tax Credit be Rei(ltroduced Mlmeo University
of California San Diego (ir progreSs)
Piggott J and J Whalley eds 1985 tJew QeyeQpments in Applied General EQuilibrium
bullenalysjs Cambridge Cambridge University Press
Preckel Pbull 1985 Alternative algorithms for computing economic equilibria In A Manne Ed
Radnor R 19amp3 Equllibrium under Uncertainty In K Arrow M Intrilligaor ods
I ibull
Robinson S 1 g86 Mullisectorai Models of Oeveloping CountrieS a Survey Oivision of
Agricullure and Resources UC Borkeley WP 401
Scarf H 1967 The approximation of fixed points of a continuous mapping ~MM Jouwal of
epplid Mathornallcs 15 pp 1328middot43
Scarf H with T Hansen 1973 00 the Computation of EconomiC Eguilibria New Haven Yale
University Press
Scarf H J Shoven Eds 1984 61l1llied Genera Elujlibrium 6nalyssect Cambridge University
~ Press Cambridge
Shoven J 1976 The Incidence and Efficiency Effects of Taxe on Income from Capital
Jurnal of ramie1 Economy 84 pp 1261-83
Shoven J 1983 Applied General Equilibrium Tax Modemng IMF Slaff PaoerS 30
Shoven J and LB Simon 1986 Share Repurchases and Acquisitions An Analysis of which
Firms Participate NBER Working Paper No 2243
bull
48
IShaven J and J Whalley 1972 ft General Equmbrium Calculation of lhe Effects of
Differential Taxation of Incoma from Capital in the US Jauroe of public Econgmics 1
pp281middot321 (
Shoven J bull and J Whalley 1973 General Equilibrium with Tax$s a computation procedure
and an existence pr00f ReviEhi of Economic Studies 40 pp 475middot490
Shoven J J Whalley 1977 Equal Yield Tax Alternatives A General Equilibrium
Computa~ui0nal Tfchnique J~HJrnal of PUblic EconomIcs 8 pp 211~224 bull
Shaven J J Whalley 1984 Applied GeneralmiddotEquilibrlum MOdels of Taxalion and
International Tr8de An Introduction and Survey Journal of Economic literature 23t pp
1007middot1051
Slemrod J 1980 A General Equilibrium Model of Capital Income Taxation Ph D bullDissertation Harvard Univnrsity
Slemrod J 1983 A General Equilibrium Model of Taxation with Endogenous Financial
Behavior in M Feldstein edbull Behavioral Simulation Methods In Tax Policy Analysis Chicago
UniVerSity of Chicago Press
Stone J 1985 Sequential Optimization and Complementarily Techniques for computing
Economic equilibrium in A Manne Ed
Summers L 1981 Taxation and Corporate Investment a q-Theory Approach BroQkinas
Summers L 1985 Taxation and the Size and ComposHion 01 the Capital Stock An Asset Price
Approach Mlmeo Harvard University
Whalley J 1975 A Geneal Equiloibrium Assessment of the 1973 United Kingdom tax
reform Economica 42 pp 139middot61
1JLTIllOS JIQll~)ltG PAPERS PUBLICADOS bull
I
I I
j
112 70 - LUCENA iJiCIIJo ~ IIEnvironrilent Honitoring and Orqanization Structure T (Setellibro 1987)
nI 71 - L~ENA Diogo A Note on the Representation of Information Stxueture (SeteJObro 1987)
nQ 72 - LUCENA D1CgO Envuorment 111)nitoring and Organization
Structure II (Setelilhro 1987)
ril 73 - AlmmE3 Antonio Pais e CAsPAR Yitor bull Tributacao Inoentivos e InvesUlllentos Analise Qualitativa noYellibro 19S7)
n2 74 - liAlltOSA rose Pedro Opt1lllal lIage R1gidity A SUggestedtlthodology to Teet tb Theory and an Application (Outubro 1987)
nI 75 - CABRAL L 11 B Three lIotes on SYJIIlletric Gemes witt itsYJIIlletric Equilibri (NoveJObro 1987)
n2 76 - LUCENlt D10gOTo Search or Not to Search (Fevereiro1967)
nQ 17 - SA Jorge asconcelos euroI A Theory of SynergyB (Fevereiro1987)
n2 76 - VILARES tanuei Jose Os Be IntenaMios IlIIportados ColIIO Factor de Produao (Julho 1987)
nQ 79 - SA J~rge VasCQncel03 e HoY to Copete and Co~unicate in tature Industrial Products_ (aio 1988)
n2 80 - ROIl Rafael Learning and Capacity Expansion in a New larket Under Uncertainty (Fevereiro 1988)
n2 81 PEREIRA Alfredo llarvilo Survey of Dynamic Computational General EquilibriUlll Hodel for Tax Policy Evaluation (Outubro 1987)
Qualquer in ormacao sabre o 1IIorlltlng Papers ja publlcados sera prestada pelo Secretariado de Apoio aos Docentes podendo os ~e3OSf
ser adquirido5 ne Secyo de Vlndas da Faculdade de EconollLia UNL na Travessa Esteviio Pinto Cltpolide - 1000 LISllOA
_
--
--- -~ ----- __-shy
_ _ - _- - Con~um9rs_ 8amphavlor
Bajla rd -Fulflonmiddot -Shaven-Whalley
1985 Anderaaon
1087
Auerba ch~ Ko1Uko 1983 1987
BaHElfdmiddotGouldor 1983 1905
Sovanborg 1985 198e
ErHchmiddotGlnaburghmiddot middotHeydeR
1987 Feltensteln 1984 1966
Gouldar 1985
GouldorSummor 1987
Jotgeneon-Yun 1984
Porelr 1966 1991
Count dynamlc dl9agg~ objecth lastes constraint ondog endog OvorlapBnqvMI of datil
year _seltlng labor no savings gonorst mollves
inlertsmporal 12 Income otility eel IntertempOfll1 yes yo no noLS 1973 Ymiddotmiddot
_ groups max ---_
0 Intertnmporal rllcufsivs
Swaden 1984 yes aggrogate utility CES eq of molion 00 yos na no consumer max for weaHh
tntortorrporal Isoeiastlc~ 00-1983
ill yo 55 poundtUS u~illy middot19831 In tcemporal yesmiddot 11El7 yes yo 00
cohorts rn~x CES-19C7 plauslblo
bYPothlC1 inlsrtamporal
19n yes 55 ago uliJity Isoelastlc Intertemporal ybullbull ye ye ysLS BFSWexpanOOd _ cohorts max
0118 inlertemporal recursivo LS plausible ye aggregate ollllly lsoolastic eq ot motion no yes no no
hYPOlhotic consumer max for weallh jnterlemporlll Boigium 1960 1983 ro aggrogate utility St()ne~Gearyjner1emporal no yo no no
consumer max Inlerl(lmporal two periodmiddot two yes
Austrs1i 1991middot1982 yes 2 income ulllly CD yeluly ybullbull tn the flrsl no no 2-perlod) groups max constraint porlod
Intertamporal LS 1973 ybullbull 55 aga utility eel intarhtmporal ye ybullbull ybullbull y
BFSW bullbullponded cohorts max
intert9mpor~1 W 1973 ves aggregate ullily 180085110 intertemporal no ybullbull no no
BFSW bullbulllMded consumer max (I~gt IntertempQ(a~
W 1955-80 ybullbull aggregate ulilily trenslog intertemporal ybullbull yes no no consumer max
Inlerlemporal recursive 1973LS yes a income ulllily CmiddotD eq of motion yos ybullbull 00 no
BFSWxpanded groups max for wealth
middot5middot
bull-- _---- -- _-----shy~ -~
I AlJlt 1 - uynafllc (lIt MOOero fur all runey tVlIIVllllQlI
dynamic dluggI utllng
I Ba~ll-a rd-fullerton
-Shoven-Vlhalley I no 19 sectors 19S5
Andarsson 1987 ye 2 sectors
~ Auerbach~KotHkof no~1gB3
1983 1967 yesmiddot1g87 1 sector
Ballard~Go uld ar 1983 1985 no 19 sectors
Bovenborg 1965 1986 yes 2 sectors
Erllch~GIn5burg hshy-Heyden no 24 seclors
1957 1__ Feltenataln 1964 196e no 30 sectors
GoUldsr 1985 no 3 sectors
~ GoutdarmiddotSumrnamprs
1987 yes 5 sectors
_ Jorgen9ClIO-Yun ona sector
1984 no produces two goods
Poralr 19S~ 1987 yos 4 sectors
~bJecllve
Produc~rs
technology behavior InpIJ1 1-0
matrix adl costs opt Imal
capital stock
sialic cost min
CES in valu$ added
110
capital labor yes 00
np Invest adjusts
to sFvnns
max 01
interlempora flot rash flow
cnsl min-lOS) Imax of valul)
of the firm~1987
ill
CD-ISBa CES-1987
capital labor
captel labor
no
no
YA quadratic in
I nO-191)3J
qutldratic in Ilk-1gB
ye
YAS
static cost min
CES in vafl1oadded
10
capita labor yes no
no invest edjuss
10 savifQs max ot capita yas ~
value ot the CD labor 00 quadratic In yo Ilrm Ilk
static CD in several no global ltlost mln vaiue added
110 ltsp1al
Jabor Ves no Inves adjusts
to MVrHIS
statle oost min
____ slatio
cost min
max of value of the
firm equivalent 10 two~stage
~ost min me) of
CO in value ad~ed
110 CES In
value added 110
CESin shyvalue added
110 dual
trBRslog price function Cobb-Douglas
capUal labor
cepHal labor
capital labor
capital labor
capital
yes
yes
yes
no
no invost adjusts no to S8vinos net of
El-xog 9011 deflcitsl no
no invest adjusts to savings r
yes quadratic In Y
Ilk
no yo I
yes Intartamporal net cash flow
In V Added 110
labof Y quadratic In I
y bullbull
- 6 -
~ --- ~ ~- -- -~------ --- --- - --~--~ ---~ shy
_
- ---
--
--------
- - - --------
H n v I taxes
olllng allowed debt Ballardfullamprton~
dynamic objective constrfnt tupendlttire (IlleUs endogenous
aU major balanced trade yoarly ybullbull rly endogenous -f ~Shovon WtlitUy US taxes with C(in~Hlt00 pubHc uility balancod cQmposilon no no debt
eiasticllv junel l 1985 rnagt eCP-REV~d1 aU fn3jO[AndtH3Son vearly balanco small opan Gcon J Swodish inturn fa to cr raturn1987 no - inaia dlJn axogonolJs 00 Ini ITnininod poundgtPfUV
-I L)~ in~h~ --~ ~~ ~pound~~---1- ----shy
runrac h~KQllfkofi ai gulD
1963 1937 US m1l3syo - intertnmponl tixogancuamp yo ys cos~d ampConoroy bandmiddotfinanced flon-oppoundIla
B iii lIa fe-Gould 9r nil llltl1gtr balanced 1rndayearly yearly crcoge)Jus
US t)X6S wiih conslanl I
1963 1965 no public utility balanced cOlYlposWon~ no nD debt max budnt ExP~RCV alastlcHy lunct
Bovenberg simp~iriad 1985middotclosod econybullbull rl~ no - balanced exogenous no no debt US laxes 19a6~two counlry1985 1986 bodgel model
Ert leh~G tn aburghmiddot yeally simplified Inlert walt funet Ishorl~run ttads Heyden no balanced exogenous 00 no Belgium
taxes _ desloullbtium flltUensleln
1967 budaet EXP=REV 6-1mprifia-dmlm costs of yoarly yo rast of tho world
1984 1985 producing a ExTax Rev exogenous financed with yo Australian ~s a Ihlrd conSUmer public good money and bonds non-optimal taxes group
Goulder yaarly annual and endogenous aU maJor
laquo1()Sed economy1985 publlc utilily In1ertemporal composillon bull yes yes U~ taxesY rna versions IlEFElQgtfpoundV bond-financed (non-opllmal balanced trademiddotGoulder-Summer yeary yoarly endogonous bull all major
1987 00 public utility balMc composition no no US taxes with constant max budGel EXPREV alasticilY fund Jl)rgenaon~Yun yoarly aU major
no balanced exogenous no1964 US taxes closed economy budGet inteftemporal recurslve endogenousPereira all major
yo public utility eq of molion lovel and yo yos1995 1987 US laxes closed economy rna for debt composition bondmiddot financed oplloll
- 7 shy
bull - -__----- _ - ------ -- ------- _-----
---- ---
------
------
ll-ltiLt 1 uynamlc Lut MOOSS fOf I ax YOtlcy tv8luallon
Inanell ~ets and maf~9ts ojumLrtm- ----_ ----c_~
privata do btl GlvldtHlI pL~lc aliocalQflo rls shy
1rlc ypa expGct assets equity retention dobl 01 savlnq$ promlums path
Sal ra rd fu Barton phys cap1 used 10 buy soquence lt
Shavon-Wha1ley no financial exogenous oxogenous no physicBl 00 bull 01 sIalic 1E myopic 1985 assets capital flqrlil -
Andersson phys cap used 10 buy 1987 no financial no no yo physical no dynamic ~ perfect
assets capital Auerbach~KolUko If phys cap usod to buy
1983 1987 neo financial no no physical cap no dyoamlc ~ perfectYmiddotmiddotnSSf)ts and Qovt debt SailBrd-Goutder phys cap
1983 1985 no financ[al assets-shy -------
Bovenberg phys cap 1985 1986 no iinenclaf
asso1s Ertlch-Glnsburgh~ phys cap1
-Heyden no financlaJ 1987 assefs
FeltoRsteln phys cap 1984 1986 Invostment is
bondmiddotfinanced
used to buy
exogenous exogenous nObull physl~al no dynamic ~ perfllcl capital
used to buy
no no no physical capital
no dynamic ~onlinuos lima
PI perfect
used 10 buy no no no physical no dynamIc PF perfect
capital jllNt) porlod) bull used 10 buy
no no ye physical cap no dynamic PI p4)rfect and govt debt (Iwo Iods) bull
Goulder phys cap used to buy 1995 no financial exogenous exogenous yenbullbull physical cap no PFlfE bull IdynamIc perfoct~assels and am dabt slaUc
- -~ - shyGoulder-Summers phy$ cap
1987 bonds and no e)(oganOU$ no used 10 buy ye dynamic PF perfect _ bullbullultv bonds MullY (exoaenous
Jorgensonyun phys cap 1984 no financial
essels -~ - ------
Pereira phys cap 1986 1967 bonds end
equHy
________
no
8xogenoos
no
exogenous
no
ye
used to buy two type$ of physical
copltal used 10 buy
go dobl bonds equity
no
no
dynamIc
-
dynamlc
~
lE
perfect
fIbullbulllble (nonperlecB
middot6middot
~-- _middot~ ___~w_ - - ___ bullbullbull___ -------- shy~
-----
------
- ---
-
IlflpnUItll tit lun pVII-y PIlUAlIV
q6H1brla oqual fax polley ovnluaU-on efUclency derlvatton Ioorllhm
calibration param~hr computation COmelfed llold tndfcaHH eHecls
Bnl jard~FuU6rtQ n base year back soll1lion steady-sta1e yo inlsrgroup inerseel or Iii
VS transilion lump-sum with Interfemporai 1965
~Shovn Whallmiddotv replication wllh exog Morrills nmamo1ers alQorilhm _~ond~middotsta1$ ears Inc CVor EV
i~noerucn sleady-sfala no agg(ogaled lnlalsectora 1967
base year back sobHon SiMNON package replication wHn oxog valiant of VS Iransilion no Indicamptor lnlertamporal
paramelsrs multiple shoo ling 6~~~l~slat8
AU8rt)ach~Kotllkot steady-slato yo Inlerguna-r a lions
1893 1llH
quatitaUve by replication a$Sium~lkml Gauss-Soidel liS Uln1WOn any rule -lilh 1111ermpOrB
sonsililJ1y I _ ~poundI~Jjnnelill $t(lfdvmiddot~luiA J3djlt15S ~~ nnrilW ~--------
a ail r nimiddotGould or bas9 9tH bac~ solution Ballard~ staadymiddotslalo )05 kltHSclo(a~
1983 1985 VS Iransllinn lump~sum dynalTlk CV inlerlemporal steadY slafe parameters alaorlthm replicaUon+ with oxog Gourder
+stead~stato (ers loc inlareel10rat BCvonborg qunmaiive by dynamic version stoadY-$Igtto yes 1985 1985 YS transition some govt eN Inlortempora
S9rHltivilv OnearizaJian method ropllcaUon assumptionl of Johansons
+steady-sfata paoo palh Erllch-Glnsbu rgh~ actual vorsus no aootagalod inlersectoral2 bftse years back solu1ion variant of Negishis
Heyden replication wIth alltogbull optlmizaUon countorfaclUal no ~odjcator inlottempora 1987 paramlJters aocroach
faltonstaln baso yeat back solution actual versus no aggrampgalfld intersElclornl replication wilh exog Merrills counlerfaclutll no indlclltor middotlntamprlempOull1984 19B6
parameters al90rithm 1981middot1982
Gouldar staady~slala Inlar5ecioralbase year back souUon Variant or 1S lronsilion no dynamic CV inlortempora
stoady state parameters Newlon 1985 feplicalion+ with iUog Fair~Taylorl
+slaildy-sla1e Intargeoorat Gtlulder~Summers base year back solution varIant of stlJadymiddotstate fixed no aggregated fntersectoral
vs trarlsition government Indicator Intertempo1al steady state oatameters Newton
repticaUon+ with exog Fair-T aylQf 1967 +ste-adv-stata bullbullbullndlno financial
Jorgenon-Yun econometric steady-stale Money melfic In1ersecloral
- - 8slimalion vs traosUian no social weUare functton intertampotal1984 +~teadlmiddotstatl)
~ shy
qtJalitaUv8 by NPSCt optimal path ye generalized inlersectorai 1986 1987 Perelr
replication 8S$umptlonl optimizatIon V$ lumpmiddotsum dynamic lntertemporat se~~ttilitf ____Jllgorilhm opllmal pah pen inc CVoclV financial
- 9 shy
bull ~ ___ - _______ - ____ bull___ OM bullbull __ bullbull __~____ ~_
--
- - - --------
--------- - - - --------
---
1nltJ1 bull I ~ 1J11ltI1 v I ~
-
Sa Ita rd-Futlertonmiddot intargration consumption lax ~StHlven~Whall~y VAT capila~ or labor tax
Ch~mqflS1985 Andersson
consumpion tax neutral corporals lax system 1987
it ue rb a c h~ i(at II Sltotf COIlSUft1ptian lax (poundpird ItS hibor Mx~on ufocto of taxation on capital fOfrnJilti1
dvnamio fiscal policv effects af OQvernmanl defioit (3tlard-Goulder
1883 1987
allocts 01 the adoption of a 1ge3 1985 pure consumpHon lax
im~rtmce of cons fCreSClht -Bovenborg analytically orianiad approach
1985 1986 effects ot pure conumption tax capllal income taxation in Opo economies
Erllch-Glnburghshy-Heyden reat wage policios in Belgium
1987 Felter1steln 1984 - theorotical
1985 ~ linnncial crcwdlng out1984 1986
Goulder effects of tax- vs bond-financed 1985 chanQes In pubtlc expondllufe
~~--
Gouidr~SUmmampfI torpora1$ lax cvts reduced 1987 ITO increase 10 gasolioe taxes
Jorgenson-Yun
effects ot several programs of tax reform on the allocation of ~~ptai
1984
-- Purelra
corporate la)( Integration effects cf romiddotinlroducinQ imreslamnl tax credits
1986 1967
10 bull
bull -~- ~- --- _ - ---____--_ - -----_-_ --~ ---- ~
1 1
2 Brief overview of non-dynamic CGE models
The origin of most or the models we survey goes back to the work of Scarf (1967 1973) whlgt
first developed a reliablo algorithm to compute equilibrium price for a ArrowmiddotDebreu economy
HiS algorithm used simpicial Bubdivision techniques and can -be shown to be the computational analog
of the fixed point theorems previously used to prove the existence of equilibrium His technique
could soW a model with an arbitrary number of consumers and commodities as long as all agents
were price takers consumers were subject to budget constraints demands were continuous andmiddot
production did not display increasing returns to scale The alQorithm while guaranteed to converge I
I I I
was relatively slow for prcb1ems involving more than S1If 20 dimensions A maior improvement in
computational speed Woo offered by Merrills algorithm (1972) which used tpe same fundamental
ideagt as Scarfs procedure
Scarfs model (as is true for the standard Arrow-Debreu model) does not include a government
sector ~ neither taxes nor public goocs At one of the mosi promising appHciUions of the new
computational techniqu w~s in Iha area of tax policy evaluation Shoven and Whalley (1973)
exlended the general equilibrium model and computational approach to include a wide rry of taxes
and a governl11ampnt spending plan The original ShavenmiddotWhalley model was static (although the
different commodi~ies CQuid be considered simitar goods available at different datest as In
Arrow-Debrau) and government was aumad to run a balanced budget The data are arranged as a
social accounling matrix The models spacification and calibration is checked by solving it in the
presence of the base set of laxes The result should ba exactly tha initial social accounting matrix
After having pasd this replication check the model is solved for a counterlactua equilibrium In
the presence of a new tux design The result is once again a social accounting matrix The two
equilibria are compared in order to tSS($S the impact of the new tax plan The first uses ot this
bull
1 Z
model for lax policy evauajon were ShovenWhalley (1972) Whalley (1975) and Shoven
(1916)
Completely static models as Shaven-Whalleys are unsatisfactory for many tax reform ISsues
These include corporate lax Integration effects of investment tax credits effects of accelerated
depreciations consumer Or e)penditure tzxes~ importance of saving subsidies like IRAs etc These
are essentialy dynamiC ksues They involve not only the allocation of capital across sectQfsbut
perhaps more Important the capital intensity of the economy But the capital accumulation and
capital reallocation take time and may involve adjustment costs Because of these issues Ballardmiddot
Fulierton~Shoven~Whaloy took the first steps towards developing a dynamic model
In this conlexl the ShovenWhalley model has been extended and implemented for Ihe US
-economy In BuIIBrdmiddotFulienonmiddotShovenmiddotWhalley (1985) While Ihis book completely documents the
mo~el it was used in several publicalions beginning in 1978 Their mode) consists of nineteem
production sectors twelvo households and fifteen consumer goods It includes a very detailed set of
taxes including the federal and state personallncome taxes federal and state corporate income taxes
SOCial Security taxes pro1J0rty taxes unemployment Insurance excise taxes sales taxes etc It
has been calibl att-d to reproduce the 1973 US economy Recentty a version corresponding to the
1983 JS economy has been developed
wThe 8alard-Fullenon-Shoven Whalley model is dynamic In the sense that consumers face a
choice between current consumption and leisure versus future consumption (which can be
purchased via s wlngs) The consumer classes act as if they were maximizing a nested CES utility
function over the domain of a CES aggregate of contempo-raneous consumer goods and leisure and
future consumption subject to their income constraint The parameters of those funcHons
determine the shares of income devo1ed to each commodity to saving and to the purchase of leisure
They also determine the two key elasticities in the models ~ the elasticity of labor supply with
bull
13
respecl to tgte rbullbull1 aoer lax wage rae and the elasticity of saving with respecl 10 the real after lax- rate of retum to capital
In the model consumerS have myopic expectations regarding lulure prices and in particular
regarding the future rale of return 10 capilal Ballard (1983) and BallardmiddotGoulder (1985)
inco~ora1e both perfect foresight and limited foresight into this model Future consumption is
acquired by buying a fixed composition portfolio of real investments that offer an infinite annuity
of returns
The production side of Ihe model is compielely static The mod1 Incorporates a constantmiddot
olasicily 01 substitution betweeen prlmary inputs in production (capital and labor) and fixed I coefficients for intermediate inputs The model distinguIshes between Industrial outputs and
eonsumer goods for the simple reason Ihat tile data are classified differenUy Industrial sectol$ I I involve such categories as forestry and fisherle metal mino9 and publishing and printing while
consumers purchas-3 fUfHi1ure 2ulomoblles and books This fact is recognized in the model by Ibull
incorporatirg a secord stage Of production which converts industrial outputs intO consumer goods
This technology is uStally mod9-led as a flxed~coefficient conversion matrix
The BallardmiddotFINton-ShovenWhalley model assumeS that the private sector finance
marginai ioveslment with the same composition of debt and equity as currently exists in each sector
This is the same arsumrtion that Herberger originally used Investors all hold debt and equity in tile
same proportion Thampreiore this ownership can be aggregated into simply capital ownership For
tax purposes however the separate treatment of debt and equity is taken into account at both the
corporate and persofa Io=vel SimllarlYJ the dividend poIicies on corporale equitY are established
exogenously There are no government bonds in the model since there are no government deficits
The modol is solved for a sequence (as many as 100) of temporary equilibria wltll consumers
allocating income between present and future conumption at each point in time The palh for 1he
14
economy ~ a sel of connected equilibria The connectIon is pra~ided by capital accumulation
Capital accumulation is endogenous and determined by saving The model stans with a sociagt
accOunting mi1rix In the bzse case the economy is assumed to be in a steaqy state growth -path (along
which all relative prices ~fe constant) The madel solves for both the naw steady state growth path
and the transltion 10 it zf~er a policy intervention rhe authors have frequently addressed the
question-of howong rt tKes to effectively settle un10 a new steady S1ate growth path
The dynamics of the model are limited however in that future consumption is collapsed into a
composite commodity Aso the absence of government deficits and the lack ofproductlon dynamics
limits the realism of the BallardmiddotFuliertonmiddotShoven-Whatley model for the analysis of dynamic
policy issues (such as the adoption 01 a consumption tax or Ihe elimination of the investment tax
credit) The models we survey try to Improve upon Ihe Ballard-FlIiertonmiddotShaven-Whalley
dynamic modelling The models Jn~orporate more forward lookTng behavior They improve the
dynamics on the conampumrtkm side of the economy and some of them introduce true 1ynamlc behavior
on the part oI the producers We turn now to the issues inY9lv_ed in developing such dynamic models
15 I I ~3 ~ssues in the modelling Of dynamic behavior
We will 100k firt at the speclficalion of economic behavior of consume(s~ producers
goyemment and Ihe r~sl of Ihe world The modemog of corporale and household financial decisions
I is then addr5sed Finally Ihe eencept of equilibrium is discussed
I
i
I 1
31 Consumers behavior
Early efforts 10 build dynamic features into the economic behavior 01 consumers are due to
Ballard (1963) and Auerbach-Kotllkoff (1983 1984) Their work has been closely followed by
several subsequent authors In faet dynamic household behavior Is the single most pervasive aspect
amohg the models surveyed
All the models Inc-O1orale some loro of lila-cycle behavior Household behavior is dewmloed
by the maximization of an addWVely separable time invariant intertemporal utillty function The
utility lune1ion is defined oyer the domain of the consumption goods In the economy In most of the
models leisure is also an argument in utility so that labor supply is optimally determined
Utility maximi4ation is subject to a lifetime intertemporal budget constraint which equalizes
the presen1 value of consumers income and expenditure More reeentiy in Andersson (1987) bull
Bovenberg (1985) and Pereira (1986b 1987d) the constraint i$ defined as a sequence 01
recursive equations of motion on wealth This has the potential advantage 01 accomodating liquidity
constraints However It should be recognized that in the absence 01 liquidity constraints (ie when
consumers are free to completely borrow against future income) the two specificallons of the
househofd ccnstraint are essentiay equivatent Furthermore tn both versions saving is optimally
determined as a way of translaring wealth intertemporally
1 6
Some lodels now have a sophisticated dynamic specification of hoosehokl In Ballard (1983) -
BallardmiddotGoulder (1985) and Auerbach-Kotlikof (1983 1984 1987) consumers behavio is
embedded in an iterQEneraional setting Each of them has S5 age cohorts simulataneously alive
Som-e measure of jntergener1ional altruism is considered in the form of bequest motives In Ballard
(1983) and Ballard-Gouldr (1985) consumers derive utility directly from bequthing parl of
their wealth Therefore bequests assume 1he form of a scrap value of terminal wealth of a
generation In AuerbachmiddotKotlikof (1983 1984 1987) each generation empathyz bullbull with Ihe
utility of future generations ovar bequethd wealth In addition Ballard (1ge3) includes detailed
demographic poiections wfill the intent of incorporating in the policy analysis the effects of the
bull Post World War tI baby-boom
In the real world household decisions also include the optimat allo~atiOn of saving among
shyalternative physical or financial assets Theories of household portfolio behavior are relatively
Iwen~establshed However they involve uncertainty as a crucial element Accordingly they have
middotnot been incorpora~ed in the context of the determinIstic dynamic models we sutyey here We will 1 Come backt-O this Issue below
32 Producer behavior
The efforts to build dynamic features into the economic behavior of producers are more recellt
and less widely adopted The first attempts are due to Bovnberg (1984 1985) and Summers
(1965) Dynamic behavior has been more fully incorporated In the recent models of Andersson
(1987) Auerbach-Kotlikoff (19B7) GouldermiddotSummers (1987) and Pereira (1986b 1987d)
Part of the reason why production side dinamlcs has been more slowty adopted is the weak
supply of accepted theories referring to the dynamic behavior of the firms In the models referred
I 17
to above gynamlc producion and investment b~havior are middotinduced by the existence of capital
adjustment costs and linked to Tobins q theory Adjuslment costs are designed to capture both the
incomplete mobility of capital aoross industries and installation costs ~e the costs of adjusting
capital towards its cpHmai level)
Adjllstment costs can be conceived as internal to the firm and measured in terms of foregone
outpul along the lines of Lucas (1967) Alternatively adjustment costs can be viewed as actual
external cests incurred together with Ihe purchase costs along the Unes of Gould (1969) With the
exception of Pereira (198Gb 1937d) all of the CGE authors follow the former approach
The firms in the economy maximize their market value as the present discountad value of the
future stream of dividends In Andersson (1981) and Pereira (198Gb 1987d) firms are seen as
maximizing the present discounted value of net cash flow Maximization is constrained by Ihe
adjustmenl cosl lechnology and ~n equation of motion describing Ihe evolution of the capital stock
It should be noted that undereogenous divldendrelention rules the two problems are equivalent II
should also be nOled Ihat a salisfectory economic rational for th bullbullistoo of dMdonds with the
present tax code is missing in the profession (Shoven (1986))
The models with static formulation of producers behavior are characterized by passive
investment behavior Investment merely accomodates to saving in the eccnomy With a dynamic
formulation induced b adjustment costs real investmen decisions are forward looking Investmenl
Is endogenously and optimally determined by the firms A fundamental difference between the
shonmiddotrun in which capital stock is given and longmiddotrun in which Ihe level of capital is allowed to be
optimally determined is emphasized
This extra richness of produc1ion dynamics is not without costs A careful look at the summary
tables will clearly show an inverse relation between the adoptlon of dynamic features in production
and the level of disagreggation of the production side of the economy In fact with production
bull
j
1 6
dynamics $e dimension of the problems is immensly increased Let IS be more specific
In a static framework with constant returns to scale production technology the output level ismiddot
indeterminate The optimal aliocation of inputs can be obtained by cost minimization with any
feasible output level generatillg zero profits Such zero profit conditions are used to solve for the
output prices in terms of the factor prices and hencemiddot to reduce the dlmensionality of the problem
from the -number of commodities and factors to the number of factors Thus even if the model deals
with 30 production sectors the computation of an economic equilibrium can take place using only
the dimensionality of the primary inputs in the economy
Now under certain regularity conditions on the production and adjustment costs technologies
leading to enough concavity of the optimality objective the intertemporal output path for the firm is
endogenously optimally and uniquely determined even with constant relums to scale technologies
(see Pereira (1985 1987a) Accordin~y with adjustment costs in the model optimal profits
will in general be non-zero This result has crucial implications for the computatiQn of equilibrium
in the model~ with production dynamics With adjustment costs no reduction of dimensionality is
possible The curse of dimensionality returns as a binding constraint
The introduction of adjustment costs adds another significant complication to the model With
capital being less that perfectly mobile in the economy different rates of return on capital will
exist in different sectors This is a difficult problem to tackle conceptually in the absence of
uncertainty Also in the real world producer maximization choices include also its choice of
financial ratios (debVequity) and payout rates (dividendretained earnings) These financial
subjects are difficult and much Clf this behavior is still taken as exogenous by the modellers We
will come back to these issues below
middot i
19
bull 33 government behavior
middot
Two issues dominate the modelling of government behavIor First govemment behavior has
Ibeen typically seen in Ihe CGE liIerature for lax poliCY evaluation as constrained by yearly balanced
I budsets (see for example 8allard-Fullrton-S~ovn-Whafley 098Sraquo The analysis of
govemment defidlS and public debl in a CelE context requires a dynamic selling Second the level of
government expenditures is either exegenouty given as in Auerbach-KoUikofi (1984 1987)
6ovenbrg (1984 1985) Feltenstein (1984 1986) and Jorgenson-Yun (1984) or
endogenously (but no optimally) determined by the balanced budget conditions as in
6allard-Fullerton-Shoven-Whalley (1985) Andersson (1987) Erllch-Ginshurgh-Heyden
~
(1987) Gouder (19aS) and Gouder-Summers (1987) In the second case the composition of
bull public expenditures is often optimally determined However the leve~ of government expendi1ures
can only 00 endogenously and optimally determined if the government is seen as an optimizing agent
and is allowed to run deficits
The first attempts to deal wiU the government defiCits in CGE tax models are due to
Auerbach-Kollilltoff (1984) and Feensain (1984) In Auerbach-Kotlikof (1984) government
expenditures afe exogenous They grow at the rate of growth of population However given 1he tax
structure and tax revenues yearly deficits and surpluses are atTowed subject to an in1ertempond
constraint that the present value of future tax revenues equals the present value of future
expenditures
In Fellenstein (1984 1986) government expenditures are also exogenously given He also
allows the government to run deficits to finance expenditures in excess to 1ax revenues However
surpluses are returned to CQnsumars in the form cf transfers Accordingly government ts not
subject to any constraint regarding the future repayment of public debt
In GouJder (19851 the 9Vernmenl maximizes a static social welfare ~nctio subject to a-
bull
20 I i
I
I I I
I bull
balanced budget conSirtlL This follows the optimal alocatkm of government expenditures alpng
the lines of Baliard-FHeiton-Shoven-Whalley (1985) Ttle model Is generallzea to allow for
exogenous changes in HIe time path of government expenditures Two financing alternatives are
consjdeft~d and contr8s1ed EldditionaJ tax revenues and bond issuance
Pereira (198Gb 1137d) auempts to address both the incorporation of deficits and the
determination of govement expenditures The path of government expenditures and the path of
delicilSlsurpluses (ano therefore the path lor debt) are endogenously and optimally determined The
government is seen as mltJximizing n intertemporal social welfare function given the 13x structure
Optimization is subject to a sequence of recursive equations of motion reflecting the evolution of the
ptlbllc debt allowing fDr government budget imbalances It should be stressathat this specification
01 the government cOMtcalnt is equivalent to the specificaUon in Auerbaeh-Kotlikoff (1983 1987)
when no liquidity constraints affect -gove~nment behavior
The extra richness in the treatment of government behavior allows tfte examination of
government debt poices in a truly dynamic setting Also financial crowding out effects induced by
government deficits can be analyzed (see Auerbach-Kottlikoff (1987) Feltenstein (1986) and
Perer (198Gb 1987e))
34 Foreign sector
Most of the open economy models follow the assumptions of balanced trade with import and
export net demands characterized by constant elasticities along the line of Ballardmiddot
-Fullerton-Shaven-Whalley (1965) Such is the case of Salfard (1963) Ballard-Goulder
(19S5) and Goulder-Sllmmers (1987)_ In Feltenstin (1985) the rest of the world is treated as
21
an addjlion~ consumer group
Bovcoberg (1986) develops a model In which two economies lIr considered each following
intertempoml perlect foresight paths These economies meet in the international forum Their
trade relaroMhips are characterized by yearly balanced trade accounts
None of the new generation of dynamic cae tax models has yet IncorPorated the lnternational
capital flows as done in aoulder-Shoven-Whalley (1983) for the earlier Ballarrshy
-Fullertonmiddot Shaven-Whalley (1985) model This is unfortunate as there is an important
Intertmporal aspect to international lending The first al1empts along Ihese lines are due to
Andersson (1987) and Erlich-Glnsburgh-Heyden (1987) Andersson (1987) in his model of the
Swedish economy adopts a closed economy appraoch in which rates of return in the domesJic
economy are largely determined by the international capital markets Fidinterest rate induce
Intlnatlonal capital flows which determine and finance the International lrade imbalance In
lurn In ErlichGinsburghHeydenmiddot (1987) foreign trade Is generaled according to an
intertemporal Irade welfare lunction with constant import and export elasticities In the shortmiddotrun
they allow inlernational trade inbalances which generate capilal ftows to the domestic households In
the long run however trade balance is assumed
35 The need for financial markets
A dynamic economic Wucture not only provides the ideal environment 10 model many features of
economic behavior it also permits one to Incorporate the financial side of the economy If the
government is allowed to run deficits the question 01 deficll financing automatically follows If
investment is optimally dewmined and returns to capital are different aClOSS sectors problems of
investment financing arise If there are seeral final1cfal assets In the economymiddot government
4-
I i
ibull
I I I I
I
j
22
bonds privpte bonds and equity (or simply physical capital installed inmiddot different sectors) the
problem of allocation of saving among assets with potentially different returns arises
The papers surveyed var greatly with respect to l~e extent of their attention to the financial
side of the economy At Clne extreme are the models in ~anardmiddotFuliertonmiddotShoven-Whalley (1985)
Andersson (1987) Ballltld (1983) Ballard-Goulder (1985) Bovenberg (1985 1986) and
Erlich-Gjnsburgh-Heyden C19B which are devoted exclusively to the real side of the economy
In turn Auerbach-KotliKoff (1984 1987) Feltenstein (1984 1986) and Goulde (1985)
allow for government debt In these models saving finances changes in government debt and physical
capital Private and publlc 2ssets are perceived by the households as perfect substitutes The
allocation of saving merely adjusts to the relative demands for funds
Feltenstein (1984 1986) is the only model surveyed which introduces money Government deficits are financed by issuing money and bonds according to an exogenously given rule Money is
demanded by consumers for transaction motives and an exogenously given fraction as a store of value
On the other hand government bonds and physical capital are the vehicles for the intertemporal
transfer of wealth
Summers (1985) and GOlldermiddotSummers (1987) introduce a whole menu of financial assets shy
firm specific equity capital Different assets earn different rates of retum However such rates
are equal up to constant and exogenous sector speCific risk premia Therefore the introduction of
constant exogenous risk premia as helpful as it may be in the context of calibration does not solve
the main issue of the non-optimality of the allocation of saving Also talking about risk premia in a
deterministic context is somewhat unsatisfactory
Pereira (1986b 1987d) also introduces a whole menu of assets private and public bonds and
firm specific equity However all the assets are expected to yield the same rate of return and
therefore perc~jved as perfect substitutes The different asset types allow consideration of
23
exogenOU$~8tUequjty triO dividendlrc~enlion rufes and therefore seve-ral sources of investment
financing bond equity and retained earnings
The nonmiddotcptimalHy of the allocation of saving and the absence or exogenelty of corporate
financial rules is _a reftectlon of the limitations of the determInistic approach Either in a context of
perfect anticipation oi the future prices Of in general within the t$atm of point price
expectations all the papers follow a deterministic approach Under such circumstances consumers
either expect diHerent rales of return (inclusive of risk premium) across assets in which case
they wilt buy only one asset (that with highest rate) Of they expect equal rat~s of relurn in which
case Ihey are indifferent about the asset composition of Iheir portfoliO There is no way of
Iradingmiddotoff rales of return and risks to obtain an optimal interiof solution to the problem of the
allocation of saving
The most advanced contribution in the modelling of saving allocation in a CGE setting is due to
Slemrod (1980 1983) in the context of a static one-period model In his model consumers act
according to a two stage separable decision process They Hrsl decide on how much to save Then
they decide on the allocation of saving according to an Indirect uttlity function dependent on the rates
of ratum and variances offered by the different assets in the eoonomy The sourCe of riskiness in the
ampConory comes from an uncertain marginal product of capital On the other hand aside from
portfolio decisions the res of the economy is insulated from uncertainty
The most complete contribution in terms of the treatment of the corporate financial rules is
FulienonmiddotGordon (1983) In a variant 01 Ihe BailardmiddotFulienon-ShovenmiddotWhelley (1985) model
Fullerton and Gordon have capital intensity and optimal financial decisions Jointly deter~ined
through a two siage process The cosl of financing capital is minimized by trading off the tax
advantages of debt against the epected real bankrupcy coSls inherenl with high debVequity ratios
Given the optimal debtequity ratio the level of inveSlment Is chosen such that at Ihe margin the
24 l
bull
rerurn pn ~uity equas the return on bonds plus an exogenous risk premium
36 Market assumptions equilibrium and expectations
Virtually all of the papers arc characterized by Walrasian market clearing assumptions All
markets are perlect1y competi1ive Atomistic competition among agents 1s assumed even thoughmiddot only
a flnile number of agents is considered Virtually no mark~t disequilibria or price stickiness afe
considered The only exception is ErlichmiddotGinsburghmiddotHeyden (1987) bull In thei model of the Belgium
economy I the _wage rate is fixed in the short-run Therefore in the shOrlwrun desequilibrium in
the labor market will generate endogenous unemployment However in lh$ long-run all prices
Including the wage rata are Ilexible and accordingly ali markets ciear
Given the dynamic nature of behavlor in the economy markelmiddotcfearing prices in each period
depend on expectations of fture prices and on tax variables In the economy There are essentially
two ways at interpreting the economic equilibrium in such a dynamic conte~bull If future prices are
perfectly anticipated (io expectations are self-fulfilling) a perfect foresight equilibrium
prevaHs Then future actions are merely the implementation of current decisions for future
periods However jf ptica expecla110ns are not perfect (Le agents make mistakes w1th re$pect 10
future prices) then a temporary or shortrun equilibrium prevails Markets ciear and clearing
prices depend on fu1ure price expectations Current plans about the future are typically not
precisely implemented They will be revised as more or better information becomes available to the
economic agents See Grandmant (1982) for a comprehensive survey 01 the temporary equilibrium
literature
Wilh the exception of Gould (1985) and Pereira (19S6b 19S7d) all the models surveyed
adopt th~ concept of a perfect foresight equilibrium In turn BaliardmiddotGouldr (1965) considers a
J
2S
flexible am9unt of foresight In terms of the number of years over which price movements are ~
foreseen Pereiras model (1986b 1987b)) is flexible in a somewhat different way in that it can
indude any range of foresight from myopia to perfect foresight The choice between the perfct foresight and lemporary equilibrium is ultimately to be made on
philosophic grounds It can be argued that less than perfect expectations Imply that agents are
irralional in some way (see Auerbach-Kctiikoff (1987 p 10) However the reverse argument can
be made One can question wiether agents are reaHy rationar and perfectly knowledgable about
future prices
Recent evidene of Balird (1987) Ballard-Goulder (19851 Goulder (19851 and Pereira
(1986b 1987d) suggests that the choice in modemng expectations is an important one They show
that the degree of foresight inl0 thc future (ranging from perfect foresight to-myopic expectations)
may have dramatic impacts on the pollcy conclusions of Ihe model Accordingly the best research
strategy may b to design models which are flexible enough 10 allow for different rules regarding the
formation of expectations With the exception of the articles just mentioned sensitivity analysis
bullhave previously not been performed along this dimension
In terms of implementation the two concepts of equilibrium - perfect foresight and temporary
equilibrium have different implications The dimensionality of the equ1ibrium soll1ion algorIthm
is involved Suppose we have a model with ten markets to be run for a period of 50 years Aside
from normalization a perfect foresight model implies computing prices In 500 dimensions while a
-t~mporary equilibrium model requires sorving SO equilibria each in ten dimensions Given that
computational speed often varies with the cube of the number of dimensions the lemporary
equilibrium formulation is potentially strikingly more feasible However Ballard (1987) and
Goulder-Ballard (1985) have devoloped techniques to greatly speed the computation of a perfect
foresight equilibrium
26
j
I
4 Jrnplemenlalion and Issues on policy evaluallon
41 Calibration and equilibrium comparlslon
CGE models at typically parameterized by the use of a calibration procedure Some parameters
are exogenously given However some crucial parameters are determined in such a way that the
model replicates the data for a given base year See MansurmiddotWhalleY 11984) for an eXlensive
discussion of 1~i$ issue
CalibratIon in a dynamic context is generaliy interpreted as requiring two properties First
replication of base year data is required Second the model is parameterized to simulate an
intertemporal balanced growth path when Ihe base policy is maintaineq This Is the approach
followed by Ballard (1983) BallardmiddotGoulder (1985) Goulder (1985) and SummersmiddotGoulde
(1986) It follows the practice of BallardmiddotFuliertonmiddotShovenmiddotWhalley (1985) with their
sequential cquilibrium dynamics
Other authors Auerbach-Kotlikof (1983 1984 1987) Boyenberg (1984 1985 1986)
and Pereira (198Gb 1987d) follow whal we cali a qualitalive calibration The slructural
palametars are exogenously chosen so that lh economy follows a reasonable path into Ihe fulure
ThaI has 10 do wllh the fact thaI given the recursive nature 01 the dynamic economy and aside tom
such structural parameters only iniUa stock values are needed to run these models Given initial
conditions on the stocks of say private wealth capltal and government debt agents will optimize
and Ihereby generate a lirst round of net demands and equilibrium conditions In lurn the
equilibrium prices will determine the evolution of the stock variables into the next perIod
There are several potential problems with calibration in the conlext of dynamic models The
assumption of a steadymiddotstate growth ~ath in the base case can be questioned First while
27 I Ibullbull
steadymiddotstaU is a possibity it certainly is not the only meaningfull solution to dynamic models
Even in the case of a perfect foresight equilibrium the model implies an equilibrium path which i may or may not involve colanced growth The model not the modeller should dictate the nature 01 I themiddotbase case path Second in the context of a temporarY equilibrium path a steady state solutlcn is Inol a likely model QutcOn1tl In fact unlessmiddot expectations are stattc short run behavior consistent
with a steady-state evolution will in general not be generated On the other hand if static I expectations are self~fufjili[1g we have in fact a perfect foresight model Third even the base year
replication requirement may cause problems In Ihe cOntexl of temporary equilibrium Any
calibration parameter would be conditional on expectation rules which is probably an undesirable
feature
The nalur of the two-requirement calibration strategy is very muchin the spirit of the bull
traditional design of the comparision of alternative equilibria comparis_ion between a steady-state
base case on one hand and alternative paths including a transition period and a finat sleadyyenstate on
the other hand Pereira (198Gb 1987d) compare different (not necessarily steadystate)
equilibrium paths The arguments against such a procedure are based on the idea that the impact of
policy changes can be observed most easily since all departures from the steadymiddotstate can be
attributed to the altemalive polioy On the other hand the base case so dEfined as a steadymiddotstate is
consist~nl with previous work in a less dynamtc setting and Iherefore allows a common standard
for comparing model results
42 Computation algorithms
We are still al a stage in which basically each author uses a different computational technique
Th development of dynamic models - in parlicular wilh adjustment costs andior perfect foresight
bull
28
has corresponded with the decline in the use of fixed point algorithms In -fact given the relative
arge dimensions inevitably involved such algorithms lend to be very inefficient at the best and
chen prohibitively stow See Stone (1985) and Preckel (1965) for a comparative assessment of
different computation techniques Among the models surveyed only Ball1rdFulle(ton~
-ShavenmiddotWhalley (1985) and Feltensteln 1985 use Merrills variant of the fixed point
algorithm 1echnique
AuerbachmiddotKoHikoff (1963 1984 1967) follow a three stage procedure They first compute a
base case steady-state then a revised case steady state and finally transition path for the economy
berveen these two steady-states tn all stages a Gauss-Seidel iterative procedure is used
Ballard (1982) Ballard-Goulder (1965) Goulder (HiSS) and Summers-Goulder (1986)
~
use a method developed by Ballard and Goutder which is similar in many awects to the FairmiddotTaylQr bull
(1983) algorl1hm Short-run equilibria are calculated (using Merrills algorithm) parametric on
nrice expeC~lions The model is then iterated to generate self-fulfilling intertemporaf expectations
and the corresponding perfect foresight equilibrium In a relatively similar approach Andersson
1987 uses a simulation program SIMNON developed In the University of Lund This program
can handle two-paint boundary problems In a fashion consistent with the multiple shooting
algcrlthm (see Lipton-Poterba-Sachs-Eummers (1982))
Boyenbergs computational approach (1985 1985) differs from the other models surveyed in
that he relies heavily on analytical techniques Computations are done by using a dynamic version of
Johansons linearization method Being essentially determined by the continuous time nature of the
I I
f
model thiS linearization model has the disadvantage of confining the analysls to infinltesinal changes
around the base case equilibrium (see Bovenberg (1985) p 53) Pereira (1986b 1987d) uses an optimization algorithm NPSOL - developped by
rGillMurray-Saunders-Wrlght (1986)) This algorithm is used 10 compute the sequence of
I
29
Shortmiddotrun temporary equilibrium which makes up the intertemporal equilibrium path The
equilibrium conditions are seen as nonlinear equality constraints in the minimization of an artlcial
objective function The prices are normalized to the unit simplex by an additional linear equality I I 1 bull constraint
II ErlichmiddotGinsburgh-Heyden (1987) follow a unique approach in that they use a variant of the
optimization technique introduced by Negishl (1960) The economic equilibrium can be generated i
as a solution of a mathemallcal program Ihe objective function of which is a weighted sum of the
utility functions of the various agents while tho constraints set consists of the market dearing 1
conditions Ginsburgh-Heyden (1985) have extended Negishis resulllO tho case of downward price
lrigidities I
Finally the paper by Jorgonson-Yun (1984) is also unique in tIlat it is the only I
econometrically estimated model Different blocks for the consumption and production Side of the
-
model are separately estimated to provide the necessary structural parameters i
I The diversity of computation techniques is yet another indicator of the exploratory nature of the
body of literature surveyed in this article
43 Equal yield comparisions
The link between a base case and counteriactual simulations Is usually provided by the concept of
equal yield Shaven-Whaley (1977) discuss the meaning 01 equal yield in a general equilibrium
context when government is not allowed to run defICits Equal yield is interpreted to moan constant
public utility Government base case utility is maintained in the counferfactual experiments With
balanced budgets the concepl of equal yield is unambiguous The new equUibrium prices and the
balanced budge condition will del ermine the minimal expendilure and taxes needed to maintain base
30
case publi~ utility Accordingly in general equal yield is inconsistent with equal nominal tax
revenue Some change in tax revenue Is necessary Different tax replacement schemes ate
considered to assure that enough tax revenue is collected This is the approach essentially followed
by most 01 the papers surveyed Only Feltenstein (1985) Goulder (198) and Jorgenson-Yun
(1984) chose not to follow an equal yield strategy
The question is of how to Interprete lila concept of equal yield when the government is alowed to
fUn deflcLts The optimal leve of expenditure for base ease public trliUty can now be taAinanced
bond financed or financed by a mix of bonds and taxation We have several versions of equal yield In
particular equal yield may now be consistent with equal tax revenUE Furthermore some meaSure
ltgtf financial crowding out effects of government deficns can be inferred from the comparislon of the
several equal yield alternatives These issues are extensively discussed ~Pereira (1987b)
44 Price expectations and the dynamic generalization of compensation
Indicators
The sum of equivalent and compensation variations over households is the most widely used
aggregated measure of efficiency gains or losses In the context of $teady~state eomparisions andor
when complete future markets exist andlor perfect foresighl is assumed there are no difficulties
associated with the use of the standard Hicksian indicators In fact correct future prices are known
in these cases
If expectations are not selfmiddot fulfilling andlor future markels are not available a dynamic
generalization is necessary 8allardmiddotGoulder (1985) provide some steps in that direction by
defining an indicator that accomodates periods far into the future when households do not have
perfect foresight but a steady~state prevails On he other hand this issue is more fundamental in
i
[
I
i I I I I [
I I
bullI
31
the contex~ of a Gner1 lemporary equilibrium ftamawor~ In such Circumstances Pereira
(1986a) dev~lops a cnamlc generalization which is obtained as the present discounted value of a i i
sequence of short-fUn optimal expenditure functions consistent with a base case expected future I stream of IlWities I
32
5 Empirical evidenee from selected policy Issues
Dynamic tax models have generated several important results which escaped Sialic modellers
The following is a seleCpoundd set of issuos whIch stress the marginal benefits of dynam~ modelHng of
economic behavior in innovltlttive areas of analysis The discussion of a consumption tax emphasizes
the benefits of dynamic tusehold behavior and intergenerational aspects The study of the
flliminatlon or the re~intToduction of the investsnemt tax credits is made meaningful by the dynamic
modelling of production betlavor Modelling governmenl in a dynamic conlexts allows the modeller
to study the irrrpact of government deflcits Finally a dynamic framework lets us appreCiate the I
relevance of consumer expectations in terms of the evaluation of policy alternatives
i
i
51 Consumption Tax
t
By Its very nature a consumption tax can not be adequately investigated with a static model bull
Fullerion-Shoven-Whaicy (1983) use Ihe model of the US economy s described in
Ballard-Fullerton-Shoven-Whalley (1935) to evaluate the movement from the currenl US tax Isys1em to a progressive consumption tax Since their modeJ Incorporates a Jaborleisure choice
where leisure is an untaxed commodlty their results reflect the fact that both the consumption tax
and Ihe present system are dslortionary
Concentrating on the rntertemporaf distortions they show that sheltering more saving from lhe
current US income tax could improve econornic efficiency even if marginal tax rate increases are
necessary in order to maintain government fevenue~ At first you have a revenue shortfall
However the economy moves to a higher sustatned growth path and uitimatefy lower tax rates
cangenerate the same revenue path Also wages increase in the long run with this policy The
bull
bullbull ~
33 shyt
i
I
I
present value of welfare gi[lS for a potiey of complete savin9 deduction wilh marginal tate
adjustments (consumption lax) is simulated to be around $500 billion to $600 billion 01 1973
dollars
They also investiga1e th 1ngth of ttme it takes th~ economy to adjust to these policy changes
Roughly they estimate the long run to b thirty years although this figure Is very sensitive to the
specification of the savings elasticity - a crucial parameler in this model
52 Investment T~x Credit
Goulder-$ummers (1 S87) address the impact of eliminating the lnvestmenl Tax Cradit (ITC) on
intersectoral capital formation and on economic growth Their model isYa1ticularly adequate to
addies$ this issu~ in that is postulates a forward looking investment m9we with adjustment costs
They show that the eliminating the ITC causes a reduction in the rale of investment Inveslment is
estimated to fall by about 7 in the short-run and by about 12 in Ihe new long run steady state
In lurn a previous policy anncuncement lowers Ihe overall al1ractiveness of Investment and leads 10
a downward shift of the Investrrent profile On the other hand the combined effect of a revenue
neutral simultaneous efimiraron of the ITC and reduction of the corporate tax rates is a long run
reduction ef the capital Sieck by 35 This pattern suggests that a revenue neutral increament in
both Ihe corporale tax rates and the inveSlment lax credits would be preferable 10 the formula
adopted in Ihe US Tax Reform Act of 1986
Pereira (1987d) addresses the impact of the ITe from Ihe oland point of Ihe curTenllax syslem
under the Tax Reform Act of 1986 Given Ihe concern over the potential depressiVe effects upon
savings and investment of the new tax system in conjunction with deficit reduC1ion mechanisms of
the GrammmiddotRudman type a pertinent question is should Ih lTC be e-introduced In the context
bull
34
-of his mod~1 of the US economy Pereira focuses on the trade-off between the distortion in the
intersectoral allocation of capital induced by the lTC the lower tax revenues it generates and
potential financial crowding-out effects on one hand and the positive effects of lowering the relative
price of new capital goods on the other hand Preliminary results confirm the qualitative results in
Goulder-Summers (1987) suggestting a welfare gain of about 2 of the present value of GNP in
the best scenario of absence of replacement taxes Furthermore preliminary results show an
important time pattern to welfare gains with the average benefits increasing the further you look
into the future This pattern is due to the presence of constraints to intersectoral mobility of capital
and inertia in the adjustment towards the optimal capital levels as reflected by the presence of
adjustment costs
I
I
bullAuerbach-Katlikaff (1gB) in the context of their intergenerational model of the US economy
consider the impact on savings capital formation and interest rates of deficit policies and balanced
budget increases in government consumption They conclude that deficit finance and government
consumption can significa-ntly crowd out capital formation and lower the welfare of future
generations However crowding out from deficit finance is a very slow process because it results
from increased government spending over potentially long horizons Also deficit policies may
substantially influence the longmiddotrun interest rates while leaving the short-run rates essentially
unaffected Finally and in opposition to the central role adjustment costs seem to play in both
53 Financial crowding outeff~cts of government deficits
Goulder-Summers and Pereiras models Auerbach-Kotlikoff suggest that the time path of interest
rates induced by a policy of deficit financing seems to be insensitive to the presence of adjustment
costs
bull
35
54 The role 01 consumer expectatlons
The expectations analysis has uncovered one of the benefits of dynamic modelling
Ballard-Goulder (1985) find that the appeal of adopting a consumption tax depends on the level of
foresight possessed by the consumers Furthermore additional foresight may be welfare worsening
This is a second best type of result This tends 10 occur under policies that lead to capital deepening
and declining rate of return to capital over time To the extent that consumers have more foresight
they will be beller equippod to anticipate the fall in the rental price of capital Consequently if a
saving incentive is enacted people save more with myopia than with perfect foresight Given the
exi~tence of taxes on capita and the discrepancy between the private an~--sccial returns to capital
the greater savings with myopia is beller socially Ballard-Goulder find that the welfare gains
from a consumption tax are reduced by about 10 when we move from myopia to farleet foresight
Therefore the attractivenness of adopting a consumption tax seems fairly [obust across these
foresight specifications
I
bull
36
6 Tbe power and weaknesses 01 dynamic modelling the example 01 corporale
tax Integrallon
The Issue of corporate tax integration allows us to sh~w the stre-ngths and weaknesses or the
dynamic approach Empirical evidence using CGE techniques indicates that depending on the precise
scheme of integration and the 1ax replacement methods integratiOn may have substantial effects In
Ihe work of FuliertonmiddotKingmiddotShoven-Whalley (1980 1981) total integration was found to yield an bull
annual static eHielency gain of $4 to $8 billion in 1973 dollars Simulated dynamic gains may be as
large as $695 billion or abcu 14 of the present value of tulUre consumption and leisure in the
US economy These gains result primarily from interindustry reallocalions of Investment and an
improved inter~emporal allocation of consumption
Mora recent work emphasizes hOw consumers asset portfolio decisions and firms finanCial
decisions affect the efficiency gains from integration Slemrod (1980) focusing on consumers
asset ponfoio deCisions finds static efficiency gains which are about twice -as Iarge as those
repOrted iiy Fulierton-King-ShovenWhalley (1961) FuUrlon-Gurdon (1983) focus on the
firms financial decisions They roporl efficiency gains of 6 of GNP from the elimination of the
tax distor1ions favoring debt However when they eliminate the corporate tax and repiace it with
increaSed persona income taxes addlHonal dls10t1ions are created in the optimal labot~leisure
decisions These distortions tend to dominate the analysis slJ(h that the overall effects of complete
integration are very modest Galpr-LuckmiddotTodr (1986) address simultaneously consumers
asset pOrlfollo decisions and the firms financial decisions They also find very modest efficiency
gains from integralion
The above results are important but may be severely biased There are severa aspects of
economic behavior and modelling crucial for the study of Income tax Integration whiCh have no been
i 37
captured innny of the above papers One aspect is the absence of government deficitsmiddot a balanced
budget is assumed It is true Ihal resource crowding OUI is caplured In these mOdels but financial
crowding out induced by government spending is nOI Second investment is not derived from
optimtztlcn behavior Investment behavior passively accomodates endogenous saving decisions As
a consequence the differential impacts of policies in the incentives to save and to invest are not
captured Aso full capital mobility across secors is assumed with instantaneous capital
adjustmenlS towards optimal levels This assumption rules qut different costs of capital across
seclors and therefore differentiated reactions to tax policies changes Finallyhe modelling of both
gov9rnmerlt deficits and of endogenous real and financial investment decisions necessitates the
~nsideraiion of a dynamilt framework and the introduction of financiaJ assets govemment bonds and
iprivate financial assets A dynamic framework also highlights the efficiency effects of Integration on r
ithe ptimal intertemporal decisions The above models are either sIalic (Slemrods and
GalpermiddotltJckemiddotTode(sj or a dynamic sequ~nce of otherwise Slatic model bullbull I Pereir a (1986b) develops a dynamic applied general equilibrium model bull to study the tbull
efficiency effects of integration on the growth and allocation of investmenl across sectors Special
Iattention is paid to the structural affeelS of resource and financial crowding oul induced by
gove mect spending and deficits and to the real and financial Investment reactions across industries
to both tax integration and flnancial crowding out His mode departs from previous work on income
lax Integraticn and for that matler from most 01 the CGIE literalure for tax policy evaluation in
several direcllons It encompasses an endogenous sequential equilibrium struelure founded on
dynamic behavior with flexible expectations Government deficits are optlmaUy determined
Investment decisions are tward looking and the resull of optimizing behavior Several financial
assetsmiddot public and privatemiddot are considered
Simulation resulls Indicale that the welfare gains from Integration under large deficits are at
32
best lJlodest when measured in terms of the GNP The elimination of the corporate tax and its I replacemenl by increased income lax rates yields long run benefits which are never larger than
2 of the presenl yalue ct telr consumption and leisure onder the previous tax regime and 1
under the current tax regime However the short run w~lare effects of full integration tend 10 be
very low and in in some scenarios even negative This is a flew intertemporar pattern of efficiency Ieffects which reflects an adjJsiment lag in the interindustry investment decisions due to the
existence of costs of adjustment On the of her hand partial integration achieved by excluding I dividends from the corporale lax base syslematically yields negalive efficlencyeffeclS In Ihe long Inm these can be as high as 3 of present value of the future value of consumption and leisure This
is a new second best effect suggesllng Ihat less than complete integration may have perverse I efficiency effects
The dynamic slructure of PerclrBs model snowing for an improved freatment of real investment
declsions and government deficts provides a potential setting for a more accurate measure of the
costs and benefits of inlegration The simulation resullS suggest lower benefits and an intertempora Ibull
pattern of growing benefits
Simulation results also clearly suggest robust negative effects from partial integration How
reliable is this result If dJvidends were deductible from the corporate lax base (as are interest
payments) the preferenlial Ireatmant of debl over equily would be ellminaled Corporallons should
be expecled 10 react by decreasing the optimal deWeqully rallo However corporale financial
gecisions are exogenously set In Pereiras model as in aft of the other models surveyed that go as far
as dealing with the issue
Also withdrawing dividends from Ihe corporate tax base is a way of eliminating Ihe double
taxation of dividends at belh Ihe corporate and personnal Income levels How will the optimal
aliocellon of household saving accomodate that change in Ihe relalive return to corporale equity
39 I
t However hOusehold portfolio decisions are exogenouly set in Pereiras model as in all of the other
models surveyeltj go as far as dealing with the issue
IIt is sate to say that the evaluation 01 the benefits 01 partial Integration as defined above are
sevrey underestima1eO Meanwhile the distortionSoenerated by the increase in the marginal
personal tax rates designed to ralse the revenue foregone by the dividend exclusion are fully
aCCltlunted for A good measure of the costs of integration together With a poor evaluation of the
benefils may very well be the reason why partial integration is Simulated 10 yield neg alive
eHikiency gains
On a different vein as most of Ihe model surveyed here Pereiras is a closed economy model It
is legitimate to wonder how the resullS would change If international capital flows were anowed
IThis almost certainly would affect financial crowding out since a good d~al of the capital inflows
could be used to finance public debt
For these various reasons the results should be treated as preliminary but they strongly I suggosl that the dynamic structure provides valuable new insights into the corporate tax integration SSU2
40
7 Concluding remarks I I I i
What has been acccmplished Ihis far The success of Ihe CGE research in laelding the challenge
of dynamic modelling can not be denied
I iGreat progress in the modelling of household behavior has been achieved Promising
developm~nts in the modeUing of production and government behavior have also been accomplished
Interesting innovations in the concept of equilibrium and computation techniques were discussed
The policy analysis was greatly enriched by considering transitional effects together with longer
run steadymiddotstate equilibrium paths generalized equal yield strategies accomodating different
financial crowding out impacts and generalized dynamic policy evaluation indicators
An important set of issues have been addresed by the models surveyed Traditional issues
focusing on the effects of capital taxation and consumption taxes have been pursued In terms of
capital taxation for example the intertemporal nature of the issue was enhanced by allowing an
optimal evolution of capital stock in the economy and forward looking optimal investment decisions
In Andersson (1987) GouldermiddotSummers (1987) and Pereira (1986b 1987d) this is coupled
with a improved treatment of several tax provisions like investment tax credits and depreciation
allowances In terms of the consumption taxes developments in the specification of Intertemporal
household behavior the introduction of overlapping generations and the modelling of
intergenerational links via bequest motives as well as a closer attention to demographic evolution
provided a much improved economic setting for the understanding of the several aspects of the
problem
Furthermore dynamic modelling has permitted the addressing of a set of new issues Policy
Issues ranging from dynamic tax policy analysis to the evaluation of exogenous changes in
government expenditures to the impact of different methods of financing government expenditure to
4 1
the importance of tinanciai croHrii1g out to 1he relevance of consumorS expectations in tne policy
results have been addressed
Despile all of the progress and in part due to such progress several avenues are wide open for
much neded additional research The following seem to be the most Interesting and promising
areasmiddot First the specincation of liquidity constraints to household behavior may help to obtain a
better understanding of policy chenges that affect directly the interest rates in the economy Second
major attention needs to be paid to the incorporafion of a roraign sector (with open capital markets)
Third some efforts are needed in terms of finding adequate computation techniques now that
dynamic modelling eally makes the curse of dimensionality worse than eVOrbull
The modelling of financial markets is the single most unsatisfactory speet of the dynamic
models surveyed here~ The problems associated with provlding a mo~ adequate treatment of
financial markets and in general endogenous and optimal financial deci$ions bull both at the corporate
and at the personal levelsmiddot have 10 be addressed That necessarily requires Introducing uncertainty
into the models To b adequate this must go well beyond the mere assumption of point expeetalions
with the whole array of modelling implementalion and evaluation problems thai generates
bull I i
42
REFERENCES
Abel A and O Blanchard 1983 An Intertemporal Model of Saving and Interest Econometrica
51 pp 675-92
~ Andersson K 1987 Tar-ation of Capital in Sweden - A General Equilibrium Model Ph D
Dissenation UnIversity of Lund Sweder
Auerbach A L KotlikoH 1983 National Savings Economic Welfare and the Structure of
Taxation in M Feldstein ed Behavioral Simulation Methods in Tax Policy- Analysis Chicago
University of Chicago Press
Auerbach A l Kotlikoff 1984 Social Security and the Economics of Demographic-
Transition in H Aaron and G Burtless eds Retirement and Economic Behavior Washington DC
Brookings Institution
Auerbach A L Kotlikoff 1987 Dynamic Fiscal Policy Cambridge Cambridge University
Press
Auerbach A L Kotlikof and J Skinner 1983 The Efficiency Gains from Dynamic Tax
Reform Internatiooaj Economic Reyiew 24 pp 81middot100
Ballard C 1983 Evaluation of the Consumption Tax with Dynamic General Equilibrium
Models Ph D Dissertation Stanford University
Ballard C 1987 Tax Policy and Consumer Foresight A General Equilibrium Simulation
Study Economic Inquiry 25 pp 267-284
Ballard C and L Goulder 1985 Consumption Taxes ForeSight and Welfare a Computatonal
General Equilibrium Analysis in J Piggott and J Whalley eds New Deyelooments in Apolied
General Equilibrium Analysis Cambridge Cambridge University Press
Ballard C D Fullerton J Shoven and J Whalley 1985 A General Equilibrium Model for
t 43
Tax Eoliey Eyalua~iQO NBER University or Chicago Press
Borges A (1987) bull A Survey of Compula1ional General Equilibnum Models CECD Economic
Slud~ 8
Bovenbig L 19B3 imperfect Capital Mobility in a Perfect Foresight Herberger Model
Mimea UC Br~eley
Bovenberg L 1984 Capitel Accumulation and Capital Immobility Qmiddottheory In a Dynamic General Equilibrium Framework PhD dissertation University of California at Berkeley t
Bovenberg L 19850 Dynamic General Equilibrium Tax Models with Adjustment Costs in A iManne ed bull 1985 EtQnomrc ECUjilbrium Model Formulation and Solution Mathematical
lProgramming Siudy 23 Arnsterdam NorthmiddotHoliand
Bovenberg L t 1986 Capital Income Taxation in Growing Open Eeondmies Journa of Pubtic
Eamp(omlcs 3t pp 347middot376 IBovenberg L ald W Keller 1981 Dynamics and Nonlinearlties in Applied General
Equilibrium Models Internal Report Mlmiddot81middot208 Department for Statistical Methods CBS bull
Vooiburg The Netherlands
Charnley C 1981 The Welfare Cost of C~pltal income Taxation In a Growing Economy
IJournal of Political ECQnQm~ 89 3 pp 468middot496 I
Charnley C 19810 Efficient Tax Reform in a Dynamic Model of General Equilibrium I Mimeo Yale University I Charnley C 1982 On the Optimal Taxation In Stylized Models 01 Inlartemporal General
Equilibrium Mimeo Yale University I Decaluw B and A Manens 1985 Pays en Oeveloppement at Modeles Calculables OEquilibre
General - Une Revue de al litterature empirique Monographie Sene 9 Volume 3 Centre de
Recherche et Developpemont Economlque Unrslt de Montreal
44
bull
I I
I
I
1
Erlich $ V Ginsburgh and L van der Heyden 1987 Wher~ do Real Wage Policies Lead
Belgium bull A General Equilibrium Analysis European Economic Reyiew (forthcoming)
air R and J Taylor 1983 Solution and Maximum Likelihood Estimation of Dynamic
Nonlinear Ratioant Expeclations Models Ecooometrica 51 pp 1169middot1186
Feltenstein A 1984 Money and Bonds In a Disaggregated Ec~nonYin Searl 1-1 J Shoven
Eds Aoolied ~eoeral EqJiliDCiulD tioalYsis Cambridge University Press Cambridge
Feltenstein A 19B6 A Dynamic General Equllibrlu Analysis of Financial Crowding Out
Theory with an Application to Australia Journal of Public Economics 31 No1
Fullerlon D R Gorden 1983 A Reexamination of Tax Distorllons in General Equilibrium
Models in M Feldstein Ed Behayora Simulation Methods in Tax pOlicy Analysis Chicago
University of Chipgo Press
Fvllerlon D J Shaven and J Whalley 1983 Replacing the US Income Tax with a
Progressive Consumption Tax JoulOal of Public ECQnomics 20 pp3middot23
Fullerton D Y Henderson J Shaven 1984 A COmparision of Methodologies in Empirical
General Equilibrium Models of Taxation in Searl H J Shoven Eds ~pQlied Geoaral Jguilibrium
lIoalyss Cambridge University Press Cambridge
Fullerton 0 A King J Shaven and J Whaney 1981 Corporate Tax Integration in the United
Statesmiddot a General Equiiibdufi1 Approach American Economic Review 71 pp~ 677~691
Galper H R Lucke E Tader 1985 Taxation portlor Choice and the Allocation of Capital A
General Equilibrium Approach Mimeo
Gill p W Murray M Saunders M Wright 1986 Users Guide for NPSOL (Version 40) A
Fortran Package for Nonlinear Programming T echnicsl Report SOL 86middot2 Deparlment of
Operations Research Stanford University
Glnsbur~h V and L van der Heyden 1985 General Equilibrium with Wage Rigidities An
45
AppJication to Belgium in A Manne ed t Economic EQumprit1m~ Made Formulation and $oJioo
Mathematical Programming Study 23 Amsterdam NorthmiddotHolland ~
Goulet J 1968 1 Adjustrent CO$1S in the Theory o(the Firm Reyiew of EconomiC Studies 35 bull
pp47middot55
Goulder L 1985 Tax-Financed versus BondmiddotFinancedGhanges in Governent Spending
bull Mimeo Harva(d UniversilYmiddot
Goulder l J Shaven and J Whalley 1983 Domestic Tax Policy and the Foreign Sector The
importance of altemative foreign policy formulations to results from a general equilibrium tax
analysis model In ~n methods in Ibe antico of income from caoia1 ed Martin
Feldstein Chicago University of Chicago riSS
Goulder l L Summers 1987 Tax Policy Asset Prices and 3rowtly ~
A General Equilibrium Analysis NBER Working Pper No 212B
Grandmont J 1982 Temporary General Equilibrium Theory Ch 20 in Handbook 01
Mathemalical Eccnornics Vol II NorthmiddotHoliand
Harberger A 1959 The Corporate Income Tax An empirical Appraisal In lax Revision
Comoendjum Voll House Cornmitee on Ways and Means Washington DC Govemment Printing
Office
Harberger A 1962 The Incidence of the Corporate Income Tax JOULn1 of Political
Economy 70 pp 215middot240
Harberger A 1964 Taxation Resource Aliocalrlon and Welfare fn lb Role of Dir bullbulll and
Indjresl Taxes io tbe Fede[al Reserve s~I~Ill Princeton Princeton Unlveresity Press
James J 1985 New Developments in the Applications of General Equilibrium models to
Economic History in Piggott aand Whalley Eds
Jorgenson D and K Yun 1984 Tax Policy and Capital Allocation Harvard Inslitute of
bull
I
I f
Economic Research WP 1107
Judd K 1965 ShonmiddotRun Analysis of Fiscal Policy in a Simple Perfect Foresight Model
Journal of poJiHcal 5CCJiW1l pp 298middot319 r
Upton D J Potbc J Sachsa nd L Summers 1982 Multiple Shooting in Rational
Expectations Models fuoometrlc 50 1329middot1333
Manne A ed 1985 fconQrnjc Eoullibrium Model FOrIDujaljQo and Solution Mathematical
Pogrammlng Study 23 Amsterdam NorthmiddotHolland
Manne A 1985 Or the Formulation and Solution 01 Economic Equilibrium Models In A
Manne ed ECQOQlli-Jr~ibmiddotum ~tQdel EQlJulailon and Solution Malhematical Programming
Study 23 Amsterdam NorthmiddotHolland
Mansur A and J Whalley 1984 Numerical Specification of Applied General Equilibrium
Models in Scarf H J Shoven Eds Aoplied General Eouillbrium AnaiysectIsect Cambridge Universily
Press Cambridge
Merrill 0 1972 Applications and Extensions of an Algorithm thaI Computes FixedmiddotPoints of
Cenain Upper Semimiddotcominuou Point to Set Mappings PhD Disserlation Unlvel$ity 01 Michigan
Negishi T 1960 Welfar Economics and Exislence I an Equnlbrlum for a Competitive
Economy MelroeCOOQ11ka 12 92middot97
Pereira A 1985 On the Existence and Uniqueness of Optimal Ou1put Path lor CRTS Firms in
Adjustment Costs Technologies Mimeo Stanford University
Pereira A 1985a Consumer Expectations and Ihe COSI 01 UllUty In an Intertemporat
Framework Mlmeo Stanford University
Pereira A 1985b Corporale Tax Integrallon and Government Deficits A Dynamic General
Equilibrium Anaiysls Mlmeo Stanford University
Pereir A 1987 On Ihe Computation of the Users Costs of Stocks Ecooomi (forthcoming)
47
Pereira A bull 1987b Equal Yield Alternatives and Government Oeficlts Mimeof Stanford
University
Pereira A 1987c A Dynamic Applied General ~uHlbrium Model for Tax Policy Evaluation
Comlletamp Docurnent2tion Mimeo University of California San Diego (in progress)
Pereira A 1987d Should the investment Tax Credit be Rei(ltroduced Mlmeo University
of California San Diego (ir progreSs)
Piggott J and J Whalley eds 1985 tJew QeyeQpments in Applied General EQuilibrium
bullenalysjs Cambridge Cambridge University Press
Preckel Pbull 1985 Alternative algorithms for computing economic equilibria In A Manne Ed
Radnor R 19amp3 Equllibrium under Uncertainty In K Arrow M Intrilligaor ods
I ibull
Robinson S 1 g86 Mullisectorai Models of Oeveloping CountrieS a Survey Oivision of
Agricullure and Resources UC Borkeley WP 401
Scarf H 1967 The approximation of fixed points of a continuous mapping ~MM Jouwal of
epplid Mathornallcs 15 pp 1328middot43
Scarf H with T Hansen 1973 00 the Computation of EconomiC Eguilibria New Haven Yale
University Press
Scarf H J Shoven Eds 1984 61l1llied Genera Elujlibrium 6nalyssect Cambridge University
~ Press Cambridge
Shoven J 1976 The Incidence and Efficiency Effects of Taxe on Income from Capital
Jurnal of ramie1 Economy 84 pp 1261-83
Shoven J 1983 Applied General Equilibrium Tax Modemng IMF Slaff PaoerS 30
Shoven J and LB Simon 1986 Share Repurchases and Acquisitions An Analysis of which
Firms Participate NBER Working Paper No 2243
bull
48
IShaven J and J Whalley 1972 ft General Equmbrium Calculation of lhe Effects of
Differential Taxation of Incoma from Capital in the US Jauroe of public Econgmics 1
pp281middot321 (
Shoven J bull and J Whalley 1973 General Equilibrium with Tax$s a computation procedure
and an existence pr00f ReviEhi of Economic Studies 40 pp 475middot490
Shoven J J Whalley 1977 Equal Yield Tax Alternatives A General Equilibrium
Computa~ui0nal Tfchnique J~HJrnal of PUblic EconomIcs 8 pp 211~224 bull
Shaven J J Whalley 1984 Applied GeneralmiddotEquilibrlum MOdels of Taxalion and
International Tr8de An Introduction and Survey Journal of Economic literature 23t pp
1007middot1051
Slemrod J 1980 A General Equilibrium Model of Capital Income Taxation Ph D bullDissertation Harvard Univnrsity
Slemrod J 1983 A General Equilibrium Model of Taxation with Endogenous Financial
Behavior in M Feldstein edbull Behavioral Simulation Methods In Tax Policy Analysis Chicago
UniVerSity of Chicago Press
Stone J 1985 Sequential Optimization and Complementarily Techniques for computing
Economic equilibrium in A Manne Ed
Summers L 1981 Taxation and Corporate Investment a q-Theory Approach BroQkinas
Summers L 1985 Taxation and the Size and ComposHion 01 the Capital Stock An Asset Price
Approach Mlmeo Harvard University
Whalley J 1975 A Geneal Equiloibrium Assessment of the 1973 United Kingdom tax
reform Economica 42 pp 139middot61
1JLTIllOS JIQll~)ltG PAPERS PUBLICADOS bull
I
I I
j
112 70 - LUCENA iJiCIIJo ~ IIEnvironrilent Honitoring and Orqanization Structure T (Setellibro 1987)
nI 71 - L~ENA Diogo A Note on the Representation of Information Stxueture (SeteJObro 1987)
nQ 72 - LUCENA D1CgO Envuorment 111)nitoring and Organization
Structure II (Setelilhro 1987)
ril 73 - AlmmE3 Antonio Pais e CAsPAR Yitor bull Tributacao Inoentivos e InvesUlllentos Analise Qualitativa noYellibro 19S7)
n2 74 - liAlltOSA rose Pedro Opt1lllal lIage R1gidity A SUggestedtlthodology to Teet tb Theory and an Application (Outubro 1987)
nI 75 - CABRAL L 11 B Three lIotes on SYJIIlletric Gemes witt itsYJIIlletric Equilibri (NoveJObro 1987)
n2 76 - LUCENlt D10gOTo Search or Not to Search (Fevereiro1967)
nQ 17 - SA Jorge asconcelos euroI A Theory of SynergyB (Fevereiro1987)
n2 76 - VILARES tanuei Jose Os Be IntenaMios IlIIportados ColIIO Factor de Produao (Julho 1987)
nQ 79 - SA J~rge VasCQncel03 e HoY to Copete and Co~unicate in tature Industrial Products_ (aio 1988)
n2 80 - ROIl Rafael Learning and Capacity Expansion in a New larket Under Uncertainty (Fevereiro 1988)
n2 81 PEREIRA Alfredo llarvilo Survey of Dynamic Computational General EquilibriUlll Hodel for Tax Policy Evaluation (Outubro 1987)
Qualquer in ormacao sabre o 1IIorlltlng Papers ja publlcados sera prestada pelo Secretariado de Apoio aos Docentes podendo os ~e3OSf
ser adquirido5 ne Secyo de Vlndas da Faculdade de EconollLia UNL na Travessa Esteviio Pinto Cltpolide - 1000 LISllOA
I AlJlt 1 - uynafllc (lIt MOOero fur all runey tVlIIVllllQlI
dynamic dluggI utllng
I Ba~ll-a rd-fullerton
-Shoven-Vlhalley I no 19 sectors 19S5
Andarsson 1987 ye 2 sectors
~ Auerbach~KotHkof no~1gB3
1983 1967 yesmiddot1g87 1 sector
Ballard~Go uld ar 1983 1985 no 19 sectors
Bovenborg 1965 1986 yes 2 sectors
Erllch~GIn5burg hshy-Heyden no 24 seclors
1957 1__ Feltenataln 1964 196e no 30 sectors
GoUldsr 1985 no 3 sectors
~ GoutdarmiddotSumrnamprs
1987 yes 5 sectors
_ Jorgen9ClIO-Yun ona sector
1984 no produces two goods
Poralr 19S~ 1987 yos 4 sectors
~bJecllve
Produc~rs
technology behavior InpIJ1 1-0
matrix adl costs opt Imal
capital stock
sialic cost min
CES in valu$ added
110
capital labor yes 00
np Invest adjusts
to sFvnns
max 01
interlempora flot rash flow
cnsl min-lOS) Imax of valul)
of the firm~1987
ill
CD-ISBa CES-1987
capital labor
captel labor
no
no
YA quadratic in
I nO-191)3J
qutldratic in Ilk-1gB
ye
YAS
static cost min
CES in vafl1oadded
10
capita labor yes no
no invest edjuss
10 savifQs max ot capita yas ~
value ot the CD labor 00 quadratic In yo Ilrm Ilk
static CD in several no global ltlost mln vaiue added
110 ltsp1al
Jabor Ves no Inves adjusts
to MVrHIS
statle oost min
____ slatio
cost min
max of value of the
firm equivalent 10 two~stage
~ost min me) of
CO in value ad~ed
110 CES In
value added 110
CESin shyvalue added
110 dual
trBRslog price function Cobb-Douglas
capUal labor
cepHal labor
capital labor
capital labor
capital
yes
yes
yes
no
no invost adjusts no to S8vinos net of
El-xog 9011 deflcitsl no
no invest adjusts to savings r
yes quadratic In Y
Ilk
no yo I
yes Intartamporal net cash flow
In V Added 110
labof Y quadratic In I
y bullbull
- 6 -
~ --- ~ ~- -- -~------ --- --- - --~--~ ---~ shy
_
- ---
--
--------
- - - --------
H n v I taxes
olllng allowed debt Ballardfullamprton~
dynamic objective constrfnt tupendlttire (IlleUs endogenous
aU major balanced trade yoarly ybullbull rly endogenous -f ~Shovon WtlitUy US taxes with C(in~Hlt00 pubHc uility balancod cQmposilon no no debt
eiasticllv junel l 1985 rnagt eCP-REV~d1 aU fn3jO[AndtH3Son vearly balanco small opan Gcon J Swodish inturn fa to cr raturn1987 no - inaia dlJn axogonolJs 00 Ini ITnininod poundgtPfUV
-I L)~ in~h~ --~ ~~ ~pound~~---1- ----shy
runrac h~KQllfkofi ai gulD
1963 1937 US m1l3syo - intertnmponl tixogancuamp yo ys cos~d ampConoroy bandmiddotfinanced flon-oppoundIla
B iii lIa fe-Gould 9r nil llltl1gtr balanced 1rndayearly yearly crcoge)Jus
US t)X6S wiih conslanl I
1963 1965 no public utility balanced cOlYlposWon~ no nD debt max budnt ExP~RCV alastlcHy lunct
Bovenberg simp~iriad 1985middotclosod econybullbull rl~ no - balanced exogenous no no debt US laxes 19a6~two counlry1985 1986 bodgel model
Ert leh~G tn aburghmiddot yeally simplified Inlert walt funet Ishorl~run ttads Heyden no balanced exogenous 00 no Belgium
taxes _ desloullbtium flltUensleln
1967 budaet EXP=REV 6-1mprifia-dmlm costs of yoarly yo rast of tho world
1984 1985 producing a ExTax Rev exogenous financed with yo Australian ~s a Ihlrd conSUmer public good money and bonds non-optimal taxes group
Goulder yaarly annual and endogenous aU maJor
laquo1()Sed economy1985 publlc utilily In1ertemporal composillon bull yes yes U~ taxesY rna versions IlEFElQgtfpoundV bond-financed (non-opllmal balanced trademiddotGoulder-Summer yeary yoarly endogonous bull all major
1987 00 public utility balMc composition no no US taxes with constant max budGel EXPREV alasticilY fund Jl)rgenaon~Yun yoarly aU major
no balanced exogenous no1964 US taxes closed economy budGet inteftemporal recurslve endogenousPereira all major
yo public utility eq of molion lovel and yo yos1995 1987 US laxes closed economy rna for debt composition bondmiddot financed oplloll
- 7 shy
bull - -__----- _ - ------ -- ------- _-----
---- ---
------
------
ll-ltiLt 1 uynamlc Lut MOOSS fOf I ax YOtlcy tv8luallon
Inanell ~ets and maf~9ts ojumLrtm- ----_ ----c_~
privata do btl GlvldtHlI pL~lc aliocalQflo rls shy
1rlc ypa expGct assets equity retention dobl 01 savlnq$ promlums path
Sal ra rd fu Barton phys cap1 used 10 buy soquence lt
Shavon-Wha1ley no financial exogenous oxogenous no physicBl 00 bull 01 sIalic 1E myopic 1985 assets capital flqrlil -
Andersson phys cap used 10 buy 1987 no financial no no yo physical no dynamic ~ perfect
assets capital Auerbach~KolUko If phys cap usod to buy
1983 1987 neo financial no no physical cap no dyoamlc ~ perfectYmiddotmiddotnSSf)ts and Qovt debt SailBrd-Goutder phys cap
1983 1985 no financ[al assets-shy -------
Bovenberg phys cap 1985 1986 no iinenclaf
asso1s Ertlch-Glnsburgh~ phys cap1
-Heyden no financlaJ 1987 assefs
FeltoRsteln phys cap 1984 1986 Invostment is
bondmiddotfinanced
used to buy
exogenous exogenous nObull physl~al no dynamic ~ perfllcl capital
used to buy
no no no physical capital
no dynamic ~onlinuos lima
PI perfect
used 10 buy no no no physical no dynamIc PF perfect
capital jllNt) porlod) bull used 10 buy
no no ye physical cap no dynamic PI p4)rfect and govt debt (Iwo Iods) bull
Goulder phys cap used to buy 1995 no financial exogenous exogenous yenbullbull physical cap no PFlfE bull IdynamIc perfoct~assels and am dabt slaUc
- -~ - shyGoulder-Summers phy$ cap
1987 bonds and no e)(oganOU$ no used 10 buy ye dynamic PF perfect _ bullbullultv bonds MullY (exoaenous
Jorgensonyun phys cap 1984 no financial
essels -~ - ------
Pereira phys cap 1986 1967 bonds end
equHy
________
no
8xogenoos
no
exogenous
no
ye
used to buy two type$ of physical
copltal used 10 buy
go dobl bonds equity
no
no
dynamIc
-
dynamlc
~
lE
perfect
fIbullbulllble (nonperlecB
middot6middot
~-- _middot~ ___~w_ - - ___ bullbullbull___ -------- shy~
-----
------
- ---
-
IlflpnUItll tit lun pVII-y PIlUAlIV
q6H1brla oqual fax polley ovnluaU-on efUclency derlvatton Ioorllhm
calibration param~hr computation COmelfed llold tndfcaHH eHecls
Bnl jard~FuU6rtQ n base year back soll1lion steady-sta1e yo inlsrgroup inerseel or Iii
VS transilion lump-sum with Interfemporai 1965
~Shovn Whallmiddotv replication wllh exog Morrills nmamo1ers alQorilhm _~ond~middotsta1$ ears Inc CVor EV
i~noerucn sleady-sfala no agg(ogaled lnlalsectora 1967
base year back sobHon SiMNON package replication wHn oxog valiant of VS Iransilion no Indicamptor lnlertamporal
paramelsrs multiple shoo ling 6~~~l~slat8
AU8rt)ach~Kotllkot steady-slato yo Inlerguna-r a lions
1893 1llH
quatitaUve by replication a$Sium~lkml Gauss-Soidel liS Uln1WOn any rule -lilh 1111ermpOrB
sonsililJ1y I _ ~poundI~Jjnnelill $t(lfdvmiddot~luiA J3djlt15S ~~ nnrilW ~--------
a ail r nimiddotGould or bas9 9tH bac~ solution Ballard~ staadymiddotslalo )05 kltHSclo(a~
1983 1985 VS Iransllinn lump~sum dynalTlk CV inlerlemporal steadY slafe parameters alaorlthm replicaUon+ with oxog Gourder
+stead~stato (ers loc inlareel10rat BCvonborg qunmaiive by dynamic version stoadY-$Igtto yes 1985 1985 YS transition some govt eN Inlortempora
S9rHltivilv OnearizaJian method ropllcaUon assumptionl of Johansons
+steady-sfata paoo palh Erllch-Glnsbu rgh~ actual vorsus no aootagalod inlersectoral2 bftse years back solu1ion variant of Negishis
Heyden replication wIth alltogbull optlmizaUon countorfaclUal no ~odjcator inlottempora 1987 paramlJters aocroach
faltonstaln baso yeat back solution actual versus no aggrampgalfld intersElclornl replication wilh exog Merrills counlerfaclutll no indlclltor middotlntamprlempOull1984 19B6
parameters al90rithm 1981middot1982
Gouldar staady~slala Inlar5ecioralbase year back souUon Variant or 1S lronsilion no dynamic CV inlortempora
stoady state parameters Newlon 1985 feplicalion+ with iUog Fair~Taylorl
+slaildy-sla1e Intargeoorat Gtlulder~Summers base year back solution varIant of stlJadymiddotstate fixed no aggregated fntersectoral
vs trarlsition government Indicator Intertempo1al steady state oatameters Newton
repticaUon+ with exog Fair-T aylQf 1967 +ste-adv-stata bullbullbullndlno financial
Jorgenon-Yun econometric steady-stale Money melfic In1ersecloral
- - 8slimalion vs traosUian no social weUare functton intertampotal1984 +~teadlmiddotstatl)
~ shy
qtJalitaUv8 by NPSCt optimal path ye generalized inlersectorai 1986 1987 Perelr
replication 8S$umptlonl optimizatIon V$ lumpmiddotsum dynamic lntertemporat se~~ttilitf ____Jllgorilhm opllmal pah pen inc CVoclV financial
- 9 shy
bull ~ ___ - _______ - ____ bull___ OM bullbull __ bullbull __~____ ~_
--
- - - --------
--------- - - - --------
---
1nltJ1 bull I ~ 1J11ltI1 v I ~
-
Sa Ita rd-Futlertonmiddot intargration consumption lax ~StHlven~Whall~y VAT capila~ or labor tax
Ch~mqflS1985 Andersson
consumpion tax neutral corporals lax system 1987
it ue rb a c h~ i(at II Sltotf COIlSUft1ptian lax (poundpird ItS hibor Mx~on ufocto of taxation on capital fOfrnJilti1
dvnamio fiscal policv effects af OQvernmanl defioit (3tlard-Goulder
1883 1987
allocts 01 the adoption of a 1ge3 1985 pure consumpHon lax
im~rtmce of cons fCreSClht -Bovenborg analytically orianiad approach
1985 1986 effects ot pure conumption tax capllal income taxation in Opo economies
Erllch-Glnburghshy-Heyden reat wage policios in Belgium
1987 Felter1steln 1984 - theorotical
1985 ~ linnncial crcwdlng out1984 1986
Goulder effects of tax- vs bond-financed 1985 chanQes In pubtlc expondllufe
~~--
Gouidr~SUmmampfI torpora1$ lax cvts reduced 1987 ITO increase 10 gasolioe taxes
Jorgenson-Yun
effects ot several programs of tax reform on the allocation of ~~ptai
1984
-- Purelra
corporate la)( Integration effects cf romiddotinlroducinQ imreslamnl tax credits
1986 1967
10 bull
bull -~- ~- --- _ - ---____--_ - -----_-_ --~ ---- ~
1 1
2 Brief overview of non-dynamic CGE models
The origin of most or the models we survey goes back to the work of Scarf (1967 1973) whlgt
first developed a reliablo algorithm to compute equilibrium price for a ArrowmiddotDebreu economy
HiS algorithm used simpicial Bubdivision techniques and can -be shown to be the computational analog
of the fixed point theorems previously used to prove the existence of equilibrium His technique
could soW a model with an arbitrary number of consumers and commodities as long as all agents
were price takers consumers were subject to budget constraints demands were continuous andmiddot
production did not display increasing returns to scale The alQorithm while guaranteed to converge I
I I I
was relatively slow for prcb1ems involving more than S1If 20 dimensions A maior improvement in
computational speed Woo offered by Merrills algorithm (1972) which used tpe same fundamental
ideagt as Scarfs procedure
Scarfs model (as is true for the standard Arrow-Debreu model) does not include a government
sector ~ neither taxes nor public goocs At one of the mosi promising appHciUions of the new
computational techniqu w~s in Iha area of tax policy evaluation Shoven and Whalley (1973)
exlended the general equilibrium model and computational approach to include a wide rry of taxes
and a governl11ampnt spending plan The original ShavenmiddotWhalley model was static (although the
different commodi~ies CQuid be considered simitar goods available at different datest as In
Arrow-Debrau) and government was aumad to run a balanced budget The data are arranged as a
social accounling matrix The models spacification and calibration is checked by solving it in the
presence of the base set of laxes The result should ba exactly tha initial social accounting matrix
After having pasd this replication check the model is solved for a counterlactua equilibrium In
the presence of a new tux design The result is once again a social accounting matrix The two
equilibria are compared in order to tSS($S the impact of the new tax plan The first uses ot this
bull
1 Z
model for lax policy evauajon were ShovenWhalley (1972) Whalley (1975) and Shoven
(1916)
Completely static models as Shaven-Whalleys are unsatisfactory for many tax reform ISsues
These include corporate lax Integration effects of investment tax credits effects of accelerated
depreciations consumer Or e)penditure tzxes~ importance of saving subsidies like IRAs etc These
are essentialy dynamiC ksues They involve not only the allocation of capital across sectQfsbut
perhaps more Important the capital intensity of the economy But the capital accumulation and
capital reallocation take time and may involve adjustment costs Because of these issues Ballardmiddot
Fulierton~Shoven~Whaloy took the first steps towards developing a dynamic model
In this conlexl the ShovenWhalley model has been extended and implemented for Ihe US
-economy In BuIIBrdmiddotFulienonmiddotShovenmiddotWhalley (1985) While Ihis book completely documents the
mo~el it was used in several publicalions beginning in 1978 Their mode) consists of nineteem
production sectors twelvo households and fifteen consumer goods It includes a very detailed set of
taxes including the federal and state personallncome taxes federal and state corporate income taxes
SOCial Security taxes pro1J0rty taxes unemployment Insurance excise taxes sales taxes etc It
has been calibl att-d to reproduce the 1973 US economy Recentty a version corresponding to the
1983 JS economy has been developed
wThe 8alard-Fullenon-Shoven Whalley model is dynamic In the sense that consumers face a
choice between current consumption and leisure versus future consumption (which can be
purchased via s wlngs) The consumer classes act as if they were maximizing a nested CES utility
function over the domain of a CES aggregate of contempo-raneous consumer goods and leisure and
future consumption subject to their income constraint The parameters of those funcHons
determine the shares of income devo1ed to each commodity to saving and to the purchase of leisure
They also determine the two key elasticities in the models ~ the elasticity of labor supply with
bull
13
respecl to tgte rbullbull1 aoer lax wage rae and the elasticity of saving with respecl 10 the real after lax- rate of retum to capital
In the model consumerS have myopic expectations regarding lulure prices and in particular
regarding the future rale of return 10 capilal Ballard (1983) and BallardmiddotGoulder (1985)
inco~ora1e both perfect foresight and limited foresight into this model Future consumption is
acquired by buying a fixed composition portfolio of real investments that offer an infinite annuity
of returns
The production side of Ihe model is compielely static The mod1 Incorporates a constantmiddot
olasicily 01 substitution betweeen prlmary inputs in production (capital and labor) and fixed I coefficients for intermediate inputs The model distinguIshes between Industrial outputs and
eonsumer goods for the simple reason Ihat tile data are classified differenUy Industrial sectol$ I I involve such categories as forestry and fisherle metal mino9 and publishing and printing while
consumers purchas-3 fUfHi1ure 2ulomoblles and books This fact is recognized in the model by Ibull
incorporatirg a secord stage Of production which converts industrial outputs intO consumer goods
This technology is uStally mod9-led as a flxed~coefficient conversion matrix
The BallardmiddotFINton-ShovenWhalley model assumeS that the private sector finance
marginai ioveslment with the same composition of debt and equity as currently exists in each sector
This is the same arsumrtion that Herberger originally used Investors all hold debt and equity in tile
same proportion Thampreiore this ownership can be aggregated into simply capital ownership For
tax purposes however the separate treatment of debt and equity is taken into account at both the
corporate and persofa Io=vel SimllarlYJ the dividend poIicies on corporale equitY are established
exogenously There are no government bonds in the model since there are no government deficits
The modol is solved for a sequence (as many as 100) of temporary equilibria wltll consumers
allocating income between present and future conumption at each point in time The palh for 1he
14
economy ~ a sel of connected equilibria The connectIon is pra~ided by capital accumulation
Capital accumulation is endogenous and determined by saving The model stans with a sociagt
accOunting mi1rix In the bzse case the economy is assumed to be in a steaqy state growth -path (along
which all relative prices ~fe constant) The madel solves for both the naw steady state growth path
and the transltion 10 it zf~er a policy intervention rhe authors have frequently addressed the
question-of howong rt tKes to effectively settle un10 a new steady S1ate growth path
The dynamics of the model are limited however in that future consumption is collapsed into a
composite commodity Aso the absence of government deficits and the lack ofproductlon dynamics
limits the realism of the BallardmiddotFuliertonmiddotShoven-Whatley model for the analysis of dynamic
policy issues (such as the adoption 01 a consumption tax or Ihe elimination of the investment tax
credit) The models we survey try to Improve upon Ihe Ballard-FlIiertonmiddotShaven-Whalley
dynamic modelling The models Jn~orporate more forward lookTng behavior They improve the
dynamics on the conampumrtkm side of the economy and some of them introduce true 1ynamlc behavior
on the part oI the producers We turn now to the issues inY9lv_ed in developing such dynamic models
15 I I ~3 ~ssues in the modelling Of dynamic behavior
We will 100k firt at the speclficalion of economic behavior of consume(s~ producers
goyemment and Ihe r~sl of Ihe world The modemog of corporale and household financial decisions
I is then addr5sed Finally Ihe eencept of equilibrium is discussed
I
i
I 1
31 Consumers behavior
Early efforts 10 build dynamic features into the economic behavior 01 consumers are due to
Ballard (1963) and Auerbach-Kotllkoff (1983 1984) Their work has been closely followed by
several subsequent authors In faet dynamic household behavior Is the single most pervasive aspect
amohg the models surveyed
All the models Inc-O1orale some loro of lila-cycle behavior Household behavior is dewmloed
by the maximization of an addWVely separable time invariant intertemporal utillty function The
utility lune1ion is defined oyer the domain of the consumption goods In the economy In most of the
models leisure is also an argument in utility so that labor supply is optimally determined
Utility maximi4ation is subject to a lifetime intertemporal budget constraint which equalizes
the presen1 value of consumers income and expenditure More reeentiy in Andersson (1987) bull
Bovenberg (1985) and Pereira (1986b 1987d) the constraint i$ defined as a sequence 01
recursive equations of motion on wealth This has the potential advantage 01 accomodating liquidity
constraints However It should be recognized that in the absence 01 liquidity constraints (ie when
consumers are free to completely borrow against future income) the two specificallons of the
househofd ccnstraint are essentiay equivatent Furthermore tn both versions saving is optimally
determined as a way of translaring wealth intertemporally
1 6
Some lodels now have a sophisticated dynamic specification of hoosehokl In Ballard (1983) -
BallardmiddotGoulder (1985) and Auerbach-Kotlikof (1983 1984 1987) consumers behavio is
embedded in an iterQEneraional setting Each of them has S5 age cohorts simulataneously alive
Som-e measure of jntergener1ional altruism is considered in the form of bequest motives In Ballard
(1983) and Ballard-Gouldr (1985) consumers derive utility directly from bequthing parl of
their wealth Therefore bequests assume 1he form of a scrap value of terminal wealth of a
generation In AuerbachmiddotKotlikof (1983 1984 1987) each generation empathyz bullbull with Ihe
utility of future generations ovar bequethd wealth In addition Ballard (1ge3) includes detailed
demographic poiections wfill the intent of incorporating in the policy analysis the effects of the
bull Post World War tI baby-boom
In the real world household decisions also include the optimat allo~atiOn of saving among
shyalternative physical or financial assets Theories of household portfolio behavior are relatively
Iwen~establshed However they involve uncertainty as a crucial element Accordingly they have
middotnot been incorpora~ed in the context of the determinIstic dynamic models we sutyey here We will 1 Come backt-O this Issue below
32 Producer behavior
The efforts to build dynamic features into the economic behavior of producers are more recellt
and less widely adopted The first attempts are due to Bovnberg (1984 1985) and Summers
(1965) Dynamic behavior has been more fully incorporated In the recent models of Andersson
(1987) Auerbach-Kotlikoff (19B7) GouldermiddotSummers (1987) and Pereira (1986b 1987d)
Part of the reason why production side dinamlcs has been more slowty adopted is the weak
supply of accepted theories referring to the dynamic behavior of the firms In the models referred
I 17
to above gynamlc producion and investment b~havior are middotinduced by the existence of capital
adjustment costs and linked to Tobins q theory Adjuslment costs are designed to capture both the
incomplete mobility of capital aoross industries and installation costs ~e the costs of adjusting
capital towards its cpHmai level)
Adjllstment costs can be conceived as internal to the firm and measured in terms of foregone
outpul along the lines of Lucas (1967) Alternatively adjustment costs can be viewed as actual
external cests incurred together with Ihe purchase costs along the Unes of Gould (1969) With the
exception of Pereira (198Gb 1937d) all of the CGE authors follow the former approach
The firms in the economy maximize their market value as the present discountad value of the
future stream of dividends In Andersson (1981) and Pereira (198Gb 1987d) firms are seen as
maximizing the present discounted value of net cash flow Maximization is constrained by Ihe
adjustmenl cosl lechnology and ~n equation of motion describing Ihe evolution of the capital stock
It should be noted that undereogenous divldendrelention rules the two problems are equivalent II
should also be nOled Ihat a salisfectory economic rational for th bullbullistoo of dMdonds with the
present tax code is missing in the profession (Shoven (1986))
The models with static formulation of producers behavior are characterized by passive
investment behavior Investment merely accomodates to saving in the eccnomy With a dynamic
formulation induced b adjustment costs real investmen decisions are forward looking Investmenl
Is endogenously and optimally determined by the firms A fundamental difference between the
shonmiddotrun in which capital stock is given and longmiddotrun in which Ihe level of capital is allowed to be
optimally determined is emphasized
This extra richness of produc1ion dynamics is not without costs A careful look at the summary
tables will clearly show an inverse relation between the adoptlon of dynamic features in production
and the level of disagreggation of the production side of the economy In fact with production
bull
j
1 6
dynamics $e dimension of the problems is immensly increased Let IS be more specific
In a static framework with constant returns to scale production technology the output level ismiddot
indeterminate The optimal aliocation of inputs can be obtained by cost minimization with any
feasible output level generatillg zero profits Such zero profit conditions are used to solve for the
output prices in terms of the factor prices and hencemiddot to reduce the dlmensionality of the problem
from the -number of commodities and factors to the number of factors Thus even if the model deals
with 30 production sectors the computation of an economic equilibrium can take place using only
the dimensionality of the primary inputs in the economy
Now under certain regularity conditions on the production and adjustment costs technologies
leading to enough concavity of the optimality objective the intertemporal output path for the firm is
endogenously optimally and uniquely determined even with constant relums to scale technologies
(see Pereira (1985 1987a) Accordin~y with adjustment costs in the model optimal profits
will in general be non-zero This result has crucial implications for the computatiQn of equilibrium
in the model~ with production dynamics With adjustment costs no reduction of dimensionality is
possible The curse of dimensionality returns as a binding constraint
The introduction of adjustment costs adds another significant complication to the model With
capital being less that perfectly mobile in the economy different rates of return on capital will
exist in different sectors This is a difficult problem to tackle conceptually in the absence of
uncertainty Also in the real world producer maximization choices include also its choice of
financial ratios (debVequity) and payout rates (dividendretained earnings) These financial
subjects are difficult and much Clf this behavior is still taken as exogenous by the modellers We
will come back to these issues below
middot i
19
bull 33 government behavior
middot
Two issues dominate the modelling of government behavIor First govemment behavior has
Ibeen typically seen in Ihe CGE liIerature for lax poliCY evaluation as constrained by yearly balanced
I budsets (see for example 8allard-Fullrton-S~ovn-Whafley 098Sraquo The analysis of
govemment defidlS and public debl in a CelE context requires a dynamic selling Second the level of
government expenditures is either exegenouty given as in Auerbach-KoUikofi (1984 1987)
6ovenbrg (1984 1985) Feltenstein (1984 1986) and Jorgenson-Yun (1984) or
endogenously (but no optimally) determined by the balanced budget conditions as in
6allard-Fullerton-Shoven-Whalley (1985) Andersson (1987) Erllch-Ginshurgh-Heyden
~
(1987) Gouder (19aS) and Gouder-Summers (1987) In the second case the composition of
bull public expenditures is often optimally determined However the leve~ of government expendi1ures
can only 00 endogenously and optimally determined if the government is seen as an optimizing agent
and is allowed to run deficits
The first attempts to deal wiU the government defiCits in CGE tax models are due to
Auerbach-Kollilltoff (1984) and Feensain (1984) In Auerbach-Kotlikof (1984) government
expenditures afe exogenous They grow at the rate of growth of population However given 1he tax
structure and tax revenues yearly deficits and surpluses are atTowed subject to an in1ertempond
constraint that the present value of future tax revenues equals the present value of future
expenditures
In Fellenstein (1984 1986) government expenditures are also exogenously given He also
allows the government to run deficits to finance expenditures in excess to 1ax revenues However
surpluses are returned to CQnsumars in the form cf transfers Accordingly government ts not
subject to any constraint regarding the future repayment of public debt
In GouJder (19851 the 9Vernmenl maximizes a static social welfare ~nctio subject to a-
bull
20 I i
I
I I I
I bull
balanced budget conSirtlL This follows the optimal alocatkm of government expenditures alpng
the lines of Baliard-FHeiton-Shoven-Whalley (1985) Ttle model Is generallzea to allow for
exogenous changes in HIe time path of government expenditures Two financing alternatives are
consjdeft~d and contr8s1ed EldditionaJ tax revenues and bond issuance
Pereira (198Gb 1137d) auempts to address both the incorporation of deficits and the
determination of govement expenditures The path of government expenditures and the path of
delicilSlsurpluses (ano therefore the path lor debt) are endogenously and optimally determined The
government is seen as mltJximizing n intertemporal social welfare function given the 13x structure
Optimization is subject to a sequence of recursive equations of motion reflecting the evolution of the
ptlbllc debt allowing fDr government budget imbalances It should be stressathat this specification
01 the government cOMtcalnt is equivalent to the specificaUon in Auerbaeh-Kotlikoff (1983 1987)
when no liquidity constraints affect -gove~nment behavior
The extra richness in the treatment of government behavior allows tfte examination of
government debt poices in a truly dynamic setting Also financial crowding out effects induced by
government deficits can be analyzed (see Auerbach-Kottlikoff (1987) Feltenstein (1986) and
Perer (198Gb 1987e))
34 Foreign sector
Most of the open economy models follow the assumptions of balanced trade with import and
export net demands characterized by constant elasticities along the line of Ballardmiddot
-Fullerton-Shaven-Whalley (1965) Such is the case of Salfard (1963) Ballard-Goulder
(19S5) and Goulder-Sllmmers (1987)_ In Feltenstin (1985) the rest of the world is treated as
21
an addjlion~ consumer group
Bovcoberg (1986) develops a model In which two economies lIr considered each following
intertempoml perlect foresight paths These economies meet in the international forum Their
trade relaroMhips are characterized by yearly balanced trade accounts
None of the new generation of dynamic cae tax models has yet IncorPorated the lnternational
capital flows as done in aoulder-Shoven-Whalley (1983) for the earlier Ballarrshy
-Fullertonmiddot Shaven-Whalley (1985) model This is unfortunate as there is an important
Intertmporal aspect to international lending The first al1empts along Ihese lines are due to
Andersson (1987) and Erlich-Glnsburgh-Heyden (1987) Andersson (1987) in his model of the
Swedish economy adopts a closed economy appraoch in which rates of return in the domesJic
economy are largely determined by the international capital markets Fidinterest rate induce
Intlnatlonal capital flows which determine and finance the International lrade imbalance In
lurn In ErlichGinsburghHeydenmiddot (1987) foreign trade Is generaled according to an
intertemporal Irade welfare lunction with constant import and export elasticities In the shortmiddotrun
they allow inlernational trade inbalances which generate capilal ftows to the domestic households In
the long run however trade balance is assumed
35 The need for financial markets
A dynamic economic Wucture not only provides the ideal environment 10 model many features of
economic behavior it also permits one to Incorporate the financial side of the economy If the
government is allowed to run deficits the question 01 deficll financing automatically follows If
investment is optimally dewmined and returns to capital are different aClOSS sectors problems of
investment financing arise If there are seeral final1cfal assets In the economymiddot government
4-
I i
ibull
I I I I
I
j
22
bonds privpte bonds and equity (or simply physical capital installed inmiddot different sectors) the
problem of allocation of saving among assets with potentially different returns arises
The papers surveyed var greatly with respect to l~e extent of their attention to the financial
side of the economy At Clne extreme are the models in ~anardmiddotFuliertonmiddotShoven-Whalley (1985)
Andersson (1987) Ballltld (1983) Ballard-Goulder (1985) Bovenberg (1985 1986) and
Erlich-Gjnsburgh-Heyden C19B which are devoted exclusively to the real side of the economy
In turn Auerbach-KotliKoff (1984 1987) Feltenstein (1984 1986) and Goulde (1985)
allow for government debt In these models saving finances changes in government debt and physical
capital Private and publlc 2ssets are perceived by the households as perfect substitutes The
allocation of saving merely adjusts to the relative demands for funds
Feltenstein (1984 1986) is the only model surveyed which introduces money Government deficits are financed by issuing money and bonds according to an exogenously given rule Money is
demanded by consumers for transaction motives and an exogenously given fraction as a store of value
On the other hand government bonds and physical capital are the vehicles for the intertemporal
transfer of wealth
Summers (1985) and GOlldermiddotSummers (1987) introduce a whole menu of financial assets shy
firm specific equity capital Different assets earn different rates of retum However such rates
are equal up to constant and exogenous sector speCific risk premia Therefore the introduction of
constant exogenous risk premia as helpful as it may be in the context of calibration does not solve
the main issue of the non-optimality of the allocation of saving Also talking about risk premia in a
deterministic context is somewhat unsatisfactory
Pereira (1986b 1987d) also introduces a whole menu of assets private and public bonds and
firm specific equity However all the assets are expected to yield the same rate of return and
therefore perc~jved as perfect substitutes The different asset types allow consideration of
23
exogenOU$~8tUequjty triO dividendlrc~enlion rufes and therefore seve-ral sources of investment
financing bond equity and retained earnings
The nonmiddotcptimalHy of the allocation of saving and the absence or exogenelty of corporate
financial rules is _a reftectlon of the limitations of the determInistic approach Either in a context of
perfect anticipation oi the future prices Of in general within the t$atm of point price
expectations all the papers follow a deterministic approach Under such circumstances consumers
either expect diHerent rales of return (inclusive of risk premium) across assets in which case
they wilt buy only one asset (that with highest rate) Of they expect equal rat~s of relurn in which
case Ihey are indifferent about the asset composition of Iheir portfoliO There is no way of
Iradingmiddotoff rales of return and risks to obtain an optimal interiof solution to the problem of the
allocation of saving
The most advanced contribution in the modelling of saving allocation in a CGE setting is due to
Slemrod (1980 1983) in the context of a static one-period model In his model consumers act
according to a two stage separable decision process They Hrsl decide on how much to save Then
they decide on the allocation of saving according to an Indirect uttlity function dependent on the rates
of ratum and variances offered by the different assets in the eoonomy The sourCe of riskiness in the
ampConory comes from an uncertain marginal product of capital On the other hand aside from
portfolio decisions the res of the economy is insulated from uncertainty
The most complete contribution in terms of the treatment of the corporate financial rules is
FulienonmiddotGordon (1983) In a variant 01 Ihe BailardmiddotFulienon-ShovenmiddotWhelley (1985) model
Fullerton and Gordon have capital intensity and optimal financial decisions Jointly deter~ined
through a two siage process The cosl of financing capital is minimized by trading off the tax
advantages of debt against the epected real bankrupcy coSls inherenl with high debVequity ratios
Given the optimal debtequity ratio the level of inveSlment Is chosen such that at Ihe margin the
24 l
bull
rerurn pn ~uity equas the return on bonds plus an exogenous risk premium
36 Market assumptions equilibrium and expectations
Virtually all of the papers arc characterized by Walrasian market clearing assumptions All
markets are perlect1y competi1ive Atomistic competition among agents 1s assumed even thoughmiddot only
a flnile number of agents is considered Virtually no mark~t disequilibria or price stickiness afe
considered The only exception is ErlichmiddotGinsburghmiddotHeyden (1987) bull In thei model of the Belgium
economy I the _wage rate is fixed in the short-run Therefore in the shOrlwrun desequilibrium in
the labor market will generate endogenous unemployment However in lh$ long-run all prices
Including the wage rata are Ilexible and accordingly ali markets ciear
Given the dynamic nature of behavlor in the economy markelmiddotcfearing prices in each period
depend on expectations of fture prices and on tax variables In the economy There are essentially
two ways at interpreting the economic equilibrium in such a dynamic conte~bull If future prices are
perfectly anticipated (io expectations are self-fulfilling) a perfect foresight equilibrium
prevaHs Then future actions are merely the implementation of current decisions for future
periods However jf ptica expecla110ns are not perfect (Le agents make mistakes w1th re$pect 10
future prices) then a temporary or shortrun equilibrium prevails Markets ciear and clearing
prices depend on fu1ure price expectations Current plans about the future are typically not
precisely implemented They will be revised as more or better information becomes available to the
economic agents See Grandmant (1982) for a comprehensive survey 01 the temporary equilibrium
literature
Wilh the exception of Gould (1985) and Pereira (19S6b 19S7d) all the models surveyed
adopt th~ concept of a perfect foresight equilibrium In turn BaliardmiddotGouldr (1965) considers a
J
2S
flexible am9unt of foresight In terms of the number of years over which price movements are ~
foreseen Pereiras model (1986b 1987b)) is flexible in a somewhat different way in that it can
indude any range of foresight from myopia to perfect foresight The choice between the perfct foresight and lemporary equilibrium is ultimately to be made on
philosophic grounds It can be argued that less than perfect expectations Imply that agents are
irralional in some way (see Auerbach-Kctiikoff (1987 p 10) However the reverse argument can
be made One can question wiether agents are reaHy rationar and perfectly knowledgable about
future prices
Recent evidene of Balird (1987) Ballard-Goulder (19851 Goulder (19851 and Pereira
(1986b 1987d) suggests that the choice in modemng expectations is an important one They show
that the degree of foresight inl0 thc future (ranging from perfect foresight to-myopic expectations)
may have dramatic impacts on the pollcy conclusions of Ihe model Accordingly the best research
strategy may b to design models which are flexible enough 10 allow for different rules regarding the
formation of expectations With the exception of the articles just mentioned sensitivity analysis
bullhave previously not been performed along this dimension
In terms of implementation the two concepts of equilibrium - perfect foresight and temporary
equilibrium have different implications The dimensionality of the equ1ibrium soll1ion algorIthm
is involved Suppose we have a model with ten markets to be run for a period of 50 years Aside
from normalization a perfect foresight model implies computing prices In 500 dimensions while a
-t~mporary equilibrium model requires sorving SO equilibria each in ten dimensions Given that
computational speed often varies with the cube of the number of dimensions the lemporary
equilibrium formulation is potentially strikingly more feasible However Ballard (1987) and
Goulder-Ballard (1985) have devoloped techniques to greatly speed the computation of a perfect
foresight equilibrium
26
j
I
4 Jrnplemenlalion and Issues on policy evaluallon
41 Calibration and equilibrium comparlslon
CGE models at typically parameterized by the use of a calibration procedure Some parameters
are exogenously given However some crucial parameters are determined in such a way that the
model replicates the data for a given base year See MansurmiddotWhalleY 11984) for an eXlensive
discussion of 1~i$ issue
CalibratIon in a dynamic context is generaliy interpreted as requiring two properties First
replication of base year data is required Second the model is parameterized to simulate an
intertemporal balanced growth path when Ihe base policy is maintaineq This Is the approach
followed by Ballard (1983) BallardmiddotGoulder (1985) Goulder (1985) and SummersmiddotGoulde
(1986) It follows the practice of BallardmiddotFuliertonmiddotShovenmiddotWhalley (1985) with their
sequential cquilibrium dynamics
Other authors Auerbach-Kotlikof (1983 1984 1987) Boyenberg (1984 1985 1986)
and Pereira (198Gb 1987d) follow whal we cali a qualitalive calibration The slructural
palametars are exogenously chosen so that lh economy follows a reasonable path into Ihe fulure
ThaI has 10 do wllh the fact thaI given the recursive nature 01 the dynamic economy and aside tom
such structural parameters only iniUa stock values are needed to run these models Given initial
conditions on the stocks of say private wealth capltal and government debt agents will optimize
and Ihereby generate a lirst round of net demands and equilibrium conditions In lurn the
equilibrium prices will determine the evolution of the stock variables into the next perIod
There are several potential problems with calibration in the conlext of dynamic models The
assumption of a steadymiddotstate growth ~ath in the base case can be questioned First while
27 I Ibullbull
steadymiddotstaU is a possibity it certainly is not the only meaningfull solution to dynamic models
Even in the case of a perfect foresight equilibrium the model implies an equilibrium path which i may or may not involve colanced growth The model not the modeller should dictate the nature 01 I themiddotbase case path Second in the context of a temporarY equilibrium path a steady state solutlcn is Inol a likely model QutcOn1tl In fact unlessmiddot expectations are stattc short run behavior consistent
with a steady-state evolution will in general not be generated On the other hand if static I expectations are self~fufjili[1g we have in fact a perfect foresight model Third even the base year
replication requirement may cause problems In Ihe cOntexl of temporary equilibrium Any
calibration parameter would be conditional on expectation rules which is probably an undesirable
feature
The nalur of the two-requirement calibration strategy is very muchin the spirit of the bull
traditional design of the comparision of alternative equilibria comparis_ion between a steady-state
base case on one hand and alternative paths including a transition period and a finat sleadyyenstate on
the other hand Pereira (198Gb 1987d) compare different (not necessarily steadystate)
equilibrium paths The arguments against such a procedure are based on the idea that the impact of
policy changes can be observed most easily since all departures from the steadymiddotstate can be
attributed to the altemalive polioy On the other hand the base case so dEfined as a steadymiddotstate is
consist~nl with previous work in a less dynamtc setting and Iherefore allows a common standard
for comparing model results
42 Computation algorithms
We are still al a stage in which basically each author uses a different computational technique
Th development of dynamic models - in parlicular wilh adjustment costs andior perfect foresight
bull
28
has corresponded with the decline in the use of fixed point algorithms In -fact given the relative
arge dimensions inevitably involved such algorithms lend to be very inefficient at the best and
chen prohibitively stow See Stone (1985) and Preckel (1965) for a comparative assessment of
different computation techniques Among the models surveyed only Ball1rdFulle(ton~
-ShavenmiddotWhalley (1985) and Feltensteln 1985 use Merrills variant of the fixed point
algorithm 1echnique
AuerbachmiddotKoHikoff (1963 1984 1967) follow a three stage procedure They first compute a
base case steady-state then a revised case steady state and finally transition path for the economy
berveen these two steady-states tn all stages a Gauss-Seidel iterative procedure is used
Ballard (1982) Ballard-Goulder (1965) Goulder (HiSS) and Summers-Goulder (1986)
~
use a method developed by Ballard and Goutder which is similar in many awects to the FairmiddotTaylQr bull
(1983) algorl1hm Short-run equilibria are calculated (using Merrills algorithm) parametric on
nrice expeC~lions The model is then iterated to generate self-fulfilling intertemporaf expectations
and the corresponding perfect foresight equilibrium In a relatively similar approach Andersson
1987 uses a simulation program SIMNON developed In the University of Lund This program
can handle two-paint boundary problems In a fashion consistent with the multiple shooting
algcrlthm (see Lipton-Poterba-Sachs-Eummers (1982))
Boyenbergs computational approach (1985 1985) differs from the other models surveyed in
that he relies heavily on analytical techniques Computations are done by using a dynamic version of
Johansons linearization method Being essentially determined by the continuous time nature of the
I I
f
model thiS linearization model has the disadvantage of confining the analysls to infinltesinal changes
around the base case equilibrium (see Bovenberg (1985) p 53) Pereira (1986b 1987d) uses an optimization algorithm NPSOL - developped by
rGillMurray-Saunders-Wrlght (1986)) This algorithm is used 10 compute the sequence of
I
29
Shortmiddotrun temporary equilibrium which makes up the intertemporal equilibrium path The
equilibrium conditions are seen as nonlinear equality constraints in the minimization of an artlcial
objective function The prices are normalized to the unit simplex by an additional linear equality I I 1 bull constraint
II ErlichmiddotGinsburgh-Heyden (1987) follow a unique approach in that they use a variant of the
optimization technique introduced by Negishl (1960) The economic equilibrium can be generated i
as a solution of a mathemallcal program Ihe objective function of which is a weighted sum of the
utility functions of the various agents while tho constraints set consists of the market dearing 1
conditions Ginsburgh-Heyden (1985) have extended Negishis resulllO tho case of downward price
lrigidities I
Finally the paper by Jorgonson-Yun (1984) is also unique in tIlat it is the only I
econometrically estimated model Different blocks for the consumption and production Side of the
-
model are separately estimated to provide the necessary structural parameters i
I The diversity of computation techniques is yet another indicator of the exploratory nature of the
body of literature surveyed in this article
43 Equal yield comparisions
The link between a base case and counteriactual simulations Is usually provided by the concept of
equal yield Shaven-Whaley (1977) discuss the meaning 01 equal yield in a general equilibrium
context when government is not allowed to run defICits Equal yield is interpreted to moan constant
public utility Government base case utility is maintained in the counferfactual experiments With
balanced budgets the concepl of equal yield is unambiguous The new equUibrium prices and the
balanced budge condition will del ermine the minimal expendilure and taxes needed to maintain base
30
case publi~ utility Accordingly in general equal yield is inconsistent with equal nominal tax
revenue Some change in tax revenue Is necessary Different tax replacement schemes ate
considered to assure that enough tax revenue is collected This is the approach essentially followed
by most 01 the papers surveyed Only Feltenstein (1985) Goulder (198) and Jorgenson-Yun
(1984) chose not to follow an equal yield strategy
The question is of how to Interprete lila concept of equal yield when the government is alowed to
fUn deflcLts The optimal leve of expenditure for base ease public trliUty can now be taAinanced
bond financed or financed by a mix of bonds and taxation We have several versions of equal yield In
particular equal yield may now be consistent with equal tax revenUE Furthermore some meaSure
ltgtf financial crowding out effects of government deficns can be inferred from the comparislon of the
several equal yield alternatives These issues are extensively discussed ~Pereira (1987b)
44 Price expectations and the dynamic generalization of compensation
Indicators
The sum of equivalent and compensation variations over households is the most widely used
aggregated measure of efficiency gains or losses In the context of $teady~state eomparisions andor
when complete future markets exist andlor perfect foresighl is assumed there are no difficulties
associated with the use of the standard Hicksian indicators In fact correct future prices are known
in these cases
If expectations are not selfmiddot fulfilling andlor future markels are not available a dynamic
generalization is necessary 8allardmiddotGoulder (1985) provide some steps in that direction by
defining an indicator that accomodates periods far into the future when households do not have
perfect foresight but a steady~state prevails On he other hand this issue is more fundamental in
i
[
I
i I I I I [
I I
bullI
31
the contex~ of a Gner1 lemporary equilibrium ftamawor~ In such Circumstances Pereira
(1986a) dev~lops a cnamlc generalization which is obtained as the present discounted value of a i i
sequence of short-fUn optimal expenditure functions consistent with a base case expected future I stream of IlWities I
32
5 Empirical evidenee from selected policy Issues
Dynamic tax models have generated several important results which escaped Sialic modellers
The following is a seleCpoundd set of issuos whIch stress the marginal benefits of dynam~ modelHng of
economic behavior in innovltlttive areas of analysis The discussion of a consumption tax emphasizes
the benefits of dynamic tusehold behavior and intergenerational aspects The study of the
flliminatlon or the re~intToduction of the investsnemt tax credits is made meaningful by the dynamic
modelling of production betlavor Modelling governmenl in a dynamic conlexts allows the modeller
to study the irrrpact of government deflcits Finally a dynamic framework lets us appreCiate the I
relevance of consumer expectations in terms of the evaluation of policy alternatives
i
i
51 Consumption Tax
t
By Its very nature a consumption tax can not be adequately investigated with a static model bull
Fullerion-Shoven-Whaicy (1983) use Ihe model of the US economy s described in
Ballard-Fullerton-Shoven-Whalley (1935) to evaluate the movement from the currenl US tax Isys1em to a progressive consumption tax Since their modeJ Incorporates a Jaborleisure choice
where leisure is an untaxed commodlty their results reflect the fact that both the consumption tax
and Ihe present system are dslortionary
Concentrating on the rntertemporaf distortions they show that sheltering more saving from lhe
current US income tax could improve econornic efficiency even if marginal tax rate increases are
necessary in order to maintain government fevenue~ At first you have a revenue shortfall
However the economy moves to a higher sustatned growth path and uitimatefy lower tax rates
cangenerate the same revenue path Also wages increase in the long run with this policy The
bull
bullbull ~
33 shyt
i
I
I
present value of welfare gi[lS for a potiey of complete savin9 deduction wilh marginal tate
adjustments (consumption lax) is simulated to be around $500 billion to $600 billion 01 1973
dollars
They also investiga1e th 1ngth of ttme it takes th~ economy to adjust to these policy changes
Roughly they estimate the long run to b thirty years although this figure Is very sensitive to the
specification of the savings elasticity - a crucial parameler in this model
52 Investment T~x Credit
Goulder-$ummers (1 S87) address the impact of eliminating the lnvestmenl Tax Cradit (ITC) on
intersectoral capital formation and on economic growth Their model isYa1ticularly adequate to
addies$ this issu~ in that is postulates a forward looking investment m9we with adjustment costs
They show that the eliminating the ITC causes a reduction in the rale of investment Inveslment is
estimated to fall by about 7 in the short-run and by about 12 in Ihe new long run steady state
In lurn a previous policy anncuncement lowers Ihe overall al1ractiveness of Investment and leads 10
a downward shift of the Investrrent profile On the other hand the combined effect of a revenue
neutral simultaneous efimiraron of the ITC and reduction of the corporate tax rates is a long run
reduction ef the capital Sieck by 35 This pattern suggests that a revenue neutral increament in
both Ihe corporale tax rates and the inveSlment lax credits would be preferable 10 the formula
adopted in Ihe US Tax Reform Act of 1986
Pereira (1987d) addresses the impact of the ITe from Ihe oland point of Ihe curTenllax syslem
under the Tax Reform Act of 1986 Given Ihe concern over the potential depressiVe effects upon
savings and investment of the new tax system in conjunction with deficit reduC1ion mechanisms of
the GrammmiddotRudman type a pertinent question is should Ih lTC be e-introduced In the context
bull
34
-of his mod~1 of the US economy Pereira focuses on the trade-off between the distortion in the
intersectoral allocation of capital induced by the lTC the lower tax revenues it generates and
potential financial crowding-out effects on one hand and the positive effects of lowering the relative
price of new capital goods on the other hand Preliminary results confirm the qualitative results in
Goulder-Summers (1987) suggestting a welfare gain of about 2 of the present value of GNP in
the best scenario of absence of replacement taxes Furthermore preliminary results show an
important time pattern to welfare gains with the average benefits increasing the further you look
into the future This pattern is due to the presence of constraints to intersectoral mobility of capital
and inertia in the adjustment towards the optimal capital levels as reflected by the presence of
adjustment costs
I
I
bullAuerbach-Katlikaff (1gB) in the context of their intergenerational model of the US economy
consider the impact on savings capital formation and interest rates of deficit policies and balanced
budget increases in government consumption They conclude that deficit finance and government
consumption can significa-ntly crowd out capital formation and lower the welfare of future
generations However crowding out from deficit finance is a very slow process because it results
from increased government spending over potentially long horizons Also deficit policies may
substantially influence the longmiddotrun interest rates while leaving the short-run rates essentially
unaffected Finally and in opposition to the central role adjustment costs seem to play in both
53 Financial crowding outeff~cts of government deficits
Goulder-Summers and Pereiras models Auerbach-Kotlikoff suggest that the time path of interest
rates induced by a policy of deficit financing seems to be insensitive to the presence of adjustment
costs
bull
35
54 The role 01 consumer expectatlons
The expectations analysis has uncovered one of the benefits of dynamic modelling
Ballard-Goulder (1985) find that the appeal of adopting a consumption tax depends on the level of
foresight possessed by the consumers Furthermore additional foresight may be welfare worsening
This is a second best type of result This tends 10 occur under policies that lead to capital deepening
and declining rate of return to capital over time To the extent that consumers have more foresight
they will be beller equippod to anticipate the fall in the rental price of capital Consequently if a
saving incentive is enacted people save more with myopia than with perfect foresight Given the
exi~tence of taxes on capita and the discrepancy between the private an~--sccial returns to capital
the greater savings with myopia is beller socially Ballard-Goulder find that the welfare gains
from a consumption tax are reduced by about 10 when we move from myopia to farleet foresight
Therefore the attractivenness of adopting a consumption tax seems fairly [obust across these
foresight specifications
I
bull
36
6 Tbe power and weaknesses 01 dynamic modelling the example 01 corporale
tax Integrallon
The Issue of corporate tax integration allows us to sh~w the stre-ngths and weaknesses or the
dynamic approach Empirical evidence using CGE techniques indicates that depending on the precise
scheme of integration and the 1ax replacement methods integratiOn may have substantial effects In
Ihe work of FuliertonmiddotKingmiddotShoven-Whalley (1980 1981) total integration was found to yield an bull
annual static eHielency gain of $4 to $8 billion in 1973 dollars Simulated dynamic gains may be as
large as $695 billion or abcu 14 of the present value of tulUre consumption and leisure in the
US economy These gains result primarily from interindustry reallocalions of Investment and an
improved inter~emporal allocation of consumption
Mora recent work emphasizes hOw consumers asset portfolio decisions and firms finanCial
decisions affect the efficiency gains from integration Slemrod (1980) focusing on consumers
asset ponfoio deCisions finds static efficiency gains which are about twice -as Iarge as those
repOrted iiy Fulierton-King-ShovenWhalley (1961) FuUrlon-Gurdon (1983) focus on the
firms financial decisions They roporl efficiency gains of 6 of GNP from the elimination of the
tax distor1ions favoring debt However when they eliminate the corporate tax and repiace it with
increaSed persona income taxes addlHonal dls10t1ions are created in the optimal labot~leisure
decisions These distortions tend to dominate the analysis slJ(h that the overall effects of complete
integration are very modest Galpr-LuckmiddotTodr (1986) address simultaneously consumers
asset pOrlfollo decisions and the firms financial decisions They also find very modest efficiency
gains from integralion
The above results are important but may be severely biased There are severa aspects of
economic behavior and modelling crucial for the study of Income tax Integration whiCh have no been
i 37
captured innny of the above papers One aspect is the absence of government deficitsmiddot a balanced
budget is assumed It is true Ihal resource crowding OUI is caplured In these mOdels but financial
crowding out induced by government spending is nOI Second investment is not derived from
optimtztlcn behavior Investment behavior passively accomodates endogenous saving decisions As
a consequence the differential impacts of policies in the incentives to save and to invest are not
captured Aso full capital mobility across secors is assumed with instantaneous capital
adjustmenlS towards optimal levels This assumption rules qut different costs of capital across
seclors and therefore differentiated reactions to tax policies changes Finallyhe modelling of both
gov9rnmerlt deficits and of endogenous real and financial investment decisions necessitates the
~nsideraiion of a dynamilt framework and the introduction of financiaJ assets govemment bonds and
iprivate financial assets A dynamic framework also highlights the efficiency effects of Integration on r
ithe ptimal intertemporal decisions The above models are either sIalic (Slemrods and
GalpermiddotltJckemiddotTode(sj or a dynamic sequ~nce of otherwise Slatic model bullbull I Pereir a (1986b) develops a dynamic applied general equilibrium model bull to study the tbull
efficiency effects of integration on the growth and allocation of investmenl across sectors Special
Iattention is paid to the structural affeelS of resource and financial crowding oul induced by
gove mect spending and deficits and to the real and financial Investment reactions across industries
to both tax integration and flnancial crowding out His mode departs from previous work on income
lax Integraticn and for that matler from most 01 the CGIE literalure for tax policy evaluation in
several direcllons It encompasses an endogenous sequential equilibrium struelure founded on
dynamic behavior with flexible expectations Government deficits are optlmaUy determined
Investment decisions are tward looking and the resull of optimizing behavior Several financial
assetsmiddot public and privatemiddot are considered
Simulation resulls Indicale that the welfare gains from Integration under large deficits are at
32
best lJlodest when measured in terms of the GNP The elimination of the corporate tax and its I replacemenl by increased income lax rates yields long run benefits which are never larger than
2 of the presenl yalue ct telr consumption and leisure onder the previous tax regime and 1
under the current tax regime However the short run w~lare effects of full integration tend 10 be
very low and in in some scenarios even negative This is a flew intertemporar pattern of efficiency Ieffects which reflects an adjJsiment lag in the interindustry investment decisions due to the
existence of costs of adjustment On the of her hand partial integration achieved by excluding I dividends from the corporale lax base syslematically yields negalive efficlencyeffeclS In Ihe long Inm these can be as high as 3 of present value of the future value of consumption and leisure This
is a new second best effect suggesllng Ihat less than complete integration may have perverse I efficiency effects
The dynamic slructure of PerclrBs model snowing for an improved freatment of real investment
declsions and government deficts provides a potential setting for a more accurate measure of the
costs and benefits of inlegration The simulation resullS suggest lower benefits and an intertempora Ibull
pattern of growing benefits
Simulation results also clearly suggest robust negative effects from partial integration How
reliable is this result If dJvidends were deductible from the corporate lax base (as are interest
payments) the preferenlial Ireatmant of debl over equily would be ellminaled Corporallons should
be expecled 10 react by decreasing the optimal deWeqully rallo However corporale financial
gecisions are exogenously set In Pereiras model as in aft of the other models surveyed that go as far
as dealing with the issue
Also withdrawing dividends from Ihe corporate tax base is a way of eliminating Ihe double
taxation of dividends at belh Ihe corporate and personnal Income levels How will the optimal
aliocellon of household saving accomodate that change in Ihe relalive return to corporale equity
39 I
t However hOusehold portfolio decisions are exogenouly set in Pereiras model as in all of the other
models surveyeltj go as far as dealing with the issue
IIt is sate to say that the evaluation 01 the benefits 01 partial Integration as defined above are
sevrey underestima1eO Meanwhile the distortionSoenerated by the increase in the marginal
personal tax rates designed to ralse the revenue foregone by the dividend exclusion are fully
aCCltlunted for A good measure of the costs of integration together With a poor evaluation of the
benefils may very well be the reason why partial integration is Simulated 10 yield neg alive
eHikiency gains
On a different vein as most of Ihe model surveyed here Pereiras is a closed economy model It
is legitimate to wonder how the resullS would change If international capital flows were anowed
IThis almost certainly would affect financial crowding out since a good d~al of the capital inflows
could be used to finance public debt
For these various reasons the results should be treated as preliminary but they strongly I suggosl that the dynamic structure provides valuable new insights into the corporate tax integration SSU2
40
7 Concluding remarks I I I i
What has been acccmplished Ihis far The success of Ihe CGE research in laelding the challenge
of dynamic modelling can not be denied
I iGreat progress in the modelling of household behavior has been achieved Promising
developm~nts in the modeUing of production and government behavior have also been accomplished
Interesting innovations in the concept of equilibrium and computation techniques were discussed
The policy analysis was greatly enriched by considering transitional effects together with longer
run steadymiddotstate equilibrium paths generalized equal yield strategies accomodating different
financial crowding out impacts and generalized dynamic policy evaluation indicators
An important set of issues have been addresed by the models surveyed Traditional issues
focusing on the effects of capital taxation and consumption taxes have been pursued In terms of
capital taxation for example the intertemporal nature of the issue was enhanced by allowing an
optimal evolution of capital stock in the economy and forward looking optimal investment decisions
In Andersson (1987) GouldermiddotSummers (1987) and Pereira (1986b 1987d) this is coupled
with a improved treatment of several tax provisions like investment tax credits and depreciation
allowances In terms of the consumption taxes developments in the specification of Intertemporal
household behavior the introduction of overlapping generations and the modelling of
intergenerational links via bequest motives as well as a closer attention to demographic evolution
provided a much improved economic setting for the understanding of the several aspects of the
problem
Furthermore dynamic modelling has permitted the addressing of a set of new issues Policy
Issues ranging from dynamic tax policy analysis to the evaluation of exogenous changes in
government expenditures to the impact of different methods of financing government expenditure to
4 1
the importance of tinanciai croHrii1g out to 1he relevance of consumorS expectations in tne policy
results have been addressed
Despile all of the progress and in part due to such progress several avenues are wide open for
much neded additional research The following seem to be the most Interesting and promising
areasmiddot First the specincation of liquidity constraints to household behavior may help to obtain a
better understanding of policy chenges that affect directly the interest rates in the economy Second
major attention needs to be paid to the incorporafion of a roraign sector (with open capital markets)
Third some efforts are needed in terms of finding adequate computation techniques now that
dynamic modelling eally makes the curse of dimensionality worse than eVOrbull
The modelling of financial markets is the single most unsatisfactory speet of the dynamic
models surveyed here~ The problems associated with provlding a mo~ adequate treatment of
financial markets and in general endogenous and optimal financial deci$ions bull both at the corporate
and at the personal levelsmiddot have 10 be addressed That necessarily requires Introducing uncertainty
into the models To b adequate this must go well beyond the mere assumption of point expeetalions
with the whole array of modelling implementalion and evaluation problems thai generates
bull I i
42
REFERENCES
Abel A and O Blanchard 1983 An Intertemporal Model of Saving and Interest Econometrica
51 pp 675-92
~ Andersson K 1987 Tar-ation of Capital in Sweden - A General Equilibrium Model Ph D
Dissenation UnIversity of Lund Sweder
Auerbach A L KotlikoH 1983 National Savings Economic Welfare and the Structure of
Taxation in M Feldstein ed Behavioral Simulation Methods in Tax Policy- Analysis Chicago
University of Chicago Press
Auerbach A l Kotlikoff 1984 Social Security and the Economics of Demographic-
Transition in H Aaron and G Burtless eds Retirement and Economic Behavior Washington DC
Brookings Institution
Auerbach A L Kotlikoff 1987 Dynamic Fiscal Policy Cambridge Cambridge University
Press
Auerbach A L Kotlikof and J Skinner 1983 The Efficiency Gains from Dynamic Tax
Reform Internatiooaj Economic Reyiew 24 pp 81middot100
Ballard C 1983 Evaluation of the Consumption Tax with Dynamic General Equilibrium
Models Ph D Dissertation Stanford University
Ballard C 1987 Tax Policy and Consumer Foresight A General Equilibrium Simulation
Study Economic Inquiry 25 pp 267-284
Ballard C and L Goulder 1985 Consumption Taxes ForeSight and Welfare a Computatonal
General Equilibrium Analysis in J Piggott and J Whalley eds New Deyelooments in Apolied
General Equilibrium Analysis Cambridge Cambridge University Press
Ballard C D Fullerton J Shoven and J Whalley 1985 A General Equilibrium Model for
t 43
Tax Eoliey Eyalua~iQO NBER University or Chicago Press
Borges A (1987) bull A Survey of Compula1ional General Equilibnum Models CECD Economic
Slud~ 8
Bovenbig L 19B3 imperfect Capital Mobility in a Perfect Foresight Herberger Model
Mimea UC Br~eley
Bovenberg L 1984 Capitel Accumulation and Capital Immobility Qmiddottheory In a Dynamic General Equilibrium Framework PhD dissertation University of California at Berkeley t
Bovenberg L 19850 Dynamic General Equilibrium Tax Models with Adjustment Costs in A iManne ed bull 1985 EtQnomrc ECUjilbrium Model Formulation and Solution Mathematical
lProgramming Siudy 23 Arnsterdam NorthmiddotHoliand
Bovenberg L t 1986 Capital Income Taxation in Growing Open Eeondmies Journa of Pubtic
Eamp(omlcs 3t pp 347middot376 IBovenberg L ald W Keller 1981 Dynamics and Nonlinearlties in Applied General
Equilibrium Models Internal Report Mlmiddot81middot208 Department for Statistical Methods CBS bull
Vooiburg The Netherlands
Charnley C 1981 The Welfare Cost of C~pltal income Taxation In a Growing Economy
IJournal of Political ECQnQm~ 89 3 pp 468middot496 I
Charnley C 19810 Efficient Tax Reform in a Dynamic Model of General Equilibrium I Mimeo Yale University I Charnley C 1982 On the Optimal Taxation In Stylized Models 01 Inlartemporal General
Equilibrium Mimeo Yale University I Decaluw B and A Manens 1985 Pays en Oeveloppement at Modeles Calculables OEquilibre
General - Une Revue de al litterature empirique Monographie Sene 9 Volume 3 Centre de
Recherche et Developpemont Economlque Unrslt de Montreal
44
bull
I I
I
I
1
Erlich $ V Ginsburgh and L van der Heyden 1987 Wher~ do Real Wage Policies Lead
Belgium bull A General Equilibrium Analysis European Economic Reyiew (forthcoming)
air R and J Taylor 1983 Solution and Maximum Likelihood Estimation of Dynamic
Nonlinear Ratioant Expeclations Models Ecooometrica 51 pp 1169middot1186
Feltenstein A 1984 Money and Bonds In a Disaggregated Ec~nonYin Searl 1-1 J Shoven
Eds Aoolied ~eoeral EqJiliDCiulD tioalYsis Cambridge University Press Cambridge
Feltenstein A 19B6 A Dynamic General Equllibrlu Analysis of Financial Crowding Out
Theory with an Application to Australia Journal of Public Economics 31 No1
Fullerlon D R Gorden 1983 A Reexamination of Tax Distorllons in General Equilibrium
Models in M Feldstein Ed Behayora Simulation Methods in Tax pOlicy Analysis Chicago
University of Chipgo Press
Fvllerlon D J Shaven and J Whalley 1983 Replacing the US Income Tax with a
Progressive Consumption Tax JoulOal of Public ECQnomics 20 pp3middot23
Fullerton D Y Henderson J Shaven 1984 A COmparision of Methodologies in Empirical
General Equilibrium Models of Taxation in Searl H J Shoven Eds ~pQlied Geoaral Jguilibrium
lIoalyss Cambridge University Press Cambridge
Fullerton 0 A King J Shaven and J Whaney 1981 Corporate Tax Integration in the United
Statesmiddot a General Equiiibdufi1 Approach American Economic Review 71 pp~ 677~691
Galper H R Lucke E Tader 1985 Taxation portlor Choice and the Allocation of Capital A
General Equilibrium Approach Mimeo
Gill p W Murray M Saunders M Wright 1986 Users Guide for NPSOL (Version 40) A
Fortran Package for Nonlinear Programming T echnicsl Report SOL 86middot2 Deparlment of
Operations Research Stanford University
Glnsbur~h V and L van der Heyden 1985 General Equilibrium with Wage Rigidities An
45
AppJication to Belgium in A Manne ed t Economic EQumprit1m~ Made Formulation and $oJioo
Mathematical Programming Study 23 Amsterdam NorthmiddotHolland ~
Goulet J 1968 1 Adjustrent CO$1S in the Theory o(the Firm Reyiew of EconomiC Studies 35 bull
pp47middot55
Goulder L 1985 Tax-Financed versus BondmiddotFinancedGhanges in Governent Spending
bull Mimeo Harva(d UniversilYmiddot
Goulder l J Shaven and J Whalley 1983 Domestic Tax Policy and the Foreign Sector The
importance of altemative foreign policy formulations to results from a general equilibrium tax
analysis model In ~n methods in Ibe antico of income from caoia1 ed Martin
Feldstein Chicago University of Chicago riSS
Goulder l L Summers 1987 Tax Policy Asset Prices and 3rowtly ~
A General Equilibrium Analysis NBER Working Pper No 212B
Grandmont J 1982 Temporary General Equilibrium Theory Ch 20 in Handbook 01
Mathemalical Eccnornics Vol II NorthmiddotHoliand
Harberger A 1959 The Corporate Income Tax An empirical Appraisal In lax Revision
Comoendjum Voll House Cornmitee on Ways and Means Washington DC Govemment Printing
Office
Harberger A 1962 The Incidence of the Corporate Income Tax JOULn1 of Political
Economy 70 pp 215middot240
Harberger A 1964 Taxation Resource Aliocalrlon and Welfare fn lb Role of Dir bullbulll and
Indjresl Taxes io tbe Fede[al Reserve s~I~Ill Princeton Princeton Unlveresity Press
James J 1985 New Developments in the Applications of General Equilibrium models to
Economic History in Piggott aand Whalley Eds
Jorgenson D and K Yun 1984 Tax Policy and Capital Allocation Harvard Inslitute of
bull
I
I f
Economic Research WP 1107
Judd K 1965 ShonmiddotRun Analysis of Fiscal Policy in a Simple Perfect Foresight Model
Journal of poJiHcal 5CCJiW1l pp 298middot319 r
Upton D J Potbc J Sachsa nd L Summers 1982 Multiple Shooting in Rational
Expectations Models fuoometrlc 50 1329middot1333
Manne A ed 1985 fconQrnjc Eoullibrium Model FOrIDujaljQo and Solution Mathematical
Pogrammlng Study 23 Amsterdam NorthmiddotHolland
Manne A 1985 Or the Formulation and Solution 01 Economic Equilibrium Models In A
Manne ed ECQOQlli-Jr~ibmiddotum ~tQdel EQlJulailon and Solution Malhematical Programming
Study 23 Amsterdam NorthmiddotHolland
Mansur A and J Whalley 1984 Numerical Specification of Applied General Equilibrium
Models in Scarf H J Shoven Eds Aoplied General Eouillbrium AnaiysectIsect Cambridge Universily
Press Cambridge
Merrill 0 1972 Applications and Extensions of an Algorithm thaI Computes FixedmiddotPoints of
Cenain Upper Semimiddotcominuou Point to Set Mappings PhD Disserlation Unlvel$ity 01 Michigan
Negishi T 1960 Welfar Economics and Exislence I an Equnlbrlum for a Competitive
Economy MelroeCOOQ11ka 12 92middot97
Pereira A 1985 On the Existence and Uniqueness of Optimal Ou1put Path lor CRTS Firms in
Adjustment Costs Technologies Mimeo Stanford University
Pereira A 1985a Consumer Expectations and Ihe COSI 01 UllUty In an Intertemporat
Framework Mlmeo Stanford University
Pereira A 1985b Corporale Tax Integrallon and Government Deficits A Dynamic General
Equilibrium Anaiysls Mlmeo Stanford University
Pereir A 1987 On Ihe Computation of the Users Costs of Stocks Ecooomi (forthcoming)
47
Pereira A bull 1987b Equal Yield Alternatives and Government Oeficlts Mimeof Stanford
University
Pereira A 1987c A Dynamic Applied General ~uHlbrium Model for Tax Policy Evaluation
Comlletamp Docurnent2tion Mimeo University of California San Diego (in progress)
Pereira A 1987d Should the investment Tax Credit be Rei(ltroduced Mlmeo University
of California San Diego (ir progreSs)
Piggott J and J Whalley eds 1985 tJew QeyeQpments in Applied General EQuilibrium
bullenalysjs Cambridge Cambridge University Press
Preckel Pbull 1985 Alternative algorithms for computing economic equilibria In A Manne Ed
Radnor R 19amp3 Equllibrium under Uncertainty In K Arrow M Intrilligaor ods
I ibull
Robinson S 1 g86 Mullisectorai Models of Oeveloping CountrieS a Survey Oivision of
Agricullure and Resources UC Borkeley WP 401
Scarf H 1967 The approximation of fixed points of a continuous mapping ~MM Jouwal of
epplid Mathornallcs 15 pp 1328middot43
Scarf H with T Hansen 1973 00 the Computation of EconomiC Eguilibria New Haven Yale
University Press
Scarf H J Shoven Eds 1984 61l1llied Genera Elujlibrium 6nalyssect Cambridge University
~ Press Cambridge
Shoven J 1976 The Incidence and Efficiency Effects of Taxe on Income from Capital
Jurnal of ramie1 Economy 84 pp 1261-83
Shoven J 1983 Applied General Equilibrium Tax Modemng IMF Slaff PaoerS 30
Shoven J and LB Simon 1986 Share Repurchases and Acquisitions An Analysis of which
Firms Participate NBER Working Paper No 2243
bull
48
IShaven J and J Whalley 1972 ft General Equmbrium Calculation of lhe Effects of
Differential Taxation of Incoma from Capital in the US Jauroe of public Econgmics 1
pp281middot321 (
Shoven J bull and J Whalley 1973 General Equilibrium with Tax$s a computation procedure
and an existence pr00f ReviEhi of Economic Studies 40 pp 475middot490
Shoven J J Whalley 1977 Equal Yield Tax Alternatives A General Equilibrium
Computa~ui0nal Tfchnique J~HJrnal of PUblic EconomIcs 8 pp 211~224 bull
Shaven J J Whalley 1984 Applied GeneralmiddotEquilibrlum MOdels of Taxalion and
International Tr8de An Introduction and Survey Journal of Economic literature 23t pp
1007middot1051
Slemrod J 1980 A General Equilibrium Model of Capital Income Taxation Ph D bullDissertation Harvard Univnrsity
Slemrod J 1983 A General Equilibrium Model of Taxation with Endogenous Financial
Behavior in M Feldstein edbull Behavioral Simulation Methods In Tax Policy Analysis Chicago
UniVerSity of Chicago Press
Stone J 1985 Sequential Optimization and Complementarily Techniques for computing
Economic equilibrium in A Manne Ed
Summers L 1981 Taxation and Corporate Investment a q-Theory Approach BroQkinas
Summers L 1985 Taxation and the Size and ComposHion 01 the Capital Stock An Asset Price
Approach Mlmeo Harvard University
Whalley J 1975 A Geneal Equiloibrium Assessment of the 1973 United Kingdom tax
reform Economica 42 pp 139middot61
1JLTIllOS JIQll~)ltG PAPERS PUBLICADOS bull
I
I I
j
112 70 - LUCENA iJiCIIJo ~ IIEnvironrilent Honitoring and Orqanization Structure T (Setellibro 1987)
nI 71 - L~ENA Diogo A Note on the Representation of Information Stxueture (SeteJObro 1987)
nQ 72 - LUCENA D1CgO Envuorment 111)nitoring and Organization
Structure II (Setelilhro 1987)
ril 73 - AlmmE3 Antonio Pais e CAsPAR Yitor bull Tributacao Inoentivos e InvesUlllentos Analise Qualitativa noYellibro 19S7)
n2 74 - liAlltOSA rose Pedro Opt1lllal lIage R1gidity A SUggestedtlthodology to Teet tb Theory and an Application (Outubro 1987)
nI 75 - CABRAL L 11 B Three lIotes on SYJIIlletric Gemes witt itsYJIIlletric Equilibri (NoveJObro 1987)
n2 76 - LUCENlt D10gOTo Search or Not to Search (Fevereiro1967)
nQ 17 - SA Jorge asconcelos euroI A Theory of SynergyB (Fevereiro1987)
n2 76 - VILARES tanuei Jose Os Be IntenaMios IlIIportados ColIIO Factor de Produao (Julho 1987)
nQ 79 - SA J~rge VasCQncel03 e HoY to Copete and Co~unicate in tature Industrial Products_ (aio 1988)
n2 80 - ROIl Rafael Learning and Capacity Expansion in a New larket Under Uncertainty (Fevereiro 1988)
n2 81 PEREIRA Alfredo llarvilo Survey of Dynamic Computational General EquilibriUlll Hodel for Tax Policy Evaluation (Outubro 1987)
Qualquer in ormacao sabre o 1IIorlltlng Papers ja publlcados sera prestada pelo Secretariado de Apoio aos Docentes podendo os ~e3OSf
ser adquirido5 ne Secyo de Vlndas da Faculdade de EconollLia UNL na Travessa Esteviio Pinto Cltpolide - 1000 LISllOA
_
- ---
--
--------
- - - --------
H n v I taxes
olllng allowed debt Ballardfullamprton~
dynamic objective constrfnt tupendlttire (IlleUs endogenous
aU major balanced trade yoarly ybullbull rly endogenous -f ~Shovon WtlitUy US taxes with C(in~Hlt00 pubHc uility balancod cQmposilon no no debt
eiasticllv junel l 1985 rnagt eCP-REV~d1 aU fn3jO[AndtH3Son vearly balanco small opan Gcon J Swodish inturn fa to cr raturn1987 no - inaia dlJn axogonolJs 00 Ini ITnininod poundgtPfUV
-I L)~ in~h~ --~ ~~ ~pound~~---1- ----shy
runrac h~KQllfkofi ai gulD
1963 1937 US m1l3syo - intertnmponl tixogancuamp yo ys cos~d ampConoroy bandmiddotfinanced flon-oppoundIla
B iii lIa fe-Gould 9r nil llltl1gtr balanced 1rndayearly yearly crcoge)Jus
US t)X6S wiih conslanl I
1963 1965 no public utility balanced cOlYlposWon~ no nD debt max budnt ExP~RCV alastlcHy lunct
Bovenberg simp~iriad 1985middotclosod econybullbull rl~ no - balanced exogenous no no debt US laxes 19a6~two counlry1985 1986 bodgel model
Ert leh~G tn aburghmiddot yeally simplified Inlert walt funet Ishorl~run ttads Heyden no balanced exogenous 00 no Belgium
taxes _ desloullbtium flltUensleln
1967 budaet EXP=REV 6-1mprifia-dmlm costs of yoarly yo rast of tho world
1984 1985 producing a ExTax Rev exogenous financed with yo Australian ~s a Ihlrd conSUmer public good money and bonds non-optimal taxes group
Goulder yaarly annual and endogenous aU maJor
laquo1()Sed economy1985 publlc utilily In1ertemporal composillon bull yes yes U~ taxesY rna versions IlEFElQgtfpoundV bond-financed (non-opllmal balanced trademiddotGoulder-Summer yeary yoarly endogonous bull all major
1987 00 public utility balMc composition no no US taxes with constant max budGel EXPREV alasticilY fund Jl)rgenaon~Yun yoarly aU major
no balanced exogenous no1964 US taxes closed economy budGet inteftemporal recurslve endogenousPereira all major
yo public utility eq of molion lovel and yo yos1995 1987 US laxes closed economy rna for debt composition bondmiddot financed oplloll
- 7 shy
bull - -__----- _ - ------ -- ------- _-----
---- ---
------
------
ll-ltiLt 1 uynamlc Lut MOOSS fOf I ax YOtlcy tv8luallon
Inanell ~ets and maf~9ts ojumLrtm- ----_ ----c_~
privata do btl GlvldtHlI pL~lc aliocalQflo rls shy
1rlc ypa expGct assets equity retention dobl 01 savlnq$ promlums path
Sal ra rd fu Barton phys cap1 used 10 buy soquence lt
Shavon-Wha1ley no financial exogenous oxogenous no physicBl 00 bull 01 sIalic 1E myopic 1985 assets capital flqrlil -
Andersson phys cap used 10 buy 1987 no financial no no yo physical no dynamic ~ perfect
assets capital Auerbach~KolUko If phys cap usod to buy
1983 1987 neo financial no no physical cap no dyoamlc ~ perfectYmiddotmiddotnSSf)ts and Qovt debt SailBrd-Goutder phys cap
1983 1985 no financ[al assets-shy -------
Bovenberg phys cap 1985 1986 no iinenclaf
asso1s Ertlch-Glnsburgh~ phys cap1
-Heyden no financlaJ 1987 assefs
FeltoRsteln phys cap 1984 1986 Invostment is
bondmiddotfinanced
used to buy
exogenous exogenous nObull physl~al no dynamic ~ perfllcl capital
used to buy
no no no physical capital
no dynamic ~onlinuos lima
PI perfect
used 10 buy no no no physical no dynamIc PF perfect
capital jllNt) porlod) bull used 10 buy
no no ye physical cap no dynamic PI p4)rfect and govt debt (Iwo Iods) bull
Goulder phys cap used to buy 1995 no financial exogenous exogenous yenbullbull physical cap no PFlfE bull IdynamIc perfoct~assels and am dabt slaUc
- -~ - shyGoulder-Summers phy$ cap
1987 bonds and no e)(oganOU$ no used 10 buy ye dynamic PF perfect _ bullbullultv bonds MullY (exoaenous
Jorgensonyun phys cap 1984 no financial
essels -~ - ------
Pereira phys cap 1986 1967 bonds end
equHy
________
no
8xogenoos
no
exogenous
no
ye
used to buy two type$ of physical
copltal used 10 buy
go dobl bonds equity
no
no
dynamIc
-
dynamlc
~
lE
perfect
fIbullbulllble (nonperlecB
middot6middot
~-- _middot~ ___~w_ - - ___ bullbullbull___ -------- shy~
-----
------
- ---
-
IlflpnUItll tit lun pVII-y PIlUAlIV
q6H1brla oqual fax polley ovnluaU-on efUclency derlvatton Ioorllhm
calibration param~hr computation COmelfed llold tndfcaHH eHecls
Bnl jard~FuU6rtQ n base year back soll1lion steady-sta1e yo inlsrgroup inerseel or Iii
VS transilion lump-sum with Interfemporai 1965
~Shovn Whallmiddotv replication wllh exog Morrills nmamo1ers alQorilhm _~ond~middotsta1$ ears Inc CVor EV
i~noerucn sleady-sfala no agg(ogaled lnlalsectora 1967
base year back sobHon SiMNON package replication wHn oxog valiant of VS Iransilion no Indicamptor lnlertamporal
paramelsrs multiple shoo ling 6~~~l~slat8
AU8rt)ach~Kotllkot steady-slato yo Inlerguna-r a lions
1893 1llH
quatitaUve by replication a$Sium~lkml Gauss-Soidel liS Uln1WOn any rule -lilh 1111ermpOrB
sonsililJ1y I _ ~poundI~Jjnnelill $t(lfdvmiddot~luiA J3djlt15S ~~ nnrilW ~--------
a ail r nimiddotGould or bas9 9tH bac~ solution Ballard~ staadymiddotslalo )05 kltHSclo(a~
1983 1985 VS Iransllinn lump~sum dynalTlk CV inlerlemporal steadY slafe parameters alaorlthm replicaUon+ with oxog Gourder
+stead~stato (ers loc inlareel10rat BCvonborg qunmaiive by dynamic version stoadY-$Igtto yes 1985 1985 YS transition some govt eN Inlortempora
S9rHltivilv OnearizaJian method ropllcaUon assumptionl of Johansons
+steady-sfata paoo palh Erllch-Glnsbu rgh~ actual vorsus no aootagalod inlersectoral2 bftse years back solu1ion variant of Negishis
Heyden replication wIth alltogbull optlmizaUon countorfaclUal no ~odjcator inlottempora 1987 paramlJters aocroach
faltonstaln baso yeat back solution actual versus no aggrampgalfld intersElclornl replication wilh exog Merrills counlerfaclutll no indlclltor middotlntamprlempOull1984 19B6
parameters al90rithm 1981middot1982
Gouldar staady~slala Inlar5ecioralbase year back souUon Variant or 1S lronsilion no dynamic CV inlortempora
stoady state parameters Newlon 1985 feplicalion+ with iUog Fair~Taylorl
+slaildy-sla1e Intargeoorat Gtlulder~Summers base year back solution varIant of stlJadymiddotstate fixed no aggregated fntersectoral
vs trarlsition government Indicator Intertempo1al steady state oatameters Newton
repticaUon+ with exog Fair-T aylQf 1967 +ste-adv-stata bullbullbullndlno financial
Jorgenon-Yun econometric steady-stale Money melfic In1ersecloral
- - 8slimalion vs traosUian no social weUare functton intertampotal1984 +~teadlmiddotstatl)
~ shy
qtJalitaUv8 by NPSCt optimal path ye generalized inlersectorai 1986 1987 Perelr
replication 8S$umptlonl optimizatIon V$ lumpmiddotsum dynamic lntertemporat se~~ttilitf ____Jllgorilhm opllmal pah pen inc CVoclV financial
- 9 shy
bull ~ ___ - _______ - ____ bull___ OM bullbull __ bullbull __~____ ~_
--
- - - --------
--------- - - - --------
---
1nltJ1 bull I ~ 1J11ltI1 v I ~
-
Sa Ita rd-Futlertonmiddot intargration consumption lax ~StHlven~Whall~y VAT capila~ or labor tax
Ch~mqflS1985 Andersson
consumpion tax neutral corporals lax system 1987
it ue rb a c h~ i(at II Sltotf COIlSUft1ptian lax (poundpird ItS hibor Mx~on ufocto of taxation on capital fOfrnJilti1
dvnamio fiscal policv effects af OQvernmanl defioit (3tlard-Goulder
1883 1987
allocts 01 the adoption of a 1ge3 1985 pure consumpHon lax
im~rtmce of cons fCreSClht -Bovenborg analytically orianiad approach
1985 1986 effects ot pure conumption tax capllal income taxation in Opo economies
Erllch-Glnburghshy-Heyden reat wage policios in Belgium
1987 Felter1steln 1984 - theorotical
1985 ~ linnncial crcwdlng out1984 1986
Goulder effects of tax- vs bond-financed 1985 chanQes In pubtlc expondllufe
~~--
Gouidr~SUmmampfI torpora1$ lax cvts reduced 1987 ITO increase 10 gasolioe taxes
Jorgenson-Yun
effects ot several programs of tax reform on the allocation of ~~ptai
1984
-- Purelra
corporate la)( Integration effects cf romiddotinlroducinQ imreslamnl tax credits
1986 1967
10 bull
bull -~- ~- --- _ - ---____--_ - -----_-_ --~ ---- ~
1 1
2 Brief overview of non-dynamic CGE models
The origin of most or the models we survey goes back to the work of Scarf (1967 1973) whlgt
first developed a reliablo algorithm to compute equilibrium price for a ArrowmiddotDebreu economy
HiS algorithm used simpicial Bubdivision techniques and can -be shown to be the computational analog
of the fixed point theorems previously used to prove the existence of equilibrium His technique
could soW a model with an arbitrary number of consumers and commodities as long as all agents
were price takers consumers were subject to budget constraints demands were continuous andmiddot
production did not display increasing returns to scale The alQorithm while guaranteed to converge I
I I I
was relatively slow for prcb1ems involving more than S1If 20 dimensions A maior improvement in
computational speed Woo offered by Merrills algorithm (1972) which used tpe same fundamental
ideagt as Scarfs procedure
Scarfs model (as is true for the standard Arrow-Debreu model) does not include a government
sector ~ neither taxes nor public goocs At one of the mosi promising appHciUions of the new
computational techniqu w~s in Iha area of tax policy evaluation Shoven and Whalley (1973)
exlended the general equilibrium model and computational approach to include a wide rry of taxes
and a governl11ampnt spending plan The original ShavenmiddotWhalley model was static (although the
different commodi~ies CQuid be considered simitar goods available at different datest as In
Arrow-Debrau) and government was aumad to run a balanced budget The data are arranged as a
social accounling matrix The models spacification and calibration is checked by solving it in the
presence of the base set of laxes The result should ba exactly tha initial social accounting matrix
After having pasd this replication check the model is solved for a counterlactua equilibrium In
the presence of a new tux design The result is once again a social accounting matrix The two
equilibria are compared in order to tSS($S the impact of the new tax plan The first uses ot this
bull
1 Z
model for lax policy evauajon were ShovenWhalley (1972) Whalley (1975) and Shoven
(1916)
Completely static models as Shaven-Whalleys are unsatisfactory for many tax reform ISsues
These include corporate lax Integration effects of investment tax credits effects of accelerated
depreciations consumer Or e)penditure tzxes~ importance of saving subsidies like IRAs etc These
are essentialy dynamiC ksues They involve not only the allocation of capital across sectQfsbut
perhaps more Important the capital intensity of the economy But the capital accumulation and
capital reallocation take time and may involve adjustment costs Because of these issues Ballardmiddot
Fulierton~Shoven~Whaloy took the first steps towards developing a dynamic model
In this conlexl the ShovenWhalley model has been extended and implemented for Ihe US
-economy In BuIIBrdmiddotFulienonmiddotShovenmiddotWhalley (1985) While Ihis book completely documents the
mo~el it was used in several publicalions beginning in 1978 Their mode) consists of nineteem
production sectors twelvo households and fifteen consumer goods It includes a very detailed set of
taxes including the federal and state personallncome taxes federal and state corporate income taxes
SOCial Security taxes pro1J0rty taxes unemployment Insurance excise taxes sales taxes etc It
has been calibl att-d to reproduce the 1973 US economy Recentty a version corresponding to the
1983 JS economy has been developed
wThe 8alard-Fullenon-Shoven Whalley model is dynamic In the sense that consumers face a
choice between current consumption and leisure versus future consumption (which can be
purchased via s wlngs) The consumer classes act as if they were maximizing a nested CES utility
function over the domain of a CES aggregate of contempo-raneous consumer goods and leisure and
future consumption subject to their income constraint The parameters of those funcHons
determine the shares of income devo1ed to each commodity to saving and to the purchase of leisure
They also determine the two key elasticities in the models ~ the elasticity of labor supply with
bull
13
respecl to tgte rbullbull1 aoer lax wage rae and the elasticity of saving with respecl 10 the real after lax- rate of retum to capital
In the model consumerS have myopic expectations regarding lulure prices and in particular
regarding the future rale of return 10 capilal Ballard (1983) and BallardmiddotGoulder (1985)
inco~ora1e both perfect foresight and limited foresight into this model Future consumption is
acquired by buying a fixed composition portfolio of real investments that offer an infinite annuity
of returns
The production side of Ihe model is compielely static The mod1 Incorporates a constantmiddot
olasicily 01 substitution betweeen prlmary inputs in production (capital and labor) and fixed I coefficients for intermediate inputs The model distinguIshes between Industrial outputs and
eonsumer goods for the simple reason Ihat tile data are classified differenUy Industrial sectol$ I I involve such categories as forestry and fisherle metal mino9 and publishing and printing while
consumers purchas-3 fUfHi1ure 2ulomoblles and books This fact is recognized in the model by Ibull
incorporatirg a secord stage Of production which converts industrial outputs intO consumer goods
This technology is uStally mod9-led as a flxed~coefficient conversion matrix
The BallardmiddotFINton-ShovenWhalley model assumeS that the private sector finance
marginai ioveslment with the same composition of debt and equity as currently exists in each sector
This is the same arsumrtion that Herberger originally used Investors all hold debt and equity in tile
same proportion Thampreiore this ownership can be aggregated into simply capital ownership For
tax purposes however the separate treatment of debt and equity is taken into account at both the
corporate and persofa Io=vel SimllarlYJ the dividend poIicies on corporale equitY are established
exogenously There are no government bonds in the model since there are no government deficits
The modol is solved for a sequence (as many as 100) of temporary equilibria wltll consumers
allocating income between present and future conumption at each point in time The palh for 1he
14
economy ~ a sel of connected equilibria The connectIon is pra~ided by capital accumulation
Capital accumulation is endogenous and determined by saving The model stans with a sociagt
accOunting mi1rix In the bzse case the economy is assumed to be in a steaqy state growth -path (along
which all relative prices ~fe constant) The madel solves for both the naw steady state growth path
and the transltion 10 it zf~er a policy intervention rhe authors have frequently addressed the
question-of howong rt tKes to effectively settle un10 a new steady S1ate growth path
The dynamics of the model are limited however in that future consumption is collapsed into a
composite commodity Aso the absence of government deficits and the lack ofproductlon dynamics
limits the realism of the BallardmiddotFuliertonmiddotShoven-Whatley model for the analysis of dynamic
policy issues (such as the adoption 01 a consumption tax or Ihe elimination of the investment tax
credit) The models we survey try to Improve upon Ihe Ballard-FlIiertonmiddotShaven-Whalley
dynamic modelling The models Jn~orporate more forward lookTng behavior They improve the
dynamics on the conampumrtkm side of the economy and some of them introduce true 1ynamlc behavior
on the part oI the producers We turn now to the issues inY9lv_ed in developing such dynamic models
15 I I ~3 ~ssues in the modelling Of dynamic behavior
We will 100k firt at the speclficalion of economic behavior of consume(s~ producers
goyemment and Ihe r~sl of Ihe world The modemog of corporale and household financial decisions
I is then addr5sed Finally Ihe eencept of equilibrium is discussed
I
i
I 1
31 Consumers behavior
Early efforts 10 build dynamic features into the economic behavior 01 consumers are due to
Ballard (1963) and Auerbach-Kotllkoff (1983 1984) Their work has been closely followed by
several subsequent authors In faet dynamic household behavior Is the single most pervasive aspect
amohg the models surveyed
All the models Inc-O1orale some loro of lila-cycle behavior Household behavior is dewmloed
by the maximization of an addWVely separable time invariant intertemporal utillty function The
utility lune1ion is defined oyer the domain of the consumption goods In the economy In most of the
models leisure is also an argument in utility so that labor supply is optimally determined
Utility maximi4ation is subject to a lifetime intertemporal budget constraint which equalizes
the presen1 value of consumers income and expenditure More reeentiy in Andersson (1987) bull
Bovenberg (1985) and Pereira (1986b 1987d) the constraint i$ defined as a sequence 01
recursive equations of motion on wealth This has the potential advantage 01 accomodating liquidity
constraints However It should be recognized that in the absence 01 liquidity constraints (ie when
consumers are free to completely borrow against future income) the two specificallons of the
househofd ccnstraint are essentiay equivatent Furthermore tn both versions saving is optimally
determined as a way of translaring wealth intertemporally
1 6
Some lodels now have a sophisticated dynamic specification of hoosehokl In Ballard (1983) -
BallardmiddotGoulder (1985) and Auerbach-Kotlikof (1983 1984 1987) consumers behavio is
embedded in an iterQEneraional setting Each of them has S5 age cohorts simulataneously alive
Som-e measure of jntergener1ional altruism is considered in the form of bequest motives In Ballard
(1983) and Ballard-Gouldr (1985) consumers derive utility directly from bequthing parl of
their wealth Therefore bequests assume 1he form of a scrap value of terminal wealth of a
generation In AuerbachmiddotKotlikof (1983 1984 1987) each generation empathyz bullbull with Ihe
utility of future generations ovar bequethd wealth In addition Ballard (1ge3) includes detailed
demographic poiections wfill the intent of incorporating in the policy analysis the effects of the
bull Post World War tI baby-boom
In the real world household decisions also include the optimat allo~atiOn of saving among
shyalternative physical or financial assets Theories of household portfolio behavior are relatively
Iwen~establshed However they involve uncertainty as a crucial element Accordingly they have
middotnot been incorpora~ed in the context of the determinIstic dynamic models we sutyey here We will 1 Come backt-O this Issue below
32 Producer behavior
The efforts to build dynamic features into the economic behavior of producers are more recellt
and less widely adopted The first attempts are due to Bovnberg (1984 1985) and Summers
(1965) Dynamic behavior has been more fully incorporated In the recent models of Andersson
(1987) Auerbach-Kotlikoff (19B7) GouldermiddotSummers (1987) and Pereira (1986b 1987d)
Part of the reason why production side dinamlcs has been more slowty adopted is the weak
supply of accepted theories referring to the dynamic behavior of the firms In the models referred
I 17
to above gynamlc producion and investment b~havior are middotinduced by the existence of capital
adjustment costs and linked to Tobins q theory Adjuslment costs are designed to capture both the
incomplete mobility of capital aoross industries and installation costs ~e the costs of adjusting
capital towards its cpHmai level)
Adjllstment costs can be conceived as internal to the firm and measured in terms of foregone
outpul along the lines of Lucas (1967) Alternatively adjustment costs can be viewed as actual
external cests incurred together with Ihe purchase costs along the Unes of Gould (1969) With the
exception of Pereira (198Gb 1937d) all of the CGE authors follow the former approach
The firms in the economy maximize their market value as the present discountad value of the
future stream of dividends In Andersson (1981) and Pereira (198Gb 1987d) firms are seen as
maximizing the present discounted value of net cash flow Maximization is constrained by Ihe
adjustmenl cosl lechnology and ~n equation of motion describing Ihe evolution of the capital stock
It should be noted that undereogenous divldendrelention rules the two problems are equivalent II
should also be nOled Ihat a salisfectory economic rational for th bullbullistoo of dMdonds with the
present tax code is missing in the profession (Shoven (1986))
The models with static formulation of producers behavior are characterized by passive
investment behavior Investment merely accomodates to saving in the eccnomy With a dynamic
formulation induced b adjustment costs real investmen decisions are forward looking Investmenl
Is endogenously and optimally determined by the firms A fundamental difference between the
shonmiddotrun in which capital stock is given and longmiddotrun in which Ihe level of capital is allowed to be
optimally determined is emphasized
This extra richness of produc1ion dynamics is not without costs A careful look at the summary
tables will clearly show an inverse relation between the adoptlon of dynamic features in production
and the level of disagreggation of the production side of the economy In fact with production
bull
j
1 6
dynamics $e dimension of the problems is immensly increased Let IS be more specific
In a static framework with constant returns to scale production technology the output level ismiddot
indeterminate The optimal aliocation of inputs can be obtained by cost minimization with any
feasible output level generatillg zero profits Such zero profit conditions are used to solve for the
output prices in terms of the factor prices and hencemiddot to reduce the dlmensionality of the problem
from the -number of commodities and factors to the number of factors Thus even if the model deals
with 30 production sectors the computation of an economic equilibrium can take place using only
the dimensionality of the primary inputs in the economy
Now under certain regularity conditions on the production and adjustment costs technologies
leading to enough concavity of the optimality objective the intertemporal output path for the firm is
endogenously optimally and uniquely determined even with constant relums to scale technologies
(see Pereira (1985 1987a) Accordin~y with adjustment costs in the model optimal profits
will in general be non-zero This result has crucial implications for the computatiQn of equilibrium
in the model~ with production dynamics With adjustment costs no reduction of dimensionality is
possible The curse of dimensionality returns as a binding constraint
The introduction of adjustment costs adds another significant complication to the model With
capital being less that perfectly mobile in the economy different rates of return on capital will
exist in different sectors This is a difficult problem to tackle conceptually in the absence of
uncertainty Also in the real world producer maximization choices include also its choice of
financial ratios (debVequity) and payout rates (dividendretained earnings) These financial
subjects are difficult and much Clf this behavior is still taken as exogenous by the modellers We
will come back to these issues below
middot i
19
bull 33 government behavior
middot
Two issues dominate the modelling of government behavIor First govemment behavior has
Ibeen typically seen in Ihe CGE liIerature for lax poliCY evaluation as constrained by yearly balanced
I budsets (see for example 8allard-Fullrton-S~ovn-Whafley 098Sraquo The analysis of
govemment defidlS and public debl in a CelE context requires a dynamic selling Second the level of
government expenditures is either exegenouty given as in Auerbach-KoUikofi (1984 1987)
6ovenbrg (1984 1985) Feltenstein (1984 1986) and Jorgenson-Yun (1984) or
endogenously (but no optimally) determined by the balanced budget conditions as in
6allard-Fullerton-Shoven-Whalley (1985) Andersson (1987) Erllch-Ginshurgh-Heyden
~
(1987) Gouder (19aS) and Gouder-Summers (1987) In the second case the composition of
bull public expenditures is often optimally determined However the leve~ of government expendi1ures
can only 00 endogenously and optimally determined if the government is seen as an optimizing agent
and is allowed to run deficits
The first attempts to deal wiU the government defiCits in CGE tax models are due to
Auerbach-Kollilltoff (1984) and Feensain (1984) In Auerbach-Kotlikof (1984) government
expenditures afe exogenous They grow at the rate of growth of population However given 1he tax
structure and tax revenues yearly deficits and surpluses are atTowed subject to an in1ertempond
constraint that the present value of future tax revenues equals the present value of future
expenditures
In Fellenstein (1984 1986) government expenditures are also exogenously given He also
allows the government to run deficits to finance expenditures in excess to 1ax revenues However
surpluses are returned to CQnsumars in the form cf transfers Accordingly government ts not
subject to any constraint regarding the future repayment of public debt
In GouJder (19851 the 9Vernmenl maximizes a static social welfare ~nctio subject to a-
bull
20 I i
I
I I I
I bull
balanced budget conSirtlL This follows the optimal alocatkm of government expenditures alpng
the lines of Baliard-FHeiton-Shoven-Whalley (1985) Ttle model Is generallzea to allow for
exogenous changes in HIe time path of government expenditures Two financing alternatives are
consjdeft~d and contr8s1ed EldditionaJ tax revenues and bond issuance
Pereira (198Gb 1137d) auempts to address both the incorporation of deficits and the
determination of govement expenditures The path of government expenditures and the path of
delicilSlsurpluses (ano therefore the path lor debt) are endogenously and optimally determined The
government is seen as mltJximizing n intertemporal social welfare function given the 13x structure
Optimization is subject to a sequence of recursive equations of motion reflecting the evolution of the
ptlbllc debt allowing fDr government budget imbalances It should be stressathat this specification
01 the government cOMtcalnt is equivalent to the specificaUon in Auerbaeh-Kotlikoff (1983 1987)
when no liquidity constraints affect -gove~nment behavior
The extra richness in the treatment of government behavior allows tfte examination of
government debt poices in a truly dynamic setting Also financial crowding out effects induced by
government deficits can be analyzed (see Auerbach-Kottlikoff (1987) Feltenstein (1986) and
Perer (198Gb 1987e))
34 Foreign sector
Most of the open economy models follow the assumptions of balanced trade with import and
export net demands characterized by constant elasticities along the line of Ballardmiddot
-Fullerton-Shaven-Whalley (1965) Such is the case of Salfard (1963) Ballard-Goulder
(19S5) and Goulder-Sllmmers (1987)_ In Feltenstin (1985) the rest of the world is treated as
21
an addjlion~ consumer group
Bovcoberg (1986) develops a model In which two economies lIr considered each following
intertempoml perlect foresight paths These economies meet in the international forum Their
trade relaroMhips are characterized by yearly balanced trade accounts
None of the new generation of dynamic cae tax models has yet IncorPorated the lnternational
capital flows as done in aoulder-Shoven-Whalley (1983) for the earlier Ballarrshy
-Fullertonmiddot Shaven-Whalley (1985) model This is unfortunate as there is an important
Intertmporal aspect to international lending The first al1empts along Ihese lines are due to
Andersson (1987) and Erlich-Glnsburgh-Heyden (1987) Andersson (1987) in his model of the
Swedish economy adopts a closed economy appraoch in which rates of return in the domesJic
economy are largely determined by the international capital markets Fidinterest rate induce
Intlnatlonal capital flows which determine and finance the International lrade imbalance In
lurn In ErlichGinsburghHeydenmiddot (1987) foreign trade Is generaled according to an
intertemporal Irade welfare lunction with constant import and export elasticities In the shortmiddotrun
they allow inlernational trade inbalances which generate capilal ftows to the domestic households In
the long run however trade balance is assumed
35 The need for financial markets
A dynamic economic Wucture not only provides the ideal environment 10 model many features of
economic behavior it also permits one to Incorporate the financial side of the economy If the
government is allowed to run deficits the question 01 deficll financing automatically follows If
investment is optimally dewmined and returns to capital are different aClOSS sectors problems of
investment financing arise If there are seeral final1cfal assets In the economymiddot government
4-
I i
ibull
I I I I
I
j
22
bonds privpte bonds and equity (or simply physical capital installed inmiddot different sectors) the
problem of allocation of saving among assets with potentially different returns arises
The papers surveyed var greatly with respect to l~e extent of their attention to the financial
side of the economy At Clne extreme are the models in ~anardmiddotFuliertonmiddotShoven-Whalley (1985)
Andersson (1987) Ballltld (1983) Ballard-Goulder (1985) Bovenberg (1985 1986) and
Erlich-Gjnsburgh-Heyden C19B which are devoted exclusively to the real side of the economy
In turn Auerbach-KotliKoff (1984 1987) Feltenstein (1984 1986) and Goulde (1985)
allow for government debt In these models saving finances changes in government debt and physical
capital Private and publlc 2ssets are perceived by the households as perfect substitutes The
allocation of saving merely adjusts to the relative demands for funds
Feltenstein (1984 1986) is the only model surveyed which introduces money Government deficits are financed by issuing money and bonds according to an exogenously given rule Money is
demanded by consumers for transaction motives and an exogenously given fraction as a store of value
On the other hand government bonds and physical capital are the vehicles for the intertemporal
transfer of wealth
Summers (1985) and GOlldermiddotSummers (1987) introduce a whole menu of financial assets shy
firm specific equity capital Different assets earn different rates of retum However such rates
are equal up to constant and exogenous sector speCific risk premia Therefore the introduction of
constant exogenous risk premia as helpful as it may be in the context of calibration does not solve
the main issue of the non-optimality of the allocation of saving Also talking about risk premia in a
deterministic context is somewhat unsatisfactory
Pereira (1986b 1987d) also introduces a whole menu of assets private and public bonds and
firm specific equity However all the assets are expected to yield the same rate of return and
therefore perc~jved as perfect substitutes The different asset types allow consideration of
23
exogenOU$~8tUequjty triO dividendlrc~enlion rufes and therefore seve-ral sources of investment
financing bond equity and retained earnings
The nonmiddotcptimalHy of the allocation of saving and the absence or exogenelty of corporate
financial rules is _a reftectlon of the limitations of the determInistic approach Either in a context of
perfect anticipation oi the future prices Of in general within the t$atm of point price
expectations all the papers follow a deterministic approach Under such circumstances consumers
either expect diHerent rales of return (inclusive of risk premium) across assets in which case
they wilt buy only one asset (that with highest rate) Of they expect equal rat~s of relurn in which
case Ihey are indifferent about the asset composition of Iheir portfoliO There is no way of
Iradingmiddotoff rales of return and risks to obtain an optimal interiof solution to the problem of the
allocation of saving
The most advanced contribution in the modelling of saving allocation in a CGE setting is due to
Slemrod (1980 1983) in the context of a static one-period model In his model consumers act
according to a two stage separable decision process They Hrsl decide on how much to save Then
they decide on the allocation of saving according to an Indirect uttlity function dependent on the rates
of ratum and variances offered by the different assets in the eoonomy The sourCe of riskiness in the
ampConory comes from an uncertain marginal product of capital On the other hand aside from
portfolio decisions the res of the economy is insulated from uncertainty
The most complete contribution in terms of the treatment of the corporate financial rules is
FulienonmiddotGordon (1983) In a variant 01 Ihe BailardmiddotFulienon-ShovenmiddotWhelley (1985) model
Fullerton and Gordon have capital intensity and optimal financial decisions Jointly deter~ined
through a two siage process The cosl of financing capital is minimized by trading off the tax
advantages of debt against the epected real bankrupcy coSls inherenl with high debVequity ratios
Given the optimal debtequity ratio the level of inveSlment Is chosen such that at Ihe margin the
24 l
bull
rerurn pn ~uity equas the return on bonds plus an exogenous risk premium
36 Market assumptions equilibrium and expectations
Virtually all of the papers arc characterized by Walrasian market clearing assumptions All
markets are perlect1y competi1ive Atomistic competition among agents 1s assumed even thoughmiddot only
a flnile number of agents is considered Virtually no mark~t disequilibria or price stickiness afe
considered The only exception is ErlichmiddotGinsburghmiddotHeyden (1987) bull In thei model of the Belgium
economy I the _wage rate is fixed in the short-run Therefore in the shOrlwrun desequilibrium in
the labor market will generate endogenous unemployment However in lh$ long-run all prices
Including the wage rata are Ilexible and accordingly ali markets ciear
Given the dynamic nature of behavlor in the economy markelmiddotcfearing prices in each period
depend on expectations of fture prices and on tax variables In the economy There are essentially
two ways at interpreting the economic equilibrium in such a dynamic conte~bull If future prices are
perfectly anticipated (io expectations are self-fulfilling) a perfect foresight equilibrium
prevaHs Then future actions are merely the implementation of current decisions for future
periods However jf ptica expecla110ns are not perfect (Le agents make mistakes w1th re$pect 10
future prices) then a temporary or shortrun equilibrium prevails Markets ciear and clearing
prices depend on fu1ure price expectations Current plans about the future are typically not
precisely implemented They will be revised as more or better information becomes available to the
economic agents See Grandmant (1982) for a comprehensive survey 01 the temporary equilibrium
literature
Wilh the exception of Gould (1985) and Pereira (19S6b 19S7d) all the models surveyed
adopt th~ concept of a perfect foresight equilibrium In turn BaliardmiddotGouldr (1965) considers a
J
2S
flexible am9unt of foresight In terms of the number of years over which price movements are ~
foreseen Pereiras model (1986b 1987b)) is flexible in a somewhat different way in that it can
indude any range of foresight from myopia to perfect foresight The choice between the perfct foresight and lemporary equilibrium is ultimately to be made on
philosophic grounds It can be argued that less than perfect expectations Imply that agents are
irralional in some way (see Auerbach-Kctiikoff (1987 p 10) However the reverse argument can
be made One can question wiether agents are reaHy rationar and perfectly knowledgable about
future prices
Recent evidene of Balird (1987) Ballard-Goulder (19851 Goulder (19851 and Pereira
(1986b 1987d) suggests that the choice in modemng expectations is an important one They show
that the degree of foresight inl0 thc future (ranging from perfect foresight to-myopic expectations)
may have dramatic impacts on the pollcy conclusions of Ihe model Accordingly the best research
strategy may b to design models which are flexible enough 10 allow for different rules regarding the
formation of expectations With the exception of the articles just mentioned sensitivity analysis
bullhave previously not been performed along this dimension
In terms of implementation the two concepts of equilibrium - perfect foresight and temporary
equilibrium have different implications The dimensionality of the equ1ibrium soll1ion algorIthm
is involved Suppose we have a model with ten markets to be run for a period of 50 years Aside
from normalization a perfect foresight model implies computing prices In 500 dimensions while a
-t~mporary equilibrium model requires sorving SO equilibria each in ten dimensions Given that
computational speed often varies with the cube of the number of dimensions the lemporary
equilibrium formulation is potentially strikingly more feasible However Ballard (1987) and
Goulder-Ballard (1985) have devoloped techniques to greatly speed the computation of a perfect
foresight equilibrium
26
j
I
4 Jrnplemenlalion and Issues on policy evaluallon
41 Calibration and equilibrium comparlslon
CGE models at typically parameterized by the use of a calibration procedure Some parameters
are exogenously given However some crucial parameters are determined in such a way that the
model replicates the data for a given base year See MansurmiddotWhalleY 11984) for an eXlensive
discussion of 1~i$ issue
CalibratIon in a dynamic context is generaliy interpreted as requiring two properties First
replication of base year data is required Second the model is parameterized to simulate an
intertemporal balanced growth path when Ihe base policy is maintaineq This Is the approach
followed by Ballard (1983) BallardmiddotGoulder (1985) Goulder (1985) and SummersmiddotGoulde
(1986) It follows the practice of BallardmiddotFuliertonmiddotShovenmiddotWhalley (1985) with their
sequential cquilibrium dynamics
Other authors Auerbach-Kotlikof (1983 1984 1987) Boyenberg (1984 1985 1986)
and Pereira (198Gb 1987d) follow whal we cali a qualitalive calibration The slructural
palametars are exogenously chosen so that lh economy follows a reasonable path into Ihe fulure
ThaI has 10 do wllh the fact thaI given the recursive nature 01 the dynamic economy and aside tom
such structural parameters only iniUa stock values are needed to run these models Given initial
conditions on the stocks of say private wealth capltal and government debt agents will optimize
and Ihereby generate a lirst round of net demands and equilibrium conditions In lurn the
equilibrium prices will determine the evolution of the stock variables into the next perIod
There are several potential problems with calibration in the conlext of dynamic models The
assumption of a steadymiddotstate growth ~ath in the base case can be questioned First while
27 I Ibullbull
steadymiddotstaU is a possibity it certainly is not the only meaningfull solution to dynamic models
Even in the case of a perfect foresight equilibrium the model implies an equilibrium path which i may or may not involve colanced growth The model not the modeller should dictate the nature 01 I themiddotbase case path Second in the context of a temporarY equilibrium path a steady state solutlcn is Inol a likely model QutcOn1tl In fact unlessmiddot expectations are stattc short run behavior consistent
with a steady-state evolution will in general not be generated On the other hand if static I expectations are self~fufjili[1g we have in fact a perfect foresight model Third even the base year
replication requirement may cause problems In Ihe cOntexl of temporary equilibrium Any
calibration parameter would be conditional on expectation rules which is probably an undesirable
feature
The nalur of the two-requirement calibration strategy is very muchin the spirit of the bull
traditional design of the comparision of alternative equilibria comparis_ion between a steady-state
base case on one hand and alternative paths including a transition period and a finat sleadyyenstate on
the other hand Pereira (198Gb 1987d) compare different (not necessarily steadystate)
equilibrium paths The arguments against such a procedure are based on the idea that the impact of
policy changes can be observed most easily since all departures from the steadymiddotstate can be
attributed to the altemalive polioy On the other hand the base case so dEfined as a steadymiddotstate is
consist~nl with previous work in a less dynamtc setting and Iherefore allows a common standard
for comparing model results
42 Computation algorithms
We are still al a stage in which basically each author uses a different computational technique
Th development of dynamic models - in parlicular wilh adjustment costs andior perfect foresight
bull
28
has corresponded with the decline in the use of fixed point algorithms In -fact given the relative
arge dimensions inevitably involved such algorithms lend to be very inefficient at the best and
chen prohibitively stow See Stone (1985) and Preckel (1965) for a comparative assessment of
different computation techniques Among the models surveyed only Ball1rdFulle(ton~
-ShavenmiddotWhalley (1985) and Feltensteln 1985 use Merrills variant of the fixed point
algorithm 1echnique
AuerbachmiddotKoHikoff (1963 1984 1967) follow a three stage procedure They first compute a
base case steady-state then a revised case steady state and finally transition path for the economy
berveen these two steady-states tn all stages a Gauss-Seidel iterative procedure is used
Ballard (1982) Ballard-Goulder (1965) Goulder (HiSS) and Summers-Goulder (1986)
~
use a method developed by Ballard and Goutder which is similar in many awects to the FairmiddotTaylQr bull
(1983) algorl1hm Short-run equilibria are calculated (using Merrills algorithm) parametric on
nrice expeC~lions The model is then iterated to generate self-fulfilling intertemporaf expectations
and the corresponding perfect foresight equilibrium In a relatively similar approach Andersson
1987 uses a simulation program SIMNON developed In the University of Lund This program
can handle two-paint boundary problems In a fashion consistent with the multiple shooting
algcrlthm (see Lipton-Poterba-Sachs-Eummers (1982))
Boyenbergs computational approach (1985 1985) differs from the other models surveyed in
that he relies heavily on analytical techniques Computations are done by using a dynamic version of
Johansons linearization method Being essentially determined by the continuous time nature of the
I I
f
model thiS linearization model has the disadvantage of confining the analysls to infinltesinal changes
around the base case equilibrium (see Bovenberg (1985) p 53) Pereira (1986b 1987d) uses an optimization algorithm NPSOL - developped by
rGillMurray-Saunders-Wrlght (1986)) This algorithm is used 10 compute the sequence of
I
29
Shortmiddotrun temporary equilibrium which makes up the intertemporal equilibrium path The
equilibrium conditions are seen as nonlinear equality constraints in the minimization of an artlcial
objective function The prices are normalized to the unit simplex by an additional linear equality I I 1 bull constraint
II ErlichmiddotGinsburgh-Heyden (1987) follow a unique approach in that they use a variant of the
optimization technique introduced by Negishl (1960) The economic equilibrium can be generated i
as a solution of a mathemallcal program Ihe objective function of which is a weighted sum of the
utility functions of the various agents while tho constraints set consists of the market dearing 1
conditions Ginsburgh-Heyden (1985) have extended Negishis resulllO tho case of downward price
lrigidities I
Finally the paper by Jorgonson-Yun (1984) is also unique in tIlat it is the only I
econometrically estimated model Different blocks for the consumption and production Side of the
-
model are separately estimated to provide the necessary structural parameters i
I The diversity of computation techniques is yet another indicator of the exploratory nature of the
body of literature surveyed in this article
43 Equal yield comparisions
The link between a base case and counteriactual simulations Is usually provided by the concept of
equal yield Shaven-Whaley (1977) discuss the meaning 01 equal yield in a general equilibrium
context when government is not allowed to run defICits Equal yield is interpreted to moan constant
public utility Government base case utility is maintained in the counferfactual experiments With
balanced budgets the concepl of equal yield is unambiguous The new equUibrium prices and the
balanced budge condition will del ermine the minimal expendilure and taxes needed to maintain base
30
case publi~ utility Accordingly in general equal yield is inconsistent with equal nominal tax
revenue Some change in tax revenue Is necessary Different tax replacement schemes ate
considered to assure that enough tax revenue is collected This is the approach essentially followed
by most 01 the papers surveyed Only Feltenstein (1985) Goulder (198) and Jorgenson-Yun
(1984) chose not to follow an equal yield strategy
The question is of how to Interprete lila concept of equal yield when the government is alowed to
fUn deflcLts The optimal leve of expenditure for base ease public trliUty can now be taAinanced
bond financed or financed by a mix of bonds and taxation We have several versions of equal yield In
particular equal yield may now be consistent with equal tax revenUE Furthermore some meaSure
ltgtf financial crowding out effects of government deficns can be inferred from the comparislon of the
several equal yield alternatives These issues are extensively discussed ~Pereira (1987b)
44 Price expectations and the dynamic generalization of compensation
Indicators
The sum of equivalent and compensation variations over households is the most widely used
aggregated measure of efficiency gains or losses In the context of $teady~state eomparisions andor
when complete future markets exist andlor perfect foresighl is assumed there are no difficulties
associated with the use of the standard Hicksian indicators In fact correct future prices are known
in these cases
If expectations are not selfmiddot fulfilling andlor future markels are not available a dynamic
generalization is necessary 8allardmiddotGoulder (1985) provide some steps in that direction by
defining an indicator that accomodates periods far into the future when households do not have
perfect foresight but a steady~state prevails On he other hand this issue is more fundamental in
i
[
I
i I I I I [
I I
bullI
31
the contex~ of a Gner1 lemporary equilibrium ftamawor~ In such Circumstances Pereira
(1986a) dev~lops a cnamlc generalization which is obtained as the present discounted value of a i i
sequence of short-fUn optimal expenditure functions consistent with a base case expected future I stream of IlWities I
32
5 Empirical evidenee from selected policy Issues
Dynamic tax models have generated several important results which escaped Sialic modellers
The following is a seleCpoundd set of issuos whIch stress the marginal benefits of dynam~ modelHng of
economic behavior in innovltlttive areas of analysis The discussion of a consumption tax emphasizes
the benefits of dynamic tusehold behavior and intergenerational aspects The study of the
flliminatlon or the re~intToduction of the investsnemt tax credits is made meaningful by the dynamic
modelling of production betlavor Modelling governmenl in a dynamic conlexts allows the modeller
to study the irrrpact of government deflcits Finally a dynamic framework lets us appreCiate the I
relevance of consumer expectations in terms of the evaluation of policy alternatives
i
i
51 Consumption Tax
t
By Its very nature a consumption tax can not be adequately investigated with a static model bull
Fullerion-Shoven-Whaicy (1983) use Ihe model of the US economy s described in
Ballard-Fullerton-Shoven-Whalley (1935) to evaluate the movement from the currenl US tax Isys1em to a progressive consumption tax Since their modeJ Incorporates a Jaborleisure choice
where leisure is an untaxed commodlty their results reflect the fact that both the consumption tax
and Ihe present system are dslortionary
Concentrating on the rntertemporaf distortions they show that sheltering more saving from lhe
current US income tax could improve econornic efficiency even if marginal tax rate increases are
necessary in order to maintain government fevenue~ At first you have a revenue shortfall
However the economy moves to a higher sustatned growth path and uitimatefy lower tax rates
cangenerate the same revenue path Also wages increase in the long run with this policy The
bull
bullbull ~
33 shyt
i
I
I
present value of welfare gi[lS for a potiey of complete savin9 deduction wilh marginal tate
adjustments (consumption lax) is simulated to be around $500 billion to $600 billion 01 1973
dollars
They also investiga1e th 1ngth of ttme it takes th~ economy to adjust to these policy changes
Roughly they estimate the long run to b thirty years although this figure Is very sensitive to the
specification of the savings elasticity - a crucial parameler in this model
52 Investment T~x Credit
Goulder-$ummers (1 S87) address the impact of eliminating the lnvestmenl Tax Cradit (ITC) on
intersectoral capital formation and on economic growth Their model isYa1ticularly adequate to
addies$ this issu~ in that is postulates a forward looking investment m9we with adjustment costs
They show that the eliminating the ITC causes a reduction in the rale of investment Inveslment is
estimated to fall by about 7 in the short-run and by about 12 in Ihe new long run steady state
In lurn a previous policy anncuncement lowers Ihe overall al1ractiveness of Investment and leads 10
a downward shift of the Investrrent profile On the other hand the combined effect of a revenue
neutral simultaneous efimiraron of the ITC and reduction of the corporate tax rates is a long run
reduction ef the capital Sieck by 35 This pattern suggests that a revenue neutral increament in
both Ihe corporale tax rates and the inveSlment lax credits would be preferable 10 the formula
adopted in Ihe US Tax Reform Act of 1986
Pereira (1987d) addresses the impact of the ITe from Ihe oland point of Ihe curTenllax syslem
under the Tax Reform Act of 1986 Given Ihe concern over the potential depressiVe effects upon
savings and investment of the new tax system in conjunction with deficit reduC1ion mechanisms of
the GrammmiddotRudman type a pertinent question is should Ih lTC be e-introduced In the context
bull
34
-of his mod~1 of the US economy Pereira focuses on the trade-off between the distortion in the
intersectoral allocation of capital induced by the lTC the lower tax revenues it generates and
potential financial crowding-out effects on one hand and the positive effects of lowering the relative
price of new capital goods on the other hand Preliminary results confirm the qualitative results in
Goulder-Summers (1987) suggestting a welfare gain of about 2 of the present value of GNP in
the best scenario of absence of replacement taxes Furthermore preliminary results show an
important time pattern to welfare gains with the average benefits increasing the further you look
into the future This pattern is due to the presence of constraints to intersectoral mobility of capital
and inertia in the adjustment towards the optimal capital levels as reflected by the presence of
adjustment costs
I
I
bullAuerbach-Katlikaff (1gB) in the context of their intergenerational model of the US economy
consider the impact on savings capital formation and interest rates of deficit policies and balanced
budget increases in government consumption They conclude that deficit finance and government
consumption can significa-ntly crowd out capital formation and lower the welfare of future
generations However crowding out from deficit finance is a very slow process because it results
from increased government spending over potentially long horizons Also deficit policies may
substantially influence the longmiddotrun interest rates while leaving the short-run rates essentially
unaffected Finally and in opposition to the central role adjustment costs seem to play in both
53 Financial crowding outeff~cts of government deficits
Goulder-Summers and Pereiras models Auerbach-Kotlikoff suggest that the time path of interest
rates induced by a policy of deficit financing seems to be insensitive to the presence of adjustment
costs
bull
35
54 The role 01 consumer expectatlons
The expectations analysis has uncovered one of the benefits of dynamic modelling
Ballard-Goulder (1985) find that the appeal of adopting a consumption tax depends on the level of
foresight possessed by the consumers Furthermore additional foresight may be welfare worsening
This is a second best type of result This tends 10 occur under policies that lead to capital deepening
and declining rate of return to capital over time To the extent that consumers have more foresight
they will be beller equippod to anticipate the fall in the rental price of capital Consequently if a
saving incentive is enacted people save more with myopia than with perfect foresight Given the
exi~tence of taxes on capita and the discrepancy between the private an~--sccial returns to capital
the greater savings with myopia is beller socially Ballard-Goulder find that the welfare gains
from a consumption tax are reduced by about 10 when we move from myopia to farleet foresight
Therefore the attractivenness of adopting a consumption tax seems fairly [obust across these
foresight specifications
I
bull
36
6 Tbe power and weaknesses 01 dynamic modelling the example 01 corporale
tax Integrallon
The Issue of corporate tax integration allows us to sh~w the stre-ngths and weaknesses or the
dynamic approach Empirical evidence using CGE techniques indicates that depending on the precise
scheme of integration and the 1ax replacement methods integratiOn may have substantial effects In
Ihe work of FuliertonmiddotKingmiddotShoven-Whalley (1980 1981) total integration was found to yield an bull
annual static eHielency gain of $4 to $8 billion in 1973 dollars Simulated dynamic gains may be as
large as $695 billion or abcu 14 of the present value of tulUre consumption and leisure in the
US economy These gains result primarily from interindustry reallocalions of Investment and an
improved inter~emporal allocation of consumption
Mora recent work emphasizes hOw consumers asset portfolio decisions and firms finanCial
decisions affect the efficiency gains from integration Slemrod (1980) focusing on consumers
asset ponfoio deCisions finds static efficiency gains which are about twice -as Iarge as those
repOrted iiy Fulierton-King-ShovenWhalley (1961) FuUrlon-Gurdon (1983) focus on the
firms financial decisions They roporl efficiency gains of 6 of GNP from the elimination of the
tax distor1ions favoring debt However when they eliminate the corporate tax and repiace it with
increaSed persona income taxes addlHonal dls10t1ions are created in the optimal labot~leisure
decisions These distortions tend to dominate the analysis slJ(h that the overall effects of complete
integration are very modest Galpr-LuckmiddotTodr (1986) address simultaneously consumers
asset pOrlfollo decisions and the firms financial decisions They also find very modest efficiency
gains from integralion
The above results are important but may be severely biased There are severa aspects of
economic behavior and modelling crucial for the study of Income tax Integration whiCh have no been
i 37
captured innny of the above papers One aspect is the absence of government deficitsmiddot a balanced
budget is assumed It is true Ihal resource crowding OUI is caplured In these mOdels but financial
crowding out induced by government spending is nOI Second investment is not derived from
optimtztlcn behavior Investment behavior passively accomodates endogenous saving decisions As
a consequence the differential impacts of policies in the incentives to save and to invest are not
captured Aso full capital mobility across secors is assumed with instantaneous capital
adjustmenlS towards optimal levels This assumption rules qut different costs of capital across
seclors and therefore differentiated reactions to tax policies changes Finallyhe modelling of both
gov9rnmerlt deficits and of endogenous real and financial investment decisions necessitates the
~nsideraiion of a dynamilt framework and the introduction of financiaJ assets govemment bonds and
iprivate financial assets A dynamic framework also highlights the efficiency effects of Integration on r
ithe ptimal intertemporal decisions The above models are either sIalic (Slemrods and
GalpermiddotltJckemiddotTode(sj or a dynamic sequ~nce of otherwise Slatic model bullbull I Pereir a (1986b) develops a dynamic applied general equilibrium model bull to study the tbull
efficiency effects of integration on the growth and allocation of investmenl across sectors Special
Iattention is paid to the structural affeelS of resource and financial crowding oul induced by
gove mect spending and deficits and to the real and financial Investment reactions across industries
to both tax integration and flnancial crowding out His mode departs from previous work on income
lax Integraticn and for that matler from most 01 the CGIE literalure for tax policy evaluation in
several direcllons It encompasses an endogenous sequential equilibrium struelure founded on
dynamic behavior with flexible expectations Government deficits are optlmaUy determined
Investment decisions are tward looking and the resull of optimizing behavior Several financial
assetsmiddot public and privatemiddot are considered
Simulation resulls Indicale that the welfare gains from Integration under large deficits are at
32
best lJlodest when measured in terms of the GNP The elimination of the corporate tax and its I replacemenl by increased income lax rates yields long run benefits which are never larger than
2 of the presenl yalue ct telr consumption and leisure onder the previous tax regime and 1
under the current tax regime However the short run w~lare effects of full integration tend 10 be
very low and in in some scenarios even negative This is a flew intertemporar pattern of efficiency Ieffects which reflects an adjJsiment lag in the interindustry investment decisions due to the
existence of costs of adjustment On the of her hand partial integration achieved by excluding I dividends from the corporale lax base syslematically yields negalive efficlencyeffeclS In Ihe long Inm these can be as high as 3 of present value of the future value of consumption and leisure This
is a new second best effect suggesllng Ihat less than complete integration may have perverse I efficiency effects
The dynamic slructure of PerclrBs model snowing for an improved freatment of real investment
declsions and government deficts provides a potential setting for a more accurate measure of the
costs and benefits of inlegration The simulation resullS suggest lower benefits and an intertempora Ibull
pattern of growing benefits
Simulation results also clearly suggest robust negative effects from partial integration How
reliable is this result If dJvidends were deductible from the corporate lax base (as are interest
payments) the preferenlial Ireatmant of debl over equily would be ellminaled Corporallons should
be expecled 10 react by decreasing the optimal deWeqully rallo However corporale financial
gecisions are exogenously set In Pereiras model as in aft of the other models surveyed that go as far
as dealing with the issue
Also withdrawing dividends from Ihe corporate tax base is a way of eliminating Ihe double
taxation of dividends at belh Ihe corporate and personnal Income levels How will the optimal
aliocellon of household saving accomodate that change in Ihe relalive return to corporale equity
39 I
t However hOusehold portfolio decisions are exogenouly set in Pereiras model as in all of the other
models surveyeltj go as far as dealing with the issue
IIt is sate to say that the evaluation 01 the benefits 01 partial Integration as defined above are
sevrey underestima1eO Meanwhile the distortionSoenerated by the increase in the marginal
personal tax rates designed to ralse the revenue foregone by the dividend exclusion are fully
aCCltlunted for A good measure of the costs of integration together With a poor evaluation of the
benefils may very well be the reason why partial integration is Simulated 10 yield neg alive
eHikiency gains
On a different vein as most of Ihe model surveyed here Pereiras is a closed economy model It
is legitimate to wonder how the resullS would change If international capital flows were anowed
IThis almost certainly would affect financial crowding out since a good d~al of the capital inflows
could be used to finance public debt
For these various reasons the results should be treated as preliminary but they strongly I suggosl that the dynamic structure provides valuable new insights into the corporate tax integration SSU2
40
7 Concluding remarks I I I i
What has been acccmplished Ihis far The success of Ihe CGE research in laelding the challenge
of dynamic modelling can not be denied
I iGreat progress in the modelling of household behavior has been achieved Promising
developm~nts in the modeUing of production and government behavior have also been accomplished
Interesting innovations in the concept of equilibrium and computation techniques were discussed
The policy analysis was greatly enriched by considering transitional effects together with longer
run steadymiddotstate equilibrium paths generalized equal yield strategies accomodating different
financial crowding out impacts and generalized dynamic policy evaluation indicators
An important set of issues have been addresed by the models surveyed Traditional issues
focusing on the effects of capital taxation and consumption taxes have been pursued In terms of
capital taxation for example the intertemporal nature of the issue was enhanced by allowing an
optimal evolution of capital stock in the economy and forward looking optimal investment decisions
In Andersson (1987) GouldermiddotSummers (1987) and Pereira (1986b 1987d) this is coupled
with a improved treatment of several tax provisions like investment tax credits and depreciation
allowances In terms of the consumption taxes developments in the specification of Intertemporal
household behavior the introduction of overlapping generations and the modelling of
intergenerational links via bequest motives as well as a closer attention to demographic evolution
provided a much improved economic setting for the understanding of the several aspects of the
problem
Furthermore dynamic modelling has permitted the addressing of a set of new issues Policy
Issues ranging from dynamic tax policy analysis to the evaluation of exogenous changes in
government expenditures to the impact of different methods of financing government expenditure to
4 1
the importance of tinanciai croHrii1g out to 1he relevance of consumorS expectations in tne policy
results have been addressed
Despile all of the progress and in part due to such progress several avenues are wide open for
much neded additional research The following seem to be the most Interesting and promising
areasmiddot First the specincation of liquidity constraints to household behavior may help to obtain a
better understanding of policy chenges that affect directly the interest rates in the economy Second
major attention needs to be paid to the incorporafion of a roraign sector (with open capital markets)
Third some efforts are needed in terms of finding adequate computation techniques now that
dynamic modelling eally makes the curse of dimensionality worse than eVOrbull
The modelling of financial markets is the single most unsatisfactory speet of the dynamic
models surveyed here~ The problems associated with provlding a mo~ adequate treatment of
financial markets and in general endogenous and optimal financial deci$ions bull both at the corporate
and at the personal levelsmiddot have 10 be addressed That necessarily requires Introducing uncertainty
into the models To b adequate this must go well beyond the mere assumption of point expeetalions
with the whole array of modelling implementalion and evaluation problems thai generates
bull I i
42
REFERENCES
Abel A and O Blanchard 1983 An Intertemporal Model of Saving and Interest Econometrica
51 pp 675-92
~ Andersson K 1987 Tar-ation of Capital in Sweden - A General Equilibrium Model Ph D
Dissenation UnIversity of Lund Sweder
Auerbach A L KotlikoH 1983 National Savings Economic Welfare and the Structure of
Taxation in M Feldstein ed Behavioral Simulation Methods in Tax Policy- Analysis Chicago
University of Chicago Press
Auerbach A l Kotlikoff 1984 Social Security and the Economics of Demographic-
Transition in H Aaron and G Burtless eds Retirement and Economic Behavior Washington DC
Brookings Institution
Auerbach A L Kotlikoff 1987 Dynamic Fiscal Policy Cambridge Cambridge University
Press
Auerbach A L Kotlikof and J Skinner 1983 The Efficiency Gains from Dynamic Tax
Reform Internatiooaj Economic Reyiew 24 pp 81middot100
Ballard C 1983 Evaluation of the Consumption Tax with Dynamic General Equilibrium
Models Ph D Dissertation Stanford University
Ballard C 1987 Tax Policy and Consumer Foresight A General Equilibrium Simulation
Study Economic Inquiry 25 pp 267-284
Ballard C and L Goulder 1985 Consumption Taxes ForeSight and Welfare a Computatonal
General Equilibrium Analysis in J Piggott and J Whalley eds New Deyelooments in Apolied
General Equilibrium Analysis Cambridge Cambridge University Press
Ballard C D Fullerton J Shoven and J Whalley 1985 A General Equilibrium Model for
t 43
Tax Eoliey Eyalua~iQO NBER University or Chicago Press
Borges A (1987) bull A Survey of Compula1ional General Equilibnum Models CECD Economic
Slud~ 8
Bovenbig L 19B3 imperfect Capital Mobility in a Perfect Foresight Herberger Model
Mimea UC Br~eley
Bovenberg L 1984 Capitel Accumulation and Capital Immobility Qmiddottheory In a Dynamic General Equilibrium Framework PhD dissertation University of California at Berkeley t
Bovenberg L 19850 Dynamic General Equilibrium Tax Models with Adjustment Costs in A iManne ed bull 1985 EtQnomrc ECUjilbrium Model Formulation and Solution Mathematical
lProgramming Siudy 23 Arnsterdam NorthmiddotHoliand
Bovenberg L t 1986 Capital Income Taxation in Growing Open Eeondmies Journa of Pubtic
Eamp(omlcs 3t pp 347middot376 IBovenberg L ald W Keller 1981 Dynamics and Nonlinearlties in Applied General
Equilibrium Models Internal Report Mlmiddot81middot208 Department for Statistical Methods CBS bull
Vooiburg The Netherlands
Charnley C 1981 The Welfare Cost of C~pltal income Taxation In a Growing Economy
IJournal of Political ECQnQm~ 89 3 pp 468middot496 I
Charnley C 19810 Efficient Tax Reform in a Dynamic Model of General Equilibrium I Mimeo Yale University I Charnley C 1982 On the Optimal Taxation In Stylized Models 01 Inlartemporal General
Equilibrium Mimeo Yale University I Decaluw B and A Manens 1985 Pays en Oeveloppement at Modeles Calculables OEquilibre
General - Une Revue de al litterature empirique Monographie Sene 9 Volume 3 Centre de
Recherche et Developpemont Economlque Unrslt de Montreal
44
bull
I I
I
I
1
Erlich $ V Ginsburgh and L van der Heyden 1987 Wher~ do Real Wage Policies Lead
Belgium bull A General Equilibrium Analysis European Economic Reyiew (forthcoming)
air R and J Taylor 1983 Solution and Maximum Likelihood Estimation of Dynamic
Nonlinear Ratioant Expeclations Models Ecooometrica 51 pp 1169middot1186
Feltenstein A 1984 Money and Bonds In a Disaggregated Ec~nonYin Searl 1-1 J Shoven
Eds Aoolied ~eoeral EqJiliDCiulD tioalYsis Cambridge University Press Cambridge
Feltenstein A 19B6 A Dynamic General Equllibrlu Analysis of Financial Crowding Out
Theory with an Application to Australia Journal of Public Economics 31 No1
Fullerlon D R Gorden 1983 A Reexamination of Tax Distorllons in General Equilibrium
Models in M Feldstein Ed Behayora Simulation Methods in Tax pOlicy Analysis Chicago
University of Chipgo Press
Fvllerlon D J Shaven and J Whalley 1983 Replacing the US Income Tax with a
Progressive Consumption Tax JoulOal of Public ECQnomics 20 pp3middot23
Fullerton D Y Henderson J Shaven 1984 A COmparision of Methodologies in Empirical
General Equilibrium Models of Taxation in Searl H J Shoven Eds ~pQlied Geoaral Jguilibrium
lIoalyss Cambridge University Press Cambridge
Fullerton 0 A King J Shaven and J Whaney 1981 Corporate Tax Integration in the United
Statesmiddot a General Equiiibdufi1 Approach American Economic Review 71 pp~ 677~691
Galper H R Lucke E Tader 1985 Taxation portlor Choice and the Allocation of Capital A
General Equilibrium Approach Mimeo
Gill p W Murray M Saunders M Wright 1986 Users Guide for NPSOL (Version 40) A
Fortran Package for Nonlinear Programming T echnicsl Report SOL 86middot2 Deparlment of
Operations Research Stanford University
Glnsbur~h V and L van der Heyden 1985 General Equilibrium with Wage Rigidities An
45
AppJication to Belgium in A Manne ed t Economic EQumprit1m~ Made Formulation and $oJioo
Mathematical Programming Study 23 Amsterdam NorthmiddotHolland ~
Goulet J 1968 1 Adjustrent CO$1S in the Theory o(the Firm Reyiew of EconomiC Studies 35 bull
pp47middot55
Goulder L 1985 Tax-Financed versus BondmiddotFinancedGhanges in Governent Spending
bull Mimeo Harva(d UniversilYmiddot
Goulder l J Shaven and J Whalley 1983 Domestic Tax Policy and the Foreign Sector The
importance of altemative foreign policy formulations to results from a general equilibrium tax
analysis model In ~n methods in Ibe antico of income from caoia1 ed Martin
Feldstein Chicago University of Chicago riSS
Goulder l L Summers 1987 Tax Policy Asset Prices and 3rowtly ~
A General Equilibrium Analysis NBER Working Pper No 212B
Grandmont J 1982 Temporary General Equilibrium Theory Ch 20 in Handbook 01
Mathemalical Eccnornics Vol II NorthmiddotHoliand
Harberger A 1959 The Corporate Income Tax An empirical Appraisal In lax Revision
Comoendjum Voll House Cornmitee on Ways and Means Washington DC Govemment Printing
Office
Harberger A 1962 The Incidence of the Corporate Income Tax JOULn1 of Political
Economy 70 pp 215middot240
Harberger A 1964 Taxation Resource Aliocalrlon and Welfare fn lb Role of Dir bullbulll and
Indjresl Taxes io tbe Fede[al Reserve s~I~Ill Princeton Princeton Unlveresity Press
James J 1985 New Developments in the Applications of General Equilibrium models to
Economic History in Piggott aand Whalley Eds
Jorgenson D and K Yun 1984 Tax Policy and Capital Allocation Harvard Inslitute of
bull
I
I f
Economic Research WP 1107
Judd K 1965 ShonmiddotRun Analysis of Fiscal Policy in a Simple Perfect Foresight Model
Journal of poJiHcal 5CCJiW1l pp 298middot319 r
Upton D J Potbc J Sachsa nd L Summers 1982 Multiple Shooting in Rational
Expectations Models fuoometrlc 50 1329middot1333
Manne A ed 1985 fconQrnjc Eoullibrium Model FOrIDujaljQo and Solution Mathematical
Pogrammlng Study 23 Amsterdam NorthmiddotHolland
Manne A 1985 Or the Formulation and Solution 01 Economic Equilibrium Models In A
Manne ed ECQOQlli-Jr~ibmiddotum ~tQdel EQlJulailon and Solution Malhematical Programming
Study 23 Amsterdam NorthmiddotHolland
Mansur A and J Whalley 1984 Numerical Specification of Applied General Equilibrium
Models in Scarf H J Shoven Eds Aoplied General Eouillbrium AnaiysectIsect Cambridge Universily
Press Cambridge
Merrill 0 1972 Applications and Extensions of an Algorithm thaI Computes FixedmiddotPoints of
Cenain Upper Semimiddotcominuou Point to Set Mappings PhD Disserlation Unlvel$ity 01 Michigan
Negishi T 1960 Welfar Economics and Exislence I an Equnlbrlum for a Competitive
Economy MelroeCOOQ11ka 12 92middot97
Pereira A 1985 On the Existence and Uniqueness of Optimal Ou1put Path lor CRTS Firms in
Adjustment Costs Technologies Mimeo Stanford University
Pereira A 1985a Consumer Expectations and Ihe COSI 01 UllUty In an Intertemporat
Framework Mlmeo Stanford University
Pereira A 1985b Corporale Tax Integrallon and Government Deficits A Dynamic General
Equilibrium Anaiysls Mlmeo Stanford University
Pereir A 1987 On Ihe Computation of the Users Costs of Stocks Ecooomi (forthcoming)
47
Pereira A bull 1987b Equal Yield Alternatives and Government Oeficlts Mimeof Stanford
University
Pereira A 1987c A Dynamic Applied General ~uHlbrium Model for Tax Policy Evaluation
Comlletamp Docurnent2tion Mimeo University of California San Diego (in progress)
Pereira A 1987d Should the investment Tax Credit be Rei(ltroduced Mlmeo University
of California San Diego (ir progreSs)
Piggott J and J Whalley eds 1985 tJew QeyeQpments in Applied General EQuilibrium
bullenalysjs Cambridge Cambridge University Press
Preckel Pbull 1985 Alternative algorithms for computing economic equilibria In A Manne Ed
Radnor R 19amp3 Equllibrium under Uncertainty In K Arrow M Intrilligaor ods
I ibull
Robinson S 1 g86 Mullisectorai Models of Oeveloping CountrieS a Survey Oivision of
Agricullure and Resources UC Borkeley WP 401
Scarf H 1967 The approximation of fixed points of a continuous mapping ~MM Jouwal of
epplid Mathornallcs 15 pp 1328middot43
Scarf H with T Hansen 1973 00 the Computation of EconomiC Eguilibria New Haven Yale
University Press
Scarf H J Shoven Eds 1984 61l1llied Genera Elujlibrium 6nalyssect Cambridge University
~ Press Cambridge
Shoven J 1976 The Incidence and Efficiency Effects of Taxe on Income from Capital
Jurnal of ramie1 Economy 84 pp 1261-83
Shoven J 1983 Applied General Equilibrium Tax Modemng IMF Slaff PaoerS 30
Shoven J and LB Simon 1986 Share Repurchases and Acquisitions An Analysis of which
Firms Participate NBER Working Paper No 2243
bull
48
IShaven J and J Whalley 1972 ft General Equmbrium Calculation of lhe Effects of
Differential Taxation of Incoma from Capital in the US Jauroe of public Econgmics 1
pp281middot321 (
Shoven J bull and J Whalley 1973 General Equilibrium with Tax$s a computation procedure
and an existence pr00f ReviEhi of Economic Studies 40 pp 475middot490
Shoven J J Whalley 1977 Equal Yield Tax Alternatives A General Equilibrium
Computa~ui0nal Tfchnique J~HJrnal of PUblic EconomIcs 8 pp 211~224 bull
Shaven J J Whalley 1984 Applied GeneralmiddotEquilibrlum MOdels of Taxalion and
International Tr8de An Introduction and Survey Journal of Economic literature 23t pp
1007middot1051
Slemrod J 1980 A General Equilibrium Model of Capital Income Taxation Ph D bullDissertation Harvard Univnrsity
Slemrod J 1983 A General Equilibrium Model of Taxation with Endogenous Financial
Behavior in M Feldstein edbull Behavioral Simulation Methods In Tax Policy Analysis Chicago
UniVerSity of Chicago Press
Stone J 1985 Sequential Optimization and Complementarily Techniques for computing
Economic equilibrium in A Manne Ed
Summers L 1981 Taxation and Corporate Investment a q-Theory Approach BroQkinas
Summers L 1985 Taxation and the Size and ComposHion 01 the Capital Stock An Asset Price
Approach Mlmeo Harvard University
Whalley J 1975 A Geneal Equiloibrium Assessment of the 1973 United Kingdom tax
reform Economica 42 pp 139middot61
1JLTIllOS JIQll~)ltG PAPERS PUBLICADOS bull
I
I I
j
112 70 - LUCENA iJiCIIJo ~ IIEnvironrilent Honitoring and Orqanization Structure T (Setellibro 1987)
nI 71 - L~ENA Diogo A Note on the Representation of Information Stxueture (SeteJObro 1987)
nQ 72 - LUCENA D1CgO Envuorment 111)nitoring and Organization
Structure II (Setelilhro 1987)
ril 73 - AlmmE3 Antonio Pais e CAsPAR Yitor bull Tributacao Inoentivos e InvesUlllentos Analise Qualitativa noYellibro 19S7)
n2 74 - liAlltOSA rose Pedro Opt1lllal lIage R1gidity A SUggestedtlthodology to Teet tb Theory and an Application (Outubro 1987)
nI 75 - CABRAL L 11 B Three lIotes on SYJIIlletric Gemes witt itsYJIIlletric Equilibri (NoveJObro 1987)
n2 76 - LUCENlt D10gOTo Search or Not to Search (Fevereiro1967)
nQ 17 - SA Jorge asconcelos euroI A Theory of SynergyB (Fevereiro1987)
n2 76 - VILARES tanuei Jose Os Be IntenaMios IlIIportados ColIIO Factor de Produao (Julho 1987)
nQ 79 - SA J~rge VasCQncel03 e HoY to Copete and Co~unicate in tature Industrial Products_ (aio 1988)
n2 80 - ROIl Rafael Learning and Capacity Expansion in a New larket Under Uncertainty (Fevereiro 1988)
n2 81 PEREIRA Alfredo llarvilo Survey of Dynamic Computational General EquilibriUlll Hodel for Tax Policy Evaluation (Outubro 1987)
Qualquer in ormacao sabre o 1IIorlltlng Papers ja publlcados sera prestada pelo Secretariado de Apoio aos Docentes podendo os ~e3OSf
ser adquirido5 ne Secyo de Vlndas da Faculdade de EconollLia UNL na Travessa Esteviio Pinto Cltpolide - 1000 LISllOA
---- ---
------
------
ll-ltiLt 1 uynamlc Lut MOOSS fOf I ax YOtlcy tv8luallon
Inanell ~ets and maf~9ts ojumLrtm- ----_ ----c_~
privata do btl GlvldtHlI pL~lc aliocalQflo rls shy
1rlc ypa expGct assets equity retention dobl 01 savlnq$ promlums path
Sal ra rd fu Barton phys cap1 used 10 buy soquence lt
Shavon-Wha1ley no financial exogenous oxogenous no physicBl 00 bull 01 sIalic 1E myopic 1985 assets capital flqrlil -
Andersson phys cap used 10 buy 1987 no financial no no yo physical no dynamic ~ perfect
assets capital Auerbach~KolUko If phys cap usod to buy
1983 1987 neo financial no no physical cap no dyoamlc ~ perfectYmiddotmiddotnSSf)ts and Qovt debt SailBrd-Goutder phys cap
1983 1985 no financ[al assets-shy -------
Bovenberg phys cap 1985 1986 no iinenclaf
asso1s Ertlch-Glnsburgh~ phys cap1
-Heyden no financlaJ 1987 assefs
FeltoRsteln phys cap 1984 1986 Invostment is
bondmiddotfinanced
used to buy
exogenous exogenous nObull physl~al no dynamic ~ perfllcl capital
used to buy
no no no physical capital
no dynamic ~onlinuos lima
PI perfect
used 10 buy no no no physical no dynamIc PF perfect
capital jllNt) porlod) bull used 10 buy
no no ye physical cap no dynamic PI p4)rfect and govt debt (Iwo Iods) bull
Goulder phys cap used to buy 1995 no financial exogenous exogenous yenbullbull physical cap no PFlfE bull IdynamIc perfoct~assels and am dabt slaUc
- -~ - shyGoulder-Summers phy$ cap
1987 bonds and no e)(oganOU$ no used 10 buy ye dynamic PF perfect _ bullbullultv bonds MullY (exoaenous
Jorgensonyun phys cap 1984 no financial
essels -~ - ------
Pereira phys cap 1986 1967 bonds end
equHy
________
no
8xogenoos
no
exogenous
no
ye
used to buy two type$ of physical
copltal used 10 buy
go dobl bonds equity
no
no
dynamIc
-
dynamlc
~
lE
perfect
fIbullbulllble (nonperlecB
middot6middot
~-- _middot~ ___~w_ - - ___ bullbullbull___ -------- shy~
-----
------
- ---
-
IlflpnUItll tit lun pVII-y PIlUAlIV
q6H1brla oqual fax polley ovnluaU-on efUclency derlvatton Ioorllhm
calibration param~hr computation COmelfed llold tndfcaHH eHecls
Bnl jard~FuU6rtQ n base year back soll1lion steady-sta1e yo inlsrgroup inerseel or Iii
VS transilion lump-sum with Interfemporai 1965
~Shovn Whallmiddotv replication wllh exog Morrills nmamo1ers alQorilhm _~ond~middotsta1$ ears Inc CVor EV
i~noerucn sleady-sfala no agg(ogaled lnlalsectora 1967
base year back sobHon SiMNON package replication wHn oxog valiant of VS Iransilion no Indicamptor lnlertamporal
paramelsrs multiple shoo ling 6~~~l~slat8
AU8rt)ach~Kotllkot steady-slato yo Inlerguna-r a lions
1893 1llH
quatitaUve by replication a$Sium~lkml Gauss-Soidel liS Uln1WOn any rule -lilh 1111ermpOrB
sonsililJ1y I _ ~poundI~Jjnnelill $t(lfdvmiddot~luiA J3djlt15S ~~ nnrilW ~--------
a ail r nimiddotGould or bas9 9tH bac~ solution Ballard~ staadymiddotslalo )05 kltHSclo(a~
1983 1985 VS Iransllinn lump~sum dynalTlk CV inlerlemporal steadY slafe parameters alaorlthm replicaUon+ with oxog Gourder
+stead~stato (ers loc inlareel10rat BCvonborg qunmaiive by dynamic version stoadY-$Igtto yes 1985 1985 YS transition some govt eN Inlortempora
S9rHltivilv OnearizaJian method ropllcaUon assumptionl of Johansons
+steady-sfata paoo palh Erllch-Glnsbu rgh~ actual vorsus no aootagalod inlersectoral2 bftse years back solu1ion variant of Negishis
Heyden replication wIth alltogbull optlmizaUon countorfaclUal no ~odjcator inlottempora 1987 paramlJters aocroach
faltonstaln baso yeat back solution actual versus no aggrampgalfld intersElclornl replication wilh exog Merrills counlerfaclutll no indlclltor middotlntamprlempOull1984 19B6
parameters al90rithm 1981middot1982
Gouldar staady~slala Inlar5ecioralbase year back souUon Variant or 1S lronsilion no dynamic CV inlortempora
stoady state parameters Newlon 1985 feplicalion+ with iUog Fair~Taylorl
+slaildy-sla1e Intargeoorat Gtlulder~Summers base year back solution varIant of stlJadymiddotstate fixed no aggregated fntersectoral
vs trarlsition government Indicator Intertempo1al steady state oatameters Newton
repticaUon+ with exog Fair-T aylQf 1967 +ste-adv-stata bullbullbullndlno financial
Jorgenon-Yun econometric steady-stale Money melfic In1ersecloral
- - 8slimalion vs traosUian no social weUare functton intertampotal1984 +~teadlmiddotstatl)
~ shy
qtJalitaUv8 by NPSCt optimal path ye generalized inlersectorai 1986 1987 Perelr
replication 8S$umptlonl optimizatIon V$ lumpmiddotsum dynamic lntertemporat se~~ttilitf ____Jllgorilhm opllmal pah pen inc CVoclV financial
- 9 shy
bull ~ ___ - _______ - ____ bull___ OM bullbull __ bullbull __~____ ~_
--
- - - --------
--------- - - - --------
---
1nltJ1 bull I ~ 1J11ltI1 v I ~
-
Sa Ita rd-Futlertonmiddot intargration consumption lax ~StHlven~Whall~y VAT capila~ or labor tax
Ch~mqflS1985 Andersson
consumpion tax neutral corporals lax system 1987
it ue rb a c h~ i(at II Sltotf COIlSUft1ptian lax (poundpird ItS hibor Mx~on ufocto of taxation on capital fOfrnJilti1
dvnamio fiscal policv effects af OQvernmanl defioit (3tlard-Goulder
1883 1987
allocts 01 the adoption of a 1ge3 1985 pure consumpHon lax
im~rtmce of cons fCreSClht -Bovenborg analytically orianiad approach
1985 1986 effects ot pure conumption tax capllal income taxation in Opo economies
Erllch-Glnburghshy-Heyden reat wage policios in Belgium
1987 Felter1steln 1984 - theorotical
1985 ~ linnncial crcwdlng out1984 1986
Goulder effects of tax- vs bond-financed 1985 chanQes In pubtlc expondllufe
~~--
Gouidr~SUmmampfI torpora1$ lax cvts reduced 1987 ITO increase 10 gasolioe taxes
Jorgenson-Yun
effects ot several programs of tax reform on the allocation of ~~ptai
1984
-- Purelra
corporate la)( Integration effects cf romiddotinlroducinQ imreslamnl tax credits
1986 1967
10 bull
bull -~- ~- --- _ - ---____--_ - -----_-_ --~ ---- ~
1 1
2 Brief overview of non-dynamic CGE models
The origin of most or the models we survey goes back to the work of Scarf (1967 1973) whlgt
first developed a reliablo algorithm to compute equilibrium price for a ArrowmiddotDebreu economy
HiS algorithm used simpicial Bubdivision techniques and can -be shown to be the computational analog
of the fixed point theorems previously used to prove the existence of equilibrium His technique
could soW a model with an arbitrary number of consumers and commodities as long as all agents
were price takers consumers were subject to budget constraints demands were continuous andmiddot
production did not display increasing returns to scale The alQorithm while guaranteed to converge I
I I I
was relatively slow for prcb1ems involving more than S1If 20 dimensions A maior improvement in
computational speed Woo offered by Merrills algorithm (1972) which used tpe same fundamental
ideagt as Scarfs procedure
Scarfs model (as is true for the standard Arrow-Debreu model) does not include a government
sector ~ neither taxes nor public goocs At one of the mosi promising appHciUions of the new
computational techniqu w~s in Iha area of tax policy evaluation Shoven and Whalley (1973)
exlended the general equilibrium model and computational approach to include a wide rry of taxes
and a governl11ampnt spending plan The original ShavenmiddotWhalley model was static (although the
different commodi~ies CQuid be considered simitar goods available at different datest as In
Arrow-Debrau) and government was aumad to run a balanced budget The data are arranged as a
social accounling matrix The models spacification and calibration is checked by solving it in the
presence of the base set of laxes The result should ba exactly tha initial social accounting matrix
After having pasd this replication check the model is solved for a counterlactua equilibrium In
the presence of a new tux design The result is once again a social accounting matrix The two
equilibria are compared in order to tSS($S the impact of the new tax plan The first uses ot this
bull
1 Z
model for lax policy evauajon were ShovenWhalley (1972) Whalley (1975) and Shoven
(1916)
Completely static models as Shaven-Whalleys are unsatisfactory for many tax reform ISsues
These include corporate lax Integration effects of investment tax credits effects of accelerated
depreciations consumer Or e)penditure tzxes~ importance of saving subsidies like IRAs etc These
are essentialy dynamiC ksues They involve not only the allocation of capital across sectQfsbut
perhaps more Important the capital intensity of the economy But the capital accumulation and
capital reallocation take time and may involve adjustment costs Because of these issues Ballardmiddot
Fulierton~Shoven~Whaloy took the first steps towards developing a dynamic model
In this conlexl the ShovenWhalley model has been extended and implemented for Ihe US
-economy In BuIIBrdmiddotFulienonmiddotShovenmiddotWhalley (1985) While Ihis book completely documents the
mo~el it was used in several publicalions beginning in 1978 Their mode) consists of nineteem
production sectors twelvo households and fifteen consumer goods It includes a very detailed set of
taxes including the federal and state personallncome taxes federal and state corporate income taxes
SOCial Security taxes pro1J0rty taxes unemployment Insurance excise taxes sales taxes etc It
has been calibl att-d to reproduce the 1973 US economy Recentty a version corresponding to the
1983 JS economy has been developed
wThe 8alard-Fullenon-Shoven Whalley model is dynamic In the sense that consumers face a
choice between current consumption and leisure versus future consumption (which can be
purchased via s wlngs) The consumer classes act as if they were maximizing a nested CES utility
function over the domain of a CES aggregate of contempo-raneous consumer goods and leisure and
future consumption subject to their income constraint The parameters of those funcHons
determine the shares of income devo1ed to each commodity to saving and to the purchase of leisure
They also determine the two key elasticities in the models ~ the elasticity of labor supply with
bull
13
respecl to tgte rbullbull1 aoer lax wage rae and the elasticity of saving with respecl 10 the real after lax- rate of retum to capital
In the model consumerS have myopic expectations regarding lulure prices and in particular
regarding the future rale of return 10 capilal Ballard (1983) and BallardmiddotGoulder (1985)
inco~ora1e both perfect foresight and limited foresight into this model Future consumption is
acquired by buying a fixed composition portfolio of real investments that offer an infinite annuity
of returns
The production side of Ihe model is compielely static The mod1 Incorporates a constantmiddot
olasicily 01 substitution betweeen prlmary inputs in production (capital and labor) and fixed I coefficients for intermediate inputs The model distinguIshes between Industrial outputs and
eonsumer goods for the simple reason Ihat tile data are classified differenUy Industrial sectol$ I I involve such categories as forestry and fisherle metal mino9 and publishing and printing while
consumers purchas-3 fUfHi1ure 2ulomoblles and books This fact is recognized in the model by Ibull
incorporatirg a secord stage Of production which converts industrial outputs intO consumer goods
This technology is uStally mod9-led as a flxed~coefficient conversion matrix
The BallardmiddotFINton-ShovenWhalley model assumeS that the private sector finance
marginai ioveslment with the same composition of debt and equity as currently exists in each sector
This is the same arsumrtion that Herberger originally used Investors all hold debt and equity in tile
same proportion Thampreiore this ownership can be aggregated into simply capital ownership For
tax purposes however the separate treatment of debt and equity is taken into account at both the
corporate and persofa Io=vel SimllarlYJ the dividend poIicies on corporale equitY are established
exogenously There are no government bonds in the model since there are no government deficits
The modol is solved for a sequence (as many as 100) of temporary equilibria wltll consumers
allocating income between present and future conumption at each point in time The palh for 1he
14
economy ~ a sel of connected equilibria The connectIon is pra~ided by capital accumulation
Capital accumulation is endogenous and determined by saving The model stans with a sociagt
accOunting mi1rix In the bzse case the economy is assumed to be in a steaqy state growth -path (along
which all relative prices ~fe constant) The madel solves for both the naw steady state growth path
and the transltion 10 it zf~er a policy intervention rhe authors have frequently addressed the
question-of howong rt tKes to effectively settle un10 a new steady S1ate growth path
The dynamics of the model are limited however in that future consumption is collapsed into a
composite commodity Aso the absence of government deficits and the lack ofproductlon dynamics
limits the realism of the BallardmiddotFuliertonmiddotShoven-Whatley model for the analysis of dynamic
policy issues (such as the adoption 01 a consumption tax or Ihe elimination of the investment tax
credit) The models we survey try to Improve upon Ihe Ballard-FlIiertonmiddotShaven-Whalley
dynamic modelling The models Jn~orporate more forward lookTng behavior They improve the
dynamics on the conampumrtkm side of the economy and some of them introduce true 1ynamlc behavior
on the part oI the producers We turn now to the issues inY9lv_ed in developing such dynamic models
15 I I ~3 ~ssues in the modelling Of dynamic behavior
We will 100k firt at the speclficalion of economic behavior of consume(s~ producers
goyemment and Ihe r~sl of Ihe world The modemog of corporale and household financial decisions
I is then addr5sed Finally Ihe eencept of equilibrium is discussed
I
i
I 1
31 Consumers behavior
Early efforts 10 build dynamic features into the economic behavior 01 consumers are due to
Ballard (1963) and Auerbach-Kotllkoff (1983 1984) Their work has been closely followed by
several subsequent authors In faet dynamic household behavior Is the single most pervasive aspect
amohg the models surveyed
All the models Inc-O1orale some loro of lila-cycle behavior Household behavior is dewmloed
by the maximization of an addWVely separable time invariant intertemporal utillty function The
utility lune1ion is defined oyer the domain of the consumption goods In the economy In most of the
models leisure is also an argument in utility so that labor supply is optimally determined
Utility maximi4ation is subject to a lifetime intertemporal budget constraint which equalizes
the presen1 value of consumers income and expenditure More reeentiy in Andersson (1987) bull
Bovenberg (1985) and Pereira (1986b 1987d) the constraint i$ defined as a sequence 01
recursive equations of motion on wealth This has the potential advantage 01 accomodating liquidity
constraints However It should be recognized that in the absence 01 liquidity constraints (ie when
consumers are free to completely borrow against future income) the two specificallons of the
househofd ccnstraint are essentiay equivatent Furthermore tn both versions saving is optimally
determined as a way of translaring wealth intertemporally
1 6
Some lodels now have a sophisticated dynamic specification of hoosehokl In Ballard (1983) -
BallardmiddotGoulder (1985) and Auerbach-Kotlikof (1983 1984 1987) consumers behavio is
embedded in an iterQEneraional setting Each of them has S5 age cohorts simulataneously alive
Som-e measure of jntergener1ional altruism is considered in the form of bequest motives In Ballard
(1983) and Ballard-Gouldr (1985) consumers derive utility directly from bequthing parl of
their wealth Therefore bequests assume 1he form of a scrap value of terminal wealth of a
generation In AuerbachmiddotKotlikof (1983 1984 1987) each generation empathyz bullbull with Ihe
utility of future generations ovar bequethd wealth In addition Ballard (1ge3) includes detailed
demographic poiections wfill the intent of incorporating in the policy analysis the effects of the
bull Post World War tI baby-boom
In the real world household decisions also include the optimat allo~atiOn of saving among
shyalternative physical or financial assets Theories of household portfolio behavior are relatively
Iwen~establshed However they involve uncertainty as a crucial element Accordingly they have
middotnot been incorpora~ed in the context of the determinIstic dynamic models we sutyey here We will 1 Come backt-O this Issue below
32 Producer behavior
The efforts to build dynamic features into the economic behavior of producers are more recellt
and less widely adopted The first attempts are due to Bovnberg (1984 1985) and Summers
(1965) Dynamic behavior has been more fully incorporated In the recent models of Andersson
(1987) Auerbach-Kotlikoff (19B7) GouldermiddotSummers (1987) and Pereira (1986b 1987d)
Part of the reason why production side dinamlcs has been more slowty adopted is the weak
supply of accepted theories referring to the dynamic behavior of the firms In the models referred
I 17
to above gynamlc producion and investment b~havior are middotinduced by the existence of capital
adjustment costs and linked to Tobins q theory Adjuslment costs are designed to capture both the
incomplete mobility of capital aoross industries and installation costs ~e the costs of adjusting
capital towards its cpHmai level)
Adjllstment costs can be conceived as internal to the firm and measured in terms of foregone
outpul along the lines of Lucas (1967) Alternatively adjustment costs can be viewed as actual
external cests incurred together with Ihe purchase costs along the Unes of Gould (1969) With the
exception of Pereira (198Gb 1937d) all of the CGE authors follow the former approach
The firms in the economy maximize their market value as the present discountad value of the
future stream of dividends In Andersson (1981) and Pereira (198Gb 1987d) firms are seen as
maximizing the present discounted value of net cash flow Maximization is constrained by Ihe
adjustmenl cosl lechnology and ~n equation of motion describing Ihe evolution of the capital stock
It should be noted that undereogenous divldendrelention rules the two problems are equivalent II
should also be nOled Ihat a salisfectory economic rational for th bullbullistoo of dMdonds with the
present tax code is missing in the profession (Shoven (1986))
The models with static formulation of producers behavior are characterized by passive
investment behavior Investment merely accomodates to saving in the eccnomy With a dynamic
formulation induced b adjustment costs real investmen decisions are forward looking Investmenl
Is endogenously and optimally determined by the firms A fundamental difference between the
shonmiddotrun in which capital stock is given and longmiddotrun in which Ihe level of capital is allowed to be
optimally determined is emphasized
This extra richness of produc1ion dynamics is not without costs A careful look at the summary
tables will clearly show an inverse relation between the adoptlon of dynamic features in production
and the level of disagreggation of the production side of the economy In fact with production
bull
j
1 6
dynamics $e dimension of the problems is immensly increased Let IS be more specific
In a static framework with constant returns to scale production technology the output level ismiddot
indeterminate The optimal aliocation of inputs can be obtained by cost minimization with any
feasible output level generatillg zero profits Such zero profit conditions are used to solve for the
output prices in terms of the factor prices and hencemiddot to reduce the dlmensionality of the problem
from the -number of commodities and factors to the number of factors Thus even if the model deals
with 30 production sectors the computation of an economic equilibrium can take place using only
the dimensionality of the primary inputs in the economy
Now under certain regularity conditions on the production and adjustment costs technologies
leading to enough concavity of the optimality objective the intertemporal output path for the firm is
endogenously optimally and uniquely determined even with constant relums to scale technologies
(see Pereira (1985 1987a) Accordin~y with adjustment costs in the model optimal profits
will in general be non-zero This result has crucial implications for the computatiQn of equilibrium
in the model~ with production dynamics With adjustment costs no reduction of dimensionality is
possible The curse of dimensionality returns as a binding constraint
The introduction of adjustment costs adds another significant complication to the model With
capital being less that perfectly mobile in the economy different rates of return on capital will
exist in different sectors This is a difficult problem to tackle conceptually in the absence of
uncertainty Also in the real world producer maximization choices include also its choice of
financial ratios (debVequity) and payout rates (dividendretained earnings) These financial
subjects are difficult and much Clf this behavior is still taken as exogenous by the modellers We
will come back to these issues below
middot i
19
bull 33 government behavior
middot
Two issues dominate the modelling of government behavIor First govemment behavior has
Ibeen typically seen in Ihe CGE liIerature for lax poliCY evaluation as constrained by yearly balanced
I budsets (see for example 8allard-Fullrton-S~ovn-Whafley 098Sraquo The analysis of
govemment defidlS and public debl in a CelE context requires a dynamic selling Second the level of
government expenditures is either exegenouty given as in Auerbach-KoUikofi (1984 1987)
6ovenbrg (1984 1985) Feltenstein (1984 1986) and Jorgenson-Yun (1984) or
endogenously (but no optimally) determined by the balanced budget conditions as in
6allard-Fullerton-Shoven-Whalley (1985) Andersson (1987) Erllch-Ginshurgh-Heyden
~
(1987) Gouder (19aS) and Gouder-Summers (1987) In the second case the composition of
bull public expenditures is often optimally determined However the leve~ of government expendi1ures
can only 00 endogenously and optimally determined if the government is seen as an optimizing agent
and is allowed to run deficits
The first attempts to deal wiU the government defiCits in CGE tax models are due to
Auerbach-Kollilltoff (1984) and Feensain (1984) In Auerbach-Kotlikof (1984) government
expenditures afe exogenous They grow at the rate of growth of population However given 1he tax
structure and tax revenues yearly deficits and surpluses are atTowed subject to an in1ertempond
constraint that the present value of future tax revenues equals the present value of future
expenditures
In Fellenstein (1984 1986) government expenditures are also exogenously given He also
allows the government to run deficits to finance expenditures in excess to 1ax revenues However
surpluses are returned to CQnsumars in the form cf transfers Accordingly government ts not
subject to any constraint regarding the future repayment of public debt
In GouJder (19851 the 9Vernmenl maximizes a static social welfare ~nctio subject to a-
bull
20 I i
I
I I I
I bull
balanced budget conSirtlL This follows the optimal alocatkm of government expenditures alpng
the lines of Baliard-FHeiton-Shoven-Whalley (1985) Ttle model Is generallzea to allow for
exogenous changes in HIe time path of government expenditures Two financing alternatives are
consjdeft~d and contr8s1ed EldditionaJ tax revenues and bond issuance
Pereira (198Gb 1137d) auempts to address both the incorporation of deficits and the
determination of govement expenditures The path of government expenditures and the path of
delicilSlsurpluses (ano therefore the path lor debt) are endogenously and optimally determined The
government is seen as mltJximizing n intertemporal social welfare function given the 13x structure
Optimization is subject to a sequence of recursive equations of motion reflecting the evolution of the
ptlbllc debt allowing fDr government budget imbalances It should be stressathat this specification
01 the government cOMtcalnt is equivalent to the specificaUon in Auerbaeh-Kotlikoff (1983 1987)
when no liquidity constraints affect -gove~nment behavior
The extra richness in the treatment of government behavior allows tfte examination of
government debt poices in a truly dynamic setting Also financial crowding out effects induced by
government deficits can be analyzed (see Auerbach-Kottlikoff (1987) Feltenstein (1986) and
Perer (198Gb 1987e))
34 Foreign sector
Most of the open economy models follow the assumptions of balanced trade with import and
export net demands characterized by constant elasticities along the line of Ballardmiddot
-Fullerton-Shaven-Whalley (1965) Such is the case of Salfard (1963) Ballard-Goulder
(19S5) and Goulder-Sllmmers (1987)_ In Feltenstin (1985) the rest of the world is treated as
21
an addjlion~ consumer group
Bovcoberg (1986) develops a model In which two economies lIr considered each following
intertempoml perlect foresight paths These economies meet in the international forum Their
trade relaroMhips are characterized by yearly balanced trade accounts
None of the new generation of dynamic cae tax models has yet IncorPorated the lnternational
capital flows as done in aoulder-Shoven-Whalley (1983) for the earlier Ballarrshy
-Fullertonmiddot Shaven-Whalley (1985) model This is unfortunate as there is an important
Intertmporal aspect to international lending The first al1empts along Ihese lines are due to
Andersson (1987) and Erlich-Glnsburgh-Heyden (1987) Andersson (1987) in his model of the
Swedish economy adopts a closed economy appraoch in which rates of return in the domesJic
economy are largely determined by the international capital markets Fidinterest rate induce
Intlnatlonal capital flows which determine and finance the International lrade imbalance In
lurn In ErlichGinsburghHeydenmiddot (1987) foreign trade Is generaled according to an
intertemporal Irade welfare lunction with constant import and export elasticities In the shortmiddotrun
they allow inlernational trade inbalances which generate capilal ftows to the domestic households In
the long run however trade balance is assumed
35 The need for financial markets
A dynamic economic Wucture not only provides the ideal environment 10 model many features of
economic behavior it also permits one to Incorporate the financial side of the economy If the
government is allowed to run deficits the question 01 deficll financing automatically follows If
investment is optimally dewmined and returns to capital are different aClOSS sectors problems of
investment financing arise If there are seeral final1cfal assets In the economymiddot government
4-
I i
ibull
I I I I
I
j
22
bonds privpte bonds and equity (or simply physical capital installed inmiddot different sectors) the
problem of allocation of saving among assets with potentially different returns arises
The papers surveyed var greatly with respect to l~e extent of their attention to the financial
side of the economy At Clne extreme are the models in ~anardmiddotFuliertonmiddotShoven-Whalley (1985)
Andersson (1987) Ballltld (1983) Ballard-Goulder (1985) Bovenberg (1985 1986) and
Erlich-Gjnsburgh-Heyden C19B which are devoted exclusively to the real side of the economy
In turn Auerbach-KotliKoff (1984 1987) Feltenstein (1984 1986) and Goulde (1985)
allow for government debt In these models saving finances changes in government debt and physical
capital Private and publlc 2ssets are perceived by the households as perfect substitutes The
allocation of saving merely adjusts to the relative demands for funds
Feltenstein (1984 1986) is the only model surveyed which introduces money Government deficits are financed by issuing money and bonds according to an exogenously given rule Money is
demanded by consumers for transaction motives and an exogenously given fraction as a store of value
On the other hand government bonds and physical capital are the vehicles for the intertemporal
transfer of wealth
Summers (1985) and GOlldermiddotSummers (1987) introduce a whole menu of financial assets shy
firm specific equity capital Different assets earn different rates of retum However such rates
are equal up to constant and exogenous sector speCific risk premia Therefore the introduction of
constant exogenous risk premia as helpful as it may be in the context of calibration does not solve
the main issue of the non-optimality of the allocation of saving Also talking about risk premia in a
deterministic context is somewhat unsatisfactory
Pereira (1986b 1987d) also introduces a whole menu of assets private and public bonds and
firm specific equity However all the assets are expected to yield the same rate of return and
therefore perc~jved as perfect substitutes The different asset types allow consideration of
23
exogenOU$~8tUequjty triO dividendlrc~enlion rufes and therefore seve-ral sources of investment
financing bond equity and retained earnings
The nonmiddotcptimalHy of the allocation of saving and the absence or exogenelty of corporate
financial rules is _a reftectlon of the limitations of the determInistic approach Either in a context of
perfect anticipation oi the future prices Of in general within the t$atm of point price
expectations all the papers follow a deterministic approach Under such circumstances consumers
either expect diHerent rales of return (inclusive of risk premium) across assets in which case
they wilt buy only one asset (that with highest rate) Of they expect equal rat~s of relurn in which
case Ihey are indifferent about the asset composition of Iheir portfoliO There is no way of
Iradingmiddotoff rales of return and risks to obtain an optimal interiof solution to the problem of the
allocation of saving
The most advanced contribution in the modelling of saving allocation in a CGE setting is due to
Slemrod (1980 1983) in the context of a static one-period model In his model consumers act
according to a two stage separable decision process They Hrsl decide on how much to save Then
they decide on the allocation of saving according to an Indirect uttlity function dependent on the rates
of ratum and variances offered by the different assets in the eoonomy The sourCe of riskiness in the
ampConory comes from an uncertain marginal product of capital On the other hand aside from
portfolio decisions the res of the economy is insulated from uncertainty
The most complete contribution in terms of the treatment of the corporate financial rules is
FulienonmiddotGordon (1983) In a variant 01 Ihe BailardmiddotFulienon-ShovenmiddotWhelley (1985) model
Fullerton and Gordon have capital intensity and optimal financial decisions Jointly deter~ined
through a two siage process The cosl of financing capital is minimized by trading off the tax
advantages of debt against the epected real bankrupcy coSls inherenl with high debVequity ratios
Given the optimal debtequity ratio the level of inveSlment Is chosen such that at Ihe margin the
24 l
bull
rerurn pn ~uity equas the return on bonds plus an exogenous risk premium
36 Market assumptions equilibrium and expectations
Virtually all of the papers arc characterized by Walrasian market clearing assumptions All
markets are perlect1y competi1ive Atomistic competition among agents 1s assumed even thoughmiddot only
a flnile number of agents is considered Virtually no mark~t disequilibria or price stickiness afe
considered The only exception is ErlichmiddotGinsburghmiddotHeyden (1987) bull In thei model of the Belgium
economy I the _wage rate is fixed in the short-run Therefore in the shOrlwrun desequilibrium in
the labor market will generate endogenous unemployment However in lh$ long-run all prices
Including the wage rata are Ilexible and accordingly ali markets ciear
Given the dynamic nature of behavlor in the economy markelmiddotcfearing prices in each period
depend on expectations of fture prices and on tax variables In the economy There are essentially
two ways at interpreting the economic equilibrium in such a dynamic conte~bull If future prices are
perfectly anticipated (io expectations are self-fulfilling) a perfect foresight equilibrium
prevaHs Then future actions are merely the implementation of current decisions for future
periods However jf ptica expecla110ns are not perfect (Le agents make mistakes w1th re$pect 10
future prices) then a temporary or shortrun equilibrium prevails Markets ciear and clearing
prices depend on fu1ure price expectations Current plans about the future are typically not
precisely implemented They will be revised as more or better information becomes available to the
economic agents See Grandmant (1982) for a comprehensive survey 01 the temporary equilibrium
literature
Wilh the exception of Gould (1985) and Pereira (19S6b 19S7d) all the models surveyed
adopt th~ concept of a perfect foresight equilibrium In turn BaliardmiddotGouldr (1965) considers a
J
2S
flexible am9unt of foresight In terms of the number of years over which price movements are ~
foreseen Pereiras model (1986b 1987b)) is flexible in a somewhat different way in that it can
indude any range of foresight from myopia to perfect foresight The choice between the perfct foresight and lemporary equilibrium is ultimately to be made on
philosophic grounds It can be argued that less than perfect expectations Imply that agents are
irralional in some way (see Auerbach-Kctiikoff (1987 p 10) However the reverse argument can
be made One can question wiether agents are reaHy rationar and perfectly knowledgable about
future prices
Recent evidene of Balird (1987) Ballard-Goulder (19851 Goulder (19851 and Pereira
(1986b 1987d) suggests that the choice in modemng expectations is an important one They show
that the degree of foresight inl0 thc future (ranging from perfect foresight to-myopic expectations)
may have dramatic impacts on the pollcy conclusions of Ihe model Accordingly the best research
strategy may b to design models which are flexible enough 10 allow for different rules regarding the
formation of expectations With the exception of the articles just mentioned sensitivity analysis
bullhave previously not been performed along this dimension
In terms of implementation the two concepts of equilibrium - perfect foresight and temporary
equilibrium have different implications The dimensionality of the equ1ibrium soll1ion algorIthm
is involved Suppose we have a model with ten markets to be run for a period of 50 years Aside
from normalization a perfect foresight model implies computing prices In 500 dimensions while a
-t~mporary equilibrium model requires sorving SO equilibria each in ten dimensions Given that
computational speed often varies with the cube of the number of dimensions the lemporary
equilibrium formulation is potentially strikingly more feasible However Ballard (1987) and
Goulder-Ballard (1985) have devoloped techniques to greatly speed the computation of a perfect
foresight equilibrium
26
j
I
4 Jrnplemenlalion and Issues on policy evaluallon
41 Calibration and equilibrium comparlslon
CGE models at typically parameterized by the use of a calibration procedure Some parameters
are exogenously given However some crucial parameters are determined in such a way that the
model replicates the data for a given base year See MansurmiddotWhalleY 11984) for an eXlensive
discussion of 1~i$ issue
CalibratIon in a dynamic context is generaliy interpreted as requiring two properties First
replication of base year data is required Second the model is parameterized to simulate an
intertemporal balanced growth path when Ihe base policy is maintaineq This Is the approach
followed by Ballard (1983) BallardmiddotGoulder (1985) Goulder (1985) and SummersmiddotGoulde
(1986) It follows the practice of BallardmiddotFuliertonmiddotShovenmiddotWhalley (1985) with their
sequential cquilibrium dynamics
Other authors Auerbach-Kotlikof (1983 1984 1987) Boyenberg (1984 1985 1986)
and Pereira (198Gb 1987d) follow whal we cali a qualitalive calibration The slructural
palametars are exogenously chosen so that lh economy follows a reasonable path into Ihe fulure
ThaI has 10 do wllh the fact thaI given the recursive nature 01 the dynamic economy and aside tom
such structural parameters only iniUa stock values are needed to run these models Given initial
conditions on the stocks of say private wealth capltal and government debt agents will optimize
and Ihereby generate a lirst round of net demands and equilibrium conditions In lurn the
equilibrium prices will determine the evolution of the stock variables into the next perIod
There are several potential problems with calibration in the conlext of dynamic models The
assumption of a steadymiddotstate growth ~ath in the base case can be questioned First while
27 I Ibullbull
steadymiddotstaU is a possibity it certainly is not the only meaningfull solution to dynamic models
Even in the case of a perfect foresight equilibrium the model implies an equilibrium path which i may or may not involve colanced growth The model not the modeller should dictate the nature 01 I themiddotbase case path Second in the context of a temporarY equilibrium path a steady state solutlcn is Inol a likely model QutcOn1tl In fact unlessmiddot expectations are stattc short run behavior consistent
with a steady-state evolution will in general not be generated On the other hand if static I expectations are self~fufjili[1g we have in fact a perfect foresight model Third even the base year
replication requirement may cause problems In Ihe cOntexl of temporary equilibrium Any
calibration parameter would be conditional on expectation rules which is probably an undesirable
feature
The nalur of the two-requirement calibration strategy is very muchin the spirit of the bull
traditional design of the comparision of alternative equilibria comparis_ion between a steady-state
base case on one hand and alternative paths including a transition period and a finat sleadyyenstate on
the other hand Pereira (198Gb 1987d) compare different (not necessarily steadystate)
equilibrium paths The arguments against such a procedure are based on the idea that the impact of
policy changes can be observed most easily since all departures from the steadymiddotstate can be
attributed to the altemalive polioy On the other hand the base case so dEfined as a steadymiddotstate is
consist~nl with previous work in a less dynamtc setting and Iherefore allows a common standard
for comparing model results
42 Computation algorithms
We are still al a stage in which basically each author uses a different computational technique
Th development of dynamic models - in parlicular wilh adjustment costs andior perfect foresight
bull
28
has corresponded with the decline in the use of fixed point algorithms In -fact given the relative
arge dimensions inevitably involved such algorithms lend to be very inefficient at the best and
chen prohibitively stow See Stone (1985) and Preckel (1965) for a comparative assessment of
different computation techniques Among the models surveyed only Ball1rdFulle(ton~
-ShavenmiddotWhalley (1985) and Feltensteln 1985 use Merrills variant of the fixed point
algorithm 1echnique
AuerbachmiddotKoHikoff (1963 1984 1967) follow a three stage procedure They first compute a
base case steady-state then a revised case steady state and finally transition path for the economy
berveen these two steady-states tn all stages a Gauss-Seidel iterative procedure is used
Ballard (1982) Ballard-Goulder (1965) Goulder (HiSS) and Summers-Goulder (1986)
~
use a method developed by Ballard and Goutder which is similar in many awects to the FairmiddotTaylQr bull
(1983) algorl1hm Short-run equilibria are calculated (using Merrills algorithm) parametric on
nrice expeC~lions The model is then iterated to generate self-fulfilling intertemporaf expectations
and the corresponding perfect foresight equilibrium In a relatively similar approach Andersson
1987 uses a simulation program SIMNON developed In the University of Lund This program
can handle two-paint boundary problems In a fashion consistent with the multiple shooting
algcrlthm (see Lipton-Poterba-Sachs-Eummers (1982))
Boyenbergs computational approach (1985 1985) differs from the other models surveyed in
that he relies heavily on analytical techniques Computations are done by using a dynamic version of
Johansons linearization method Being essentially determined by the continuous time nature of the
I I
f
model thiS linearization model has the disadvantage of confining the analysls to infinltesinal changes
around the base case equilibrium (see Bovenberg (1985) p 53) Pereira (1986b 1987d) uses an optimization algorithm NPSOL - developped by
rGillMurray-Saunders-Wrlght (1986)) This algorithm is used 10 compute the sequence of
I
29
Shortmiddotrun temporary equilibrium which makes up the intertemporal equilibrium path The
equilibrium conditions are seen as nonlinear equality constraints in the minimization of an artlcial
objective function The prices are normalized to the unit simplex by an additional linear equality I I 1 bull constraint
II ErlichmiddotGinsburgh-Heyden (1987) follow a unique approach in that they use a variant of the
optimization technique introduced by Negishl (1960) The economic equilibrium can be generated i
as a solution of a mathemallcal program Ihe objective function of which is a weighted sum of the
utility functions of the various agents while tho constraints set consists of the market dearing 1
conditions Ginsburgh-Heyden (1985) have extended Negishis resulllO tho case of downward price
lrigidities I
Finally the paper by Jorgonson-Yun (1984) is also unique in tIlat it is the only I
econometrically estimated model Different blocks for the consumption and production Side of the
-
model are separately estimated to provide the necessary structural parameters i
I The diversity of computation techniques is yet another indicator of the exploratory nature of the
body of literature surveyed in this article
43 Equal yield comparisions
The link between a base case and counteriactual simulations Is usually provided by the concept of
equal yield Shaven-Whaley (1977) discuss the meaning 01 equal yield in a general equilibrium
context when government is not allowed to run defICits Equal yield is interpreted to moan constant
public utility Government base case utility is maintained in the counferfactual experiments With
balanced budgets the concepl of equal yield is unambiguous The new equUibrium prices and the
balanced budge condition will del ermine the minimal expendilure and taxes needed to maintain base
30
case publi~ utility Accordingly in general equal yield is inconsistent with equal nominal tax
revenue Some change in tax revenue Is necessary Different tax replacement schemes ate
considered to assure that enough tax revenue is collected This is the approach essentially followed
by most 01 the papers surveyed Only Feltenstein (1985) Goulder (198) and Jorgenson-Yun
(1984) chose not to follow an equal yield strategy
The question is of how to Interprete lila concept of equal yield when the government is alowed to
fUn deflcLts The optimal leve of expenditure for base ease public trliUty can now be taAinanced
bond financed or financed by a mix of bonds and taxation We have several versions of equal yield In
particular equal yield may now be consistent with equal tax revenUE Furthermore some meaSure
ltgtf financial crowding out effects of government deficns can be inferred from the comparislon of the
several equal yield alternatives These issues are extensively discussed ~Pereira (1987b)
44 Price expectations and the dynamic generalization of compensation
Indicators
The sum of equivalent and compensation variations over households is the most widely used
aggregated measure of efficiency gains or losses In the context of $teady~state eomparisions andor
when complete future markets exist andlor perfect foresighl is assumed there are no difficulties
associated with the use of the standard Hicksian indicators In fact correct future prices are known
in these cases
If expectations are not selfmiddot fulfilling andlor future markels are not available a dynamic
generalization is necessary 8allardmiddotGoulder (1985) provide some steps in that direction by
defining an indicator that accomodates periods far into the future when households do not have
perfect foresight but a steady~state prevails On he other hand this issue is more fundamental in
i
[
I
i I I I I [
I I
bullI
31
the contex~ of a Gner1 lemporary equilibrium ftamawor~ In such Circumstances Pereira
(1986a) dev~lops a cnamlc generalization which is obtained as the present discounted value of a i i
sequence of short-fUn optimal expenditure functions consistent with a base case expected future I stream of IlWities I
32
5 Empirical evidenee from selected policy Issues
Dynamic tax models have generated several important results which escaped Sialic modellers
The following is a seleCpoundd set of issuos whIch stress the marginal benefits of dynam~ modelHng of
economic behavior in innovltlttive areas of analysis The discussion of a consumption tax emphasizes
the benefits of dynamic tusehold behavior and intergenerational aspects The study of the
flliminatlon or the re~intToduction of the investsnemt tax credits is made meaningful by the dynamic
modelling of production betlavor Modelling governmenl in a dynamic conlexts allows the modeller
to study the irrrpact of government deflcits Finally a dynamic framework lets us appreCiate the I
relevance of consumer expectations in terms of the evaluation of policy alternatives
i
i
51 Consumption Tax
t
By Its very nature a consumption tax can not be adequately investigated with a static model bull
Fullerion-Shoven-Whaicy (1983) use Ihe model of the US economy s described in
Ballard-Fullerton-Shoven-Whalley (1935) to evaluate the movement from the currenl US tax Isys1em to a progressive consumption tax Since their modeJ Incorporates a Jaborleisure choice
where leisure is an untaxed commodlty their results reflect the fact that both the consumption tax
and Ihe present system are dslortionary
Concentrating on the rntertemporaf distortions they show that sheltering more saving from lhe
current US income tax could improve econornic efficiency even if marginal tax rate increases are
necessary in order to maintain government fevenue~ At first you have a revenue shortfall
However the economy moves to a higher sustatned growth path and uitimatefy lower tax rates
cangenerate the same revenue path Also wages increase in the long run with this policy The
bull
bullbull ~
33 shyt
i
I
I
present value of welfare gi[lS for a potiey of complete savin9 deduction wilh marginal tate
adjustments (consumption lax) is simulated to be around $500 billion to $600 billion 01 1973
dollars
They also investiga1e th 1ngth of ttme it takes th~ economy to adjust to these policy changes
Roughly they estimate the long run to b thirty years although this figure Is very sensitive to the
specification of the savings elasticity - a crucial parameler in this model
52 Investment T~x Credit
Goulder-$ummers (1 S87) address the impact of eliminating the lnvestmenl Tax Cradit (ITC) on
intersectoral capital formation and on economic growth Their model isYa1ticularly adequate to
addies$ this issu~ in that is postulates a forward looking investment m9we with adjustment costs
They show that the eliminating the ITC causes a reduction in the rale of investment Inveslment is
estimated to fall by about 7 in the short-run and by about 12 in Ihe new long run steady state
In lurn a previous policy anncuncement lowers Ihe overall al1ractiveness of Investment and leads 10
a downward shift of the Investrrent profile On the other hand the combined effect of a revenue
neutral simultaneous efimiraron of the ITC and reduction of the corporate tax rates is a long run
reduction ef the capital Sieck by 35 This pattern suggests that a revenue neutral increament in
both Ihe corporale tax rates and the inveSlment lax credits would be preferable 10 the formula
adopted in Ihe US Tax Reform Act of 1986
Pereira (1987d) addresses the impact of the ITe from Ihe oland point of Ihe curTenllax syslem
under the Tax Reform Act of 1986 Given Ihe concern over the potential depressiVe effects upon
savings and investment of the new tax system in conjunction with deficit reduC1ion mechanisms of
the GrammmiddotRudman type a pertinent question is should Ih lTC be e-introduced In the context
bull
34
-of his mod~1 of the US economy Pereira focuses on the trade-off between the distortion in the
intersectoral allocation of capital induced by the lTC the lower tax revenues it generates and
potential financial crowding-out effects on one hand and the positive effects of lowering the relative
price of new capital goods on the other hand Preliminary results confirm the qualitative results in
Goulder-Summers (1987) suggestting a welfare gain of about 2 of the present value of GNP in
the best scenario of absence of replacement taxes Furthermore preliminary results show an
important time pattern to welfare gains with the average benefits increasing the further you look
into the future This pattern is due to the presence of constraints to intersectoral mobility of capital
and inertia in the adjustment towards the optimal capital levels as reflected by the presence of
adjustment costs
I
I
bullAuerbach-Katlikaff (1gB) in the context of their intergenerational model of the US economy
consider the impact on savings capital formation and interest rates of deficit policies and balanced
budget increases in government consumption They conclude that deficit finance and government
consumption can significa-ntly crowd out capital formation and lower the welfare of future
generations However crowding out from deficit finance is a very slow process because it results
from increased government spending over potentially long horizons Also deficit policies may
substantially influence the longmiddotrun interest rates while leaving the short-run rates essentially
unaffected Finally and in opposition to the central role adjustment costs seem to play in both
53 Financial crowding outeff~cts of government deficits
Goulder-Summers and Pereiras models Auerbach-Kotlikoff suggest that the time path of interest
rates induced by a policy of deficit financing seems to be insensitive to the presence of adjustment
costs
bull
35
54 The role 01 consumer expectatlons
The expectations analysis has uncovered one of the benefits of dynamic modelling
Ballard-Goulder (1985) find that the appeal of adopting a consumption tax depends on the level of
foresight possessed by the consumers Furthermore additional foresight may be welfare worsening
This is a second best type of result This tends 10 occur under policies that lead to capital deepening
and declining rate of return to capital over time To the extent that consumers have more foresight
they will be beller equippod to anticipate the fall in the rental price of capital Consequently if a
saving incentive is enacted people save more with myopia than with perfect foresight Given the
exi~tence of taxes on capita and the discrepancy between the private an~--sccial returns to capital
the greater savings with myopia is beller socially Ballard-Goulder find that the welfare gains
from a consumption tax are reduced by about 10 when we move from myopia to farleet foresight
Therefore the attractivenness of adopting a consumption tax seems fairly [obust across these
foresight specifications
I
bull
36
6 Tbe power and weaknesses 01 dynamic modelling the example 01 corporale
tax Integrallon
The Issue of corporate tax integration allows us to sh~w the stre-ngths and weaknesses or the
dynamic approach Empirical evidence using CGE techniques indicates that depending on the precise
scheme of integration and the 1ax replacement methods integratiOn may have substantial effects In
Ihe work of FuliertonmiddotKingmiddotShoven-Whalley (1980 1981) total integration was found to yield an bull
annual static eHielency gain of $4 to $8 billion in 1973 dollars Simulated dynamic gains may be as
large as $695 billion or abcu 14 of the present value of tulUre consumption and leisure in the
US economy These gains result primarily from interindustry reallocalions of Investment and an
improved inter~emporal allocation of consumption
Mora recent work emphasizes hOw consumers asset portfolio decisions and firms finanCial
decisions affect the efficiency gains from integration Slemrod (1980) focusing on consumers
asset ponfoio deCisions finds static efficiency gains which are about twice -as Iarge as those
repOrted iiy Fulierton-King-ShovenWhalley (1961) FuUrlon-Gurdon (1983) focus on the
firms financial decisions They roporl efficiency gains of 6 of GNP from the elimination of the
tax distor1ions favoring debt However when they eliminate the corporate tax and repiace it with
increaSed persona income taxes addlHonal dls10t1ions are created in the optimal labot~leisure
decisions These distortions tend to dominate the analysis slJ(h that the overall effects of complete
integration are very modest Galpr-LuckmiddotTodr (1986) address simultaneously consumers
asset pOrlfollo decisions and the firms financial decisions They also find very modest efficiency
gains from integralion
The above results are important but may be severely biased There are severa aspects of
economic behavior and modelling crucial for the study of Income tax Integration whiCh have no been
i 37
captured innny of the above papers One aspect is the absence of government deficitsmiddot a balanced
budget is assumed It is true Ihal resource crowding OUI is caplured In these mOdels but financial
crowding out induced by government spending is nOI Second investment is not derived from
optimtztlcn behavior Investment behavior passively accomodates endogenous saving decisions As
a consequence the differential impacts of policies in the incentives to save and to invest are not
captured Aso full capital mobility across secors is assumed with instantaneous capital
adjustmenlS towards optimal levels This assumption rules qut different costs of capital across
seclors and therefore differentiated reactions to tax policies changes Finallyhe modelling of both
gov9rnmerlt deficits and of endogenous real and financial investment decisions necessitates the
~nsideraiion of a dynamilt framework and the introduction of financiaJ assets govemment bonds and
iprivate financial assets A dynamic framework also highlights the efficiency effects of Integration on r
ithe ptimal intertemporal decisions The above models are either sIalic (Slemrods and
GalpermiddotltJckemiddotTode(sj or a dynamic sequ~nce of otherwise Slatic model bullbull I Pereir a (1986b) develops a dynamic applied general equilibrium model bull to study the tbull
efficiency effects of integration on the growth and allocation of investmenl across sectors Special
Iattention is paid to the structural affeelS of resource and financial crowding oul induced by
gove mect spending and deficits and to the real and financial Investment reactions across industries
to both tax integration and flnancial crowding out His mode departs from previous work on income
lax Integraticn and for that matler from most 01 the CGIE literalure for tax policy evaluation in
several direcllons It encompasses an endogenous sequential equilibrium struelure founded on
dynamic behavior with flexible expectations Government deficits are optlmaUy determined
Investment decisions are tward looking and the resull of optimizing behavior Several financial
assetsmiddot public and privatemiddot are considered
Simulation resulls Indicale that the welfare gains from Integration under large deficits are at
32
best lJlodest when measured in terms of the GNP The elimination of the corporate tax and its I replacemenl by increased income lax rates yields long run benefits which are never larger than
2 of the presenl yalue ct telr consumption and leisure onder the previous tax regime and 1
under the current tax regime However the short run w~lare effects of full integration tend 10 be
very low and in in some scenarios even negative This is a flew intertemporar pattern of efficiency Ieffects which reflects an adjJsiment lag in the interindustry investment decisions due to the
existence of costs of adjustment On the of her hand partial integration achieved by excluding I dividends from the corporale lax base syslematically yields negalive efficlencyeffeclS In Ihe long Inm these can be as high as 3 of present value of the future value of consumption and leisure This
is a new second best effect suggesllng Ihat less than complete integration may have perverse I efficiency effects
The dynamic slructure of PerclrBs model snowing for an improved freatment of real investment
declsions and government deficts provides a potential setting for a more accurate measure of the
costs and benefits of inlegration The simulation resullS suggest lower benefits and an intertempora Ibull
pattern of growing benefits
Simulation results also clearly suggest robust negative effects from partial integration How
reliable is this result If dJvidends were deductible from the corporate lax base (as are interest
payments) the preferenlial Ireatmant of debl over equily would be ellminaled Corporallons should
be expecled 10 react by decreasing the optimal deWeqully rallo However corporale financial
gecisions are exogenously set In Pereiras model as in aft of the other models surveyed that go as far
as dealing with the issue
Also withdrawing dividends from Ihe corporate tax base is a way of eliminating Ihe double
taxation of dividends at belh Ihe corporate and personnal Income levels How will the optimal
aliocellon of household saving accomodate that change in Ihe relalive return to corporale equity
39 I
t However hOusehold portfolio decisions are exogenouly set in Pereiras model as in all of the other
models surveyeltj go as far as dealing with the issue
IIt is sate to say that the evaluation 01 the benefits 01 partial Integration as defined above are
sevrey underestima1eO Meanwhile the distortionSoenerated by the increase in the marginal
personal tax rates designed to ralse the revenue foregone by the dividend exclusion are fully
aCCltlunted for A good measure of the costs of integration together With a poor evaluation of the
benefils may very well be the reason why partial integration is Simulated 10 yield neg alive
eHikiency gains
On a different vein as most of Ihe model surveyed here Pereiras is a closed economy model It
is legitimate to wonder how the resullS would change If international capital flows were anowed
IThis almost certainly would affect financial crowding out since a good d~al of the capital inflows
could be used to finance public debt
For these various reasons the results should be treated as preliminary but they strongly I suggosl that the dynamic structure provides valuable new insights into the corporate tax integration SSU2
40
7 Concluding remarks I I I i
What has been acccmplished Ihis far The success of Ihe CGE research in laelding the challenge
of dynamic modelling can not be denied
I iGreat progress in the modelling of household behavior has been achieved Promising
developm~nts in the modeUing of production and government behavior have also been accomplished
Interesting innovations in the concept of equilibrium and computation techniques were discussed
The policy analysis was greatly enriched by considering transitional effects together with longer
run steadymiddotstate equilibrium paths generalized equal yield strategies accomodating different
financial crowding out impacts and generalized dynamic policy evaluation indicators
An important set of issues have been addresed by the models surveyed Traditional issues
focusing on the effects of capital taxation and consumption taxes have been pursued In terms of
capital taxation for example the intertemporal nature of the issue was enhanced by allowing an
optimal evolution of capital stock in the economy and forward looking optimal investment decisions
In Andersson (1987) GouldermiddotSummers (1987) and Pereira (1986b 1987d) this is coupled
with a improved treatment of several tax provisions like investment tax credits and depreciation
allowances In terms of the consumption taxes developments in the specification of Intertemporal
household behavior the introduction of overlapping generations and the modelling of
intergenerational links via bequest motives as well as a closer attention to demographic evolution
provided a much improved economic setting for the understanding of the several aspects of the
problem
Furthermore dynamic modelling has permitted the addressing of a set of new issues Policy
Issues ranging from dynamic tax policy analysis to the evaluation of exogenous changes in
government expenditures to the impact of different methods of financing government expenditure to
4 1
the importance of tinanciai croHrii1g out to 1he relevance of consumorS expectations in tne policy
results have been addressed
Despile all of the progress and in part due to such progress several avenues are wide open for
much neded additional research The following seem to be the most Interesting and promising
areasmiddot First the specincation of liquidity constraints to household behavior may help to obtain a
better understanding of policy chenges that affect directly the interest rates in the economy Second
major attention needs to be paid to the incorporafion of a roraign sector (with open capital markets)
Third some efforts are needed in terms of finding adequate computation techniques now that
dynamic modelling eally makes the curse of dimensionality worse than eVOrbull
The modelling of financial markets is the single most unsatisfactory speet of the dynamic
models surveyed here~ The problems associated with provlding a mo~ adequate treatment of
financial markets and in general endogenous and optimal financial deci$ions bull both at the corporate
and at the personal levelsmiddot have 10 be addressed That necessarily requires Introducing uncertainty
into the models To b adequate this must go well beyond the mere assumption of point expeetalions
with the whole array of modelling implementalion and evaluation problems thai generates
bull I i
42
REFERENCES
Abel A and O Blanchard 1983 An Intertemporal Model of Saving and Interest Econometrica
51 pp 675-92
~ Andersson K 1987 Tar-ation of Capital in Sweden - A General Equilibrium Model Ph D
Dissenation UnIversity of Lund Sweder
Auerbach A L KotlikoH 1983 National Savings Economic Welfare and the Structure of
Taxation in M Feldstein ed Behavioral Simulation Methods in Tax Policy- Analysis Chicago
University of Chicago Press
Auerbach A l Kotlikoff 1984 Social Security and the Economics of Demographic-
Transition in H Aaron and G Burtless eds Retirement and Economic Behavior Washington DC
Brookings Institution
Auerbach A L Kotlikoff 1987 Dynamic Fiscal Policy Cambridge Cambridge University
Press
Auerbach A L Kotlikof and J Skinner 1983 The Efficiency Gains from Dynamic Tax
Reform Internatiooaj Economic Reyiew 24 pp 81middot100
Ballard C 1983 Evaluation of the Consumption Tax with Dynamic General Equilibrium
Models Ph D Dissertation Stanford University
Ballard C 1987 Tax Policy and Consumer Foresight A General Equilibrium Simulation
Study Economic Inquiry 25 pp 267-284
Ballard C and L Goulder 1985 Consumption Taxes ForeSight and Welfare a Computatonal
General Equilibrium Analysis in J Piggott and J Whalley eds New Deyelooments in Apolied
General Equilibrium Analysis Cambridge Cambridge University Press
Ballard C D Fullerton J Shoven and J Whalley 1985 A General Equilibrium Model for
t 43
Tax Eoliey Eyalua~iQO NBER University or Chicago Press
Borges A (1987) bull A Survey of Compula1ional General Equilibnum Models CECD Economic
Slud~ 8
Bovenbig L 19B3 imperfect Capital Mobility in a Perfect Foresight Herberger Model
Mimea UC Br~eley
Bovenberg L 1984 Capitel Accumulation and Capital Immobility Qmiddottheory In a Dynamic General Equilibrium Framework PhD dissertation University of California at Berkeley t
Bovenberg L 19850 Dynamic General Equilibrium Tax Models with Adjustment Costs in A iManne ed bull 1985 EtQnomrc ECUjilbrium Model Formulation and Solution Mathematical
lProgramming Siudy 23 Arnsterdam NorthmiddotHoliand
Bovenberg L t 1986 Capital Income Taxation in Growing Open Eeondmies Journa of Pubtic
Eamp(omlcs 3t pp 347middot376 IBovenberg L ald W Keller 1981 Dynamics and Nonlinearlties in Applied General
Equilibrium Models Internal Report Mlmiddot81middot208 Department for Statistical Methods CBS bull
Vooiburg The Netherlands
Charnley C 1981 The Welfare Cost of C~pltal income Taxation In a Growing Economy
IJournal of Political ECQnQm~ 89 3 pp 468middot496 I
Charnley C 19810 Efficient Tax Reform in a Dynamic Model of General Equilibrium I Mimeo Yale University I Charnley C 1982 On the Optimal Taxation In Stylized Models 01 Inlartemporal General
Equilibrium Mimeo Yale University I Decaluw B and A Manens 1985 Pays en Oeveloppement at Modeles Calculables OEquilibre
General - Une Revue de al litterature empirique Monographie Sene 9 Volume 3 Centre de
Recherche et Developpemont Economlque Unrslt de Montreal
44
bull
I I
I
I
1
Erlich $ V Ginsburgh and L van der Heyden 1987 Wher~ do Real Wage Policies Lead
Belgium bull A General Equilibrium Analysis European Economic Reyiew (forthcoming)
air R and J Taylor 1983 Solution and Maximum Likelihood Estimation of Dynamic
Nonlinear Ratioant Expeclations Models Ecooometrica 51 pp 1169middot1186
Feltenstein A 1984 Money and Bonds In a Disaggregated Ec~nonYin Searl 1-1 J Shoven
Eds Aoolied ~eoeral EqJiliDCiulD tioalYsis Cambridge University Press Cambridge
Feltenstein A 19B6 A Dynamic General Equllibrlu Analysis of Financial Crowding Out
Theory with an Application to Australia Journal of Public Economics 31 No1
Fullerlon D R Gorden 1983 A Reexamination of Tax Distorllons in General Equilibrium
Models in M Feldstein Ed Behayora Simulation Methods in Tax pOlicy Analysis Chicago
University of Chipgo Press
Fvllerlon D J Shaven and J Whalley 1983 Replacing the US Income Tax with a
Progressive Consumption Tax JoulOal of Public ECQnomics 20 pp3middot23
Fullerton D Y Henderson J Shaven 1984 A COmparision of Methodologies in Empirical
General Equilibrium Models of Taxation in Searl H J Shoven Eds ~pQlied Geoaral Jguilibrium
lIoalyss Cambridge University Press Cambridge
Fullerton 0 A King J Shaven and J Whaney 1981 Corporate Tax Integration in the United
Statesmiddot a General Equiiibdufi1 Approach American Economic Review 71 pp~ 677~691
Galper H R Lucke E Tader 1985 Taxation portlor Choice and the Allocation of Capital A
General Equilibrium Approach Mimeo
Gill p W Murray M Saunders M Wright 1986 Users Guide for NPSOL (Version 40) A
Fortran Package for Nonlinear Programming T echnicsl Report SOL 86middot2 Deparlment of
Operations Research Stanford University
Glnsbur~h V and L van der Heyden 1985 General Equilibrium with Wage Rigidities An
45
AppJication to Belgium in A Manne ed t Economic EQumprit1m~ Made Formulation and $oJioo
Mathematical Programming Study 23 Amsterdam NorthmiddotHolland ~
Goulet J 1968 1 Adjustrent CO$1S in the Theory o(the Firm Reyiew of EconomiC Studies 35 bull
pp47middot55
Goulder L 1985 Tax-Financed versus BondmiddotFinancedGhanges in Governent Spending
bull Mimeo Harva(d UniversilYmiddot
Goulder l J Shaven and J Whalley 1983 Domestic Tax Policy and the Foreign Sector The
importance of altemative foreign policy formulations to results from a general equilibrium tax
analysis model In ~n methods in Ibe antico of income from caoia1 ed Martin
Feldstein Chicago University of Chicago riSS
Goulder l L Summers 1987 Tax Policy Asset Prices and 3rowtly ~
A General Equilibrium Analysis NBER Working Pper No 212B
Grandmont J 1982 Temporary General Equilibrium Theory Ch 20 in Handbook 01
Mathemalical Eccnornics Vol II NorthmiddotHoliand
Harberger A 1959 The Corporate Income Tax An empirical Appraisal In lax Revision
Comoendjum Voll House Cornmitee on Ways and Means Washington DC Govemment Printing
Office
Harberger A 1962 The Incidence of the Corporate Income Tax JOULn1 of Political
Economy 70 pp 215middot240
Harberger A 1964 Taxation Resource Aliocalrlon and Welfare fn lb Role of Dir bullbulll and
Indjresl Taxes io tbe Fede[al Reserve s~I~Ill Princeton Princeton Unlveresity Press
James J 1985 New Developments in the Applications of General Equilibrium models to
Economic History in Piggott aand Whalley Eds
Jorgenson D and K Yun 1984 Tax Policy and Capital Allocation Harvard Inslitute of
bull
I
I f
Economic Research WP 1107
Judd K 1965 ShonmiddotRun Analysis of Fiscal Policy in a Simple Perfect Foresight Model
Journal of poJiHcal 5CCJiW1l pp 298middot319 r
Upton D J Potbc J Sachsa nd L Summers 1982 Multiple Shooting in Rational
Expectations Models fuoometrlc 50 1329middot1333
Manne A ed 1985 fconQrnjc Eoullibrium Model FOrIDujaljQo and Solution Mathematical
Pogrammlng Study 23 Amsterdam NorthmiddotHolland
Manne A 1985 Or the Formulation and Solution 01 Economic Equilibrium Models In A
Manne ed ECQOQlli-Jr~ibmiddotum ~tQdel EQlJulailon and Solution Malhematical Programming
Study 23 Amsterdam NorthmiddotHolland
Mansur A and J Whalley 1984 Numerical Specification of Applied General Equilibrium
Models in Scarf H J Shoven Eds Aoplied General Eouillbrium AnaiysectIsect Cambridge Universily
Press Cambridge
Merrill 0 1972 Applications and Extensions of an Algorithm thaI Computes FixedmiddotPoints of
Cenain Upper Semimiddotcominuou Point to Set Mappings PhD Disserlation Unlvel$ity 01 Michigan
Negishi T 1960 Welfar Economics and Exislence I an Equnlbrlum for a Competitive
Economy MelroeCOOQ11ka 12 92middot97
Pereira A 1985 On the Existence and Uniqueness of Optimal Ou1put Path lor CRTS Firms in
Adjustment Costs Technologies Mimeo Stanford University
Pereira A 1985a Consumer Expectations and Ihe COSI 01 UllUty In an Intertemporat
Framework Mlmeo Stanford University
Pereira A 1985b Corporale Tax Integrallon and Government Deficits A Dynamic General
Equilibrium Anaiysls Mlmeo Stanford University
Pereir A 1987 On Ihe Computation of the Users Costs of Stocks Ecooomi (forthcoming)
47
Pereira A bull 1987b Equal Yield Alternatives and Government Oeficlts Mimeof Stanford
University
Pereira A 1987c A Dynamic Applied General ~uHlbrium Model for Tax Policy Evaluation
Comlletamp Docurnent2tion Mimeo University of California San Diego (in progress)
Pereira A 1987d Should the investment Tax Credit be Rei(ltroduced Mlmeo University
of California San Diego (ir progreSs)
Piggott J and J Whalley eds 1985 tJew QeyeQpments in Applied General EQuilibrium
bullenalysjs Cambridge Cambridge University Press
Preckel Pbull 1985 Alternative algorithms for computing economic equilibria In A Manne Ed
Radnor R 19amp3 Equllibrium under Uncertainty In K Arrow M Intrilligaor ods
I ibull
Robinson S 1 g86 Mullisectorai Models of Oeveloping CountrieS a Survey Oivision of
Agricullure and Resources UC Borkeley WP 401
Scarf H 1967 The approximation of fixed points of a continuous mapping ~MM Jouwal of
epplid Mathornallcs 15 pp 1328middot43
Scarf H with T Hansen 1973 00 the Computation of EconomiC Eguilibria New Haven Yale
University Press
Scarf H J Shoven Eds 1984 61l1llied Genera Elujlibrium 6nalyssect Cambridge University
~ Press Cambridge
Shoven J 1976 The Incidence and Efficiency Effects of Taxe on Income from Capital
Jurnal of ramie1 Economy 84 pp 1261-83
Shoven J 1983 Applied General Equilibrium Tax Modemng IMF Slaff PaoerS 30
Shoven J and LB Simon 1986 Share Repurchases and Acquisitions An Analysis of which
Firms Participate NBER Working Paper No 2243
bull
48
IShaven J and J Whalley 1972 ft General Equmbrium Calculation of lhe Effects of
Differential Taxation of Incoma from Capital in the US Jauroe of public Econgmics 1
pp281middot321 (
Shoven J bull and J Whalley 1973 General Equilibrium with Tax$s a computation procedure
and an existence pr00f ReviEhi of Economic Studies 40 pp 475middot490
Shoven J J Whalley 1977 Equal Yield Tax Alternatives A General Equilibrium
Computa~ui0nal Tfchnique J~HJrnal of PUblic EconomIcs 8 pp 211~224 bull
Shaven J J Whalley 1984 Applied GeneralmiddotEquilibrlum MOdels of Taxalion and
International Tr8de An Introduction and Survey Journal of Economic literature 23t pp
1007middot1051
Slemrod J 1980 A General Equilibrium Model of Capital Income Taxation Ph D bullDissertation Harvard Univnrsity
Slemrod J 1983 A General Equilibrium Model of Taxation with Endogenous Financial
Behavior in M Feldstein edbull Behavioral Simulation Methods In Tax Policy Analysis Chicago
UniVerSity of Chicago Press
Stone J 1985 Sequential Optimization and Complementarily Techniques for computing
Economic equilibrium in A Manne Ed
Summers L 1981 Taxation and Corporate Investment a q-Theory Approach BroQkinas
Summers L 1985 Taxation and the Size and ComposHion 01 the Capital Stock An Asset Price
Approach Mlmeo Harvard University
Whalley J 1975 A Geneal Equiloibrium Assessment of the 1973 United Kingdom tax
reform Economica 42 pp 139middot61
1JLTIllOS JIQll~)ltG PAPERS PUBLICADOS bull
I
I I
j
112 70 - LUCENA iJiCIIJo ~ IIEnvironrilent Honitoring and Orqanization Structure T (Setellibro 1987)
nI 71 - L~ENA Diogo A Note on the Representation of Information Stxueture (SeteJObro 1987)
nQ 72 - LUCENA D1CgO Envuorment 111)nitoring and Organization
Structure II (Setelilhro 1987)
ril 73 - AlmmE3 Antonio Pais e CAsPAR Yitor bull Tributacao Inoentivos e InvesUlllentos Analise Qualitativa noYellibro 19S7)
n2 74 - liAlltOSA rose Pedro Opt1lllal lIage R1gidity A SUggestedtlthodology to Teet tb Theory and an Application (Outubro 1987)
nI 75 - CABRAL L 11 B Three lIotes on SYJIIlletric Gemes witt itsYJIIlletric Equilibri (NoveJObro 1987)
n2 76 - LUCENlt D10gOTo Search or Not to Search (Fevereiro1967)
nQ 17 - SA Jorge asconcelos euroI A Theory of SynergyB (Fevereiro1987)
n2 76 - VILARES tanuei Jose Os Be IntenaMios IlIIportados ColIIO Factor de Produao (Julho 1987)
nQ 79 - SA J~rge VasCQncel03 e HoY to Copete and Co~unicate in tature Industrial Products_ (aio 1988)
n2 80 - ROIl Rafael Learning and Capacity Expansion in a New larket Under Uncertainty (Fevereiro 1988)
n2 81 PEREIRA Alfredo llarvilo Survey of Dynamic Computational General EquilibriUlll Hodel for Tax Policy Evaluation (Outubro 1987)
Qualquer in ormacao sabre o 1IIorlltlng Papers ja publlcados sera prestada pelo Secretariado de Apoio aos Docentes podendo os ~e3OSf
ser adquirido5 ne Secyo de Vlndas da Faculdade de EconollLia UNL na Travessa Esteviio Pinto Cltpolide - 1000 LISllOA
-----
------
- ---
-
IlflpnUItll tit lun pVII-y PIlUAlIV
q6H1brla oqual fax polley ovnluaU-on efUclency derlvatton Ioorllhm
calibration param~hr computation COmelfed llold tndfcaHH eHecls
Bnl jard~FuU6rtQ n base year back soll1lion steady-sta1e yo inlsrgroup inerseel or Iii
VS transilion lump-sum with Interfemporai 1965
~Shovn Whallmiddotv replication wllh exog Morrills nmamo1ers alQorilhm _~ond~middotsta1$ ears Inc CVor EV
i~noerucn sleady-sfala no agg(ogaled lnlalsectora 1967
base year back sobHon SiMNON package replication wHn oxog valiant of VS Iransilion no Indicamptor lnlertamporal
paramelsrs multiple shoo ling 6~~~l~slat8
AU8rt)ach~Kotllkot steady-slato yo Inlerguna-r a lions
1893 1llH
quatitaUve by replication a$Sium~lkml Gauss-Soidel liS Uln1WOn any rule -lilh 1111ermpOrB
sonsililJ1y I _ ~poundI~Jjnnelill $t(lfdvmiddot~luiA J3djlt15S ~~ nnrilW ~--------
a ail r nimiddotGould or bas9 9tH bac~ solution Ballard~ staadymiddotslalo )05 kltHSclo(a~
1983 1985 VS Iransllinn lump~sum dynalTlk CV inlerlemporal steadY slafe parameters alaorlthm replicaUon+ with oxog Gourder
+stead~stato (ers loc inlareel10rat BCvonborg qunmaiive by dynamic version stoadY-$Igtto yes 1985 1985 YS transition some govt eN Inlortempora
S9rHltivilv OnearizaJian method ropllcaUon assumptionl of Johansons
+steady-sfata paoo palh Erllch-Glnsbu rgh~ actual vorsus no aootagalod inlersectoral2 bftse years back solu1ion variant of Negishis
Heyden replication wIth alltogbull optlmizaUon countorfaclUal no ~odjcator inlottempora 1987 paramlJters aocroach
faltonstaln baso yeat back solution actual versus no aggrampgalfld intersElclornl replication wilh exog Merrills counlerfaclutll no indlclltor middotlntamprlempOull1984 19B6
parameters al90rithm 1981middot1982
Gouldar staady~slala Inlar5ecioralbase year back souUon Variant or 1S lronsilion no dynamic CV inlortempora
stoady state parameters Newlon 1985 feplicalion+ with iUog Fair~Taylorl
+slaildy-sla1e Intargeoorat Gtlulder~Summers base year back solution varIant of stlJadymiddotstate fixed no aggregated fntersectoral
vs trarlsition government Indicator Intertempo1al steady state oatameters Newton
repticaUon+ with exog Fair-T aylQf 1967 +ste-adv-stata bullbullbullndlno financial
Jorgenon-Yun econometric steady-stale Money melfic In1ersecloral
- - 8slimalion vs traosUian no social weUare functton intertampotal1984 +~teadlmiddotstatl)
~ shy
qtJalitaUv8 by NPSCt optimal path ye generalized inlersectorai 1986 1987 Perelr
replication 8S$umptlonl optimizatIon V$ lumpmiddotsum dynamic lntertemporat se~~ttilitf ____Jllgorilhm opllmal pah pen inc CVoclV financial
- 9 shy
bull ~ ___ - _______ - ____ bull___ OM bullbull __ bullbull __~____ ~_
--
- - - --------
--------- - - - --------
---
1nltJ1 bull I ~ 1J11ltI1 v I ~
-
Sa Ita rd-Futlertonmiddot intargration consumption lax ~StHlven~Whall~y VAT capila~ or labor tax
Ch~mqflS1985 Andersson
consumpion tax neutral corporals lax system 1987
it ue rb a c h~ i(at II Sltotf COIlSUft1ptian lax (poundpird ItS hibor Mx~on ufocto of taxation on capital fOfrnJilti1
dvnamio fiscal policv effects af OQvernmanl defioit (3tlard-Goulder
1883 1987
allocts 01 the adoption of a 1ge3 1985 pure consumpHon lax
im~rtmce of cons fCreSClht -Bovenborg analytically orianiad approach
1985 1986 effects ot pure conumption tax capllal income taxation in Opo economies
Erllch-Glnburghshy-Heyden reat wage policios in Belgium
1987 Felter1steln 1984 - theorotical
1985 ~ linnncial crcwdlng out1984 1986
Goulder effects of tax- vs bond-financed 1985 chanQes In pubtlc expondllufe
~~--
Gouidr~SUmmampfI torpora1$ lax cvts reduced 1987 ITO increase 10 gasolioe taxes
Jorgenson-Yun
effects ot several programs of tax reform on the allocation of ~~ptai
1984
-- Purelra
corporate la)( Integration effects cf romiddotinlroducinQ imreslamnl tax credits
1986 1967
10 bull
bull -~- ~- --- _ - ---____--_ - -----_-_ --~ ---- ~
1 1
2 Brief overview of non-dynamic CGE models
The origin of most or the models we survey goes back to the work of Scarf (1967 1973) whlgt
first developed a reliablo algorithm to compute equilibrium price for a ArrowmiddotDebreu economy
HiS algorithm used simpicial Bubdivision techniques and can -be shown to be the computational analog
of the fixed point theorems previously used to prove the existence of equilibrium His technique
could soW a model with an arbitrary number of consumers and commodities as long as all agents
were price takers consumers were subject to budget constraints demands were continuous andmiddot
production did not display increasing returns to scale The alQorithm while guaranteed to converge I
I I I
was relatively slow for prcb1ems involving more than S1If 20 dimensions A maior improvement in
computational speed Woo offered by Merrills algorithm (1972) which used tpe same fundamental
ideagt as Scarfs procedure
Scarfs model (as is true for the standard Arrow-Debreu model) does not include a government
sector ~ neither taxes nor public goocs At one of the mosi promising appHciUions of the new
computational techniqu w~s in Iha area of tax policy evaluation Shoven and Whalley (1973)
exlended the general equilibrium model and computational approach to include a wide rry of taxes
and a governl11ampnt spending plan The original ShavenmiddotWhalley model was static (although the
different commodi~ies CQuid be considered simitar goods available at different datest as In
Arrow-Debrau) and government was aumad to run a balanced budget The data are arranged as a
social accounling matrix The models spacification and calibration is checked by solving it in the
presence of the base set of laxes The result should ba exactly tha initial social accounting matrix
After having pasd this replication check the model is solved for a counterlactua equilibrium In
the presence of a new tux design The result is once again a social accounting matrix The two
equilibria are compared in order to tSS($S the impact of the new tax plan The first uses ot this
bull
1 Z
model for lax policy evauajon were ShovenWhalley (1972) Whalley (1975) and Shoven
(1916)
Completely static models as Shaven-Whalleys are unsatisfactory for many tax reform ISsues
These include corporate lax Integration effects of investment tax credits effects of accelerated
depreciations consumer Or e)penditure tzxes~ importance of saving subsidies like IRAs etc These
are essentialy dynamiC ksues They involve not only the allocation of capital across sectQfsbut
perhaps more Important the capital intensity of the economy But the capital accumulation and
capital reallocation take time and may involve adjustment costs Because of these issues Ballardmiddot
Fulierton~Shoven~Whaloy took the first steps towards developing a dynamic model
In this conlexl the ShovenWhalley model has been extended and implemented for Ihe US
-economy In BuIIBrdmiddotFulienonmiddotShovenmiddotWhalley (1985) While Ihis book completely documents the
mo~el it was used in several publicalions beginning in 1978 Their mode) consists of nineteem
production sectors twelvo households and fifteen consumer goods It includes a very detailed set of
taxes including the federal and state personallncome taxes federal and state corporate income taxes
SOCial Security taxes pro1J0rty taxes unemployment Insurance excise taxes sales taxes etc It
has been calibl att-d to reproduce the 1973 US economy Recentty a version corresponding to the
1983 JS economy has been developed
wThe 8alard-Fullenon-Shoven Whalley model is dynamic In the sense that consumers face a
choice between current consumption and leisure versus future consumption (which can be
purchased via s wlngs) The consumer classes act as if they were maximizing a nested CES utility
function over the domain of a CES aggregate of contempo-raneous consumer goods and leisure and
future consumption subject to their income constraint The parameters of those funcHons
determine the shares of income devo1ed to each commodity to saving and to the purchase of leisure
They also determine the two key elasticities in the models ~ the elasticity of labor supply with
bull
13
respecl to tgte rbullbull1 aoer lax wage rae and the elasticity of saving with respecl 10 the real after lax- rate of retum to capital
In the model consumerS have myopic expectations regarding lulure prices and in particular
regarding the future rale of return 10 capilal Ballard (1983) and BallardmiddotGoulder (1985)
inco~ora1e both perfect foresight and limited foresight into this model Future consumption is
acquired by buying a fixed composition portfolio of real investments that offer an infinite annuity
of returns
The production side of Ihe model is compielely static The mod1 Incorporates a constantmiddot
olasicily 01 substitution betweeen prlmary inputs in production (capital and labor) and fixed I coefficients for intermediate inputs The model distinguIshes between Industrial outputs and
eonsumer goods for the simple reason Ihat tile data are classified differenUy Industrial sectol$ I I involve such categories as forestry and fisherle metal mino9 and publishing and printing while
consumers purchas-3 fUfHi1ure 2ulomoblles and books This fact is recognized in the model by Ibull
incorporatirg a secord stage Of production which converts industrial outputs intO consumer goods
This technology is uStally mod9-led as a flxed~coefficient conversion matrix
The BallardmiddotFINton-ShovenWhalley model assumeS that the private sector finance
marginai ioveslment with the same composition of debt and equity as currently exists in each sector
This is the same arsumrtion that Herberger originally used Investors all hold debt and equity in tile
same proportion Thampreiore this ownership can be aggregated into simply capital ownership For
tax purposes however the separate treatment of debt and equity is taken into account at both the
corporate and persofa Io=vel SimllarlYJ the dividend poIicies on corporale equitY are established
exogenously There are no government bonds in the model since there are no government deficits
The modol is solved for a sequence (as many as 100) of temporary equilibria wltll consumers
allocating income between present and future conumption at each point in time The palh for 1he
14
economy ~ a sel of connected equilibria The connectIon is pra~ided by capital accumulation
Capital accumulation is endogenous and determined by saving The model stans with a sociagt
accOunting mi1rix In the bzse case the economy is assumed to be in a steaqy state growth -path (along
which all relative prices ~fe constant) The madel solves for both the naw steady state growth path
and the transltion 10 it zf~er a policy intervention rhe authors have frequently addressed the
question-of howong rt tKes to effectively settle un10 a new steady S1ate growth path
The dynamics of the model are limited however in that future consumption is collapsed into a
composite commodity Aso the absence of government deficits and the lack ofproductlon dynamics
limits the realism of the BallardmiddotFuliertonmiddotShoven-Whatley model for the analysis of dynamic
policy issues (such as the adoption 01 a consumption tax or Ihe elimination of the investment tax
credit) The models we survey try to Improve upon Ihe Ballard-FlIiertonmiddotShaven-Whalley
dynamic modelling The models Jn~orporate more forward lookTng behavior They improve the
dynamics on the conampumrtkm side of the economy and some of them introduce true 1ynamlc behavior
on the part oI the producers We turn now to the issues inY9lv_ed in developing such dynamic models
15 I I ~3 ~ssues in the modelling Of dynamic behavior
We will 100k firt at the speclficalion of economic behavior of consume(s~ producers
goyemment and Ihe r~sl of Ihe world The modemog of corporale and household financial decisions
I is then addr5sed Finally Ihe eencept of equilibrium is discussed
I
i
I 1
31 Consumers behavior
Early efforts 10 build dynamic features into the economic behavior 01 consumers are due to
Ballard (1963) and Auerbach-Kotllkoff (1983 1984) Their work has been closely followed by
several subsequent authors In faet dynamic household behavior Is the single most pervasive aspect
amohg the models surveyed
All the models Inc-O1orale some loro of lila-cycle behavior Household behavior is dewmloed
by the maximization of an addWVely separable time invariant intertemporal utillty function The
utility lune1ion is defined oyer the domain of the consumption goods In the economy In most of the
models leisure is also an argument in utility so that labor supply is optimally determined
Utility maximi4ation is subject to a lifetime intertemporal budget constraint which equalizes
the presen1 value of consumers income and expenditure More reeentiy in Andersson (1987) bull
Bovenberg (1985) and Pereira (1986b 1987d) the constraint i$ defined as a sequence 01
recursive equations of motion on wealth This has the potential advantage 01 accomodating liquidity
constraints However It should be recognized that in the absence 01 liquidity constraints (ie when
consumers are free to completely borrow against future income) the two specificallons of the
househofd ccnstraint are essentiay equivatent Furthermore tn both versions saving is optimally
determined as a way of translaring wealth intertemporally
1 6
Some lodels now have a sophisticated dynamic specification of hoosehokl In Ballard (1983) -
BallardmiddotGoulder (1985) and Auerbach-Kotlikof (1983 1984 1987) consumers behavio is
embedded in an iterQEneraional setting Each of them has S5 age cohorts simulataneously alive
Som-e measure of jntergener1ional altruism is considered in the form of bequest motives In Ballard
(1983) and Ballard-Gouldr (1985) consumers derive utility directly from bequthing parl of
their wealth Therefore bequests assume 1he form of a scrap value of terminal wealth of a
generation In AuerbachmiddotKotlikof (1983 1984 1987) each generation empathyz bullbull with Ihe
utility of future generations ovar bequethd wealth In addition Ballard (1ge3) includes detailed
demographic poiections wfill the intent of incorporating in the policy analysis the effects of the
bull Post World War tI baby-boom
In the real world household decisions also include the optimat allo~atiOn of saving among
shyalternative physical or financial assets Theories of household portfolio behavior are relatively
Iwen~establshed However they involve uncertainty as a crucial element Accordingly they have
middotnot been incorpora~ed in the context of the determinIstic dynamic models we sutyey here We will 1 Come backt-O this Issue below
32 Producer behavior
The efforts to build dynamic features into the economic behavior of producers are more recellt
and less widely adopted The first attempts are due to Bovnberg (1984 1985) and Summers
(1965) Dynamic behavior has been more fully incorporated In the recent models of Andersson
(1987) Auerbach-Kotlikoff (19B7) GouldermiddotSummers (1987) and Pereira (1986b 1987d)
Part of the reason why production side dinamlcs has been more slowty adopted is the weak
supply of accepted theories referring to the dynamic behavior of the firms In the models referred
I 17
to above gynamlc producion and investment b~havior are middotinduced by the existence of capital
adjustment costs and linked to Tobins q theory Adjuslment costs are designed to capture both the
incomplete mobility of capital aoross industries and installation costs ~e the costs of adjusting
capital towards its cpHmai level)
Adjllstment costs can be conceived as internal to the firm and measured in terms of foregone
outpul along the lines of Lucas (1967) Alternatively adjustment costs can be viewed as actual
external cests incurred together with Ihe purchase costs along the Unes of Gould (1969) With the
exception of Pereira (198Gb 1937d) all of the CGE authors follow the former approach
The firms in the economy maximize their market value as the present discountad value of the
future stream of dividends In Andersson (1981) and Pereira (198Gb 1987d) firms are seen as
maximizing the present discounted value of net cash flow Maximization is constrained by Ihe
adjustmenl cosl lechnology and ~n equation of motion describing Ihe evolution of the capital stock
It should be noted that undereogenous divldendrelention rules the two problems are equivalent II
should also be nOled Ihat a salisfectory economic rational for th bullbullistoo of dMdonds with the
present tax code is missing in the profession (Shoven (1986))
The models with static formulation of producers behavior are characterized by passive
investment behavior Investment merely accomodates to saving in the eccnomy With a dynamic
formulation induced b adjustment costs real investmen decisions are forward looking Investmenl
Is endogenously and optimally determined by the firms A fundamental difference between the
shonmiddotrun in which capital stock is given and longmiddotrun in which Ihe level of capital is allowed to be
optimally determined is emphasized
This extra richness of produc1ion dynamics is not without costs A careful look at the summary
tables will clearly show an inverse relation between the adoptlon of dynamic features in production
and the level of disagreggation of the production side of the economy In fact with production
bull
j
1 6
dynamics $e dimension of the problems is immensly increased Let IS be more specific
In a static framework with constant returns to scale production technology the output level ismiddot
indeterminate The optimal aliocation of inputs can be obtained by cost minimization with any
feasible output level generatillg zero profits Such zero profit conditions are used to solve for the
output prices in terms of the factor prices and hencemiddot to reduce the dlmensionality of the problem
from the -number of commodities and factors to the number of factors Thus even if the model deals
with 30 production sectors the computation of an economic equilibrium can take place using only
the dimensionality of the primary inputs in the economy
Now under certain regularity conditions on the production and adjustment costs technologies
leading to enough concavity of the optimality objective the intertemporal output path for the firm is
endogenously optimally and uniquely determined even with constant relums to scale technologies
(see Pereira (1985 1987a) Accordin~y with adjustment costs in the model optimal profits
will in general be non-zero This result has crucial implications for the computatiQn of equilibrium
in the model~ with production dynamics With adjustment costs no reduction of dimensionality is
possible The curse of dimensionality returns as a binding constraint
The introduction of adjustment costs adds another significant complication to the model With
capital being less that perfectly mobile in the economy different rates of return on capital will
exist in different sectors This is a difficult problem to tackle conceptually in the absence of
uncertainty Also in the real world producer maximization choices include also its choice of
financial ratios (debVequity) and payout rates (dividendretained earnings) These financial
subjects are difficult and much Clf this behavior is still taken as exogenous by the modellers We
will come back to these issues below
middot i
19
bull 33 government behavior
middot
Two issues dominate the modelling of government behavIor First govemment behavior has
Ibeen typically seen in Ihe CGE liIerature for lax poliCY evaluation as constrained by yearly balanced
I budsets (see for example 8allard-Fullrton-S~ovn-Whafley 098Sraquo The analysis of
govemment defidlS and public debl in a CelE context requires a dynamic selling Second the level of
government expenditures is either exegenouty given as in Auerbach-KoUikofi (1984 1987)
6ovenbrg (1984 1985) Feltenstein (1984 1986) and Jorgenson-Yun (1984) or
endogenously (but no optimally) determined by the balanced budget conditions as in
6allard-Fullerton-Shoven-Whalley (1985) Andersson (1987) Erllch-Ginshurgh-Heyden
~
(1987) Gouder (19aS) and Gouder-Summers (1987) In the second case the composition of
bull public expenditures is often optimally determined However the leve~ of government expendi1ures
can only 00 endogenously and optimally determined if the government is seen as an optimizing agent
and is allowed to run deficits
The first attempts to deal wiU the government defiCits in CGE tax models are due to
Auerbach-Kollilltoff (1984) and Feensain (1984) In Auerbach-Kotlikof (1984) government
expenditures afe exogenous They grow at the rate of growth of population However given 1he tax
structure and tax revenues yearly deficits and surpluses are atTowed subject to an in1ertempond
constraint that the present value of future tax revenues equals the present value of future
expenditures
In Fellenstein (1984 1986) government expenditures are also exogenously given He also
allows the government to run deficits to finance expenditures in excess to 1ax revenues However
surpluses are returned to CQnsumars in the form cf transfers Accordingly government ts not
subject to any constraint regarding the future repayment of public debt
In GouJder (19851 the 9Vernmenl maximizes a static social welfare ~nctio subject to a-
bull
20 I i
I
I I I
I bull
balanced budget conSirtlL This follows the optimal alocatkm of government expenditures alpng
the lines of Baliard-FHeiton-Shoven-Whalley (1985) Ttle model Is generallzea to allow for
exogenous changes in HIe time path of government expenditures Two financing alternatives are
consjdeft~d and contr8s1ed EldditionaJ tax revenues and bond issuance
Pereira (198Gb 1137d) auempts to address both the incorporation of deficits and the
determination of govement expenditures The path of government expenditures and the path of
delicilSlsurpluses (ano therefore the path lor debt) are endogenously and optimally determined The
government is seen as mltJximizing n intertemporal social welfare function given the 13x structure
Optimization is subject to a sequence of recursive equations of motion reflecting the evolution of the
ptlbllc debt allowing fDr government budget imbalances It should be stressathat this specification
01 the government cOMtcalnt is equivalent to the specificaUon in Auerbaeh-Kotlikoff (1983 1987)
when no liquidity constraints affect -gove~nment behavior
The extra richness in the treatment of government behavior allows tfte examination of
government debt poices in a truly dynamic setting Also financial crowding out effects induced by
government deficits can be analyzed (see Auerbach-Kottlikoff (1987) Feltenstein (1986) and
Perer (198Gb 1987e))
34 Foreign sector
Most of the open economy models follow the assumptions of balanced trade with import and
export net demands characterized by constant elasticities along the line of Ballardmiddot
-Fullerton-Shaven-Whalley (1965) Such is the case of Salfard (1963) Ballard-Goulder
(19S5) and Goulder-Sllmmers (1987)_ In Feltenstin (1985) the rest of the world is treated as
21
an addjlion~ consumer group
Bovcoberg (1986) develops a model In which two economies lIr considered each following
intertempoml perlect foresight paths These economies meet in the international forum Their
trade relaroMhips are characterized by yearly balanced trade accounts
None of the new generation of dynamic cae tax models has yet IncorPorated the lnternational
capital flows as done in aoulder-Shoven-Whalley (1983) for the earlier Ballarrshy
-Fullertonmiddot Shaven-Whalley (1985) model This is unfortunate as there is an important
Intertmporal aspect to international lending The first al1empts along Ihese lines are due to
Andersson (1987) and Erlich-Glnsburgh-Heyden (1987) Andersson (1987) in his model of the
Swedish economy adopts a closed economy appraoch in which rates of return in the domesJic
economy are largely determined by the international capital markets Fidinterest rate induce
Intlnatlonal capital flows which determine and finance the International lrade imbalance In
lurn In ErlichGinsburghHeydenmiddot (1987) foreign trade Is generaled according to an
intertemporal Irade welfare lunction with constant import and export elasticities In the shortmiddotrun
they allow inlernational trade inbalances which generate capilal ftows to the domestic households In
the long run however trade balance is assumed
35 The need for financial markets
A dynamic economic Wucture not only provides the ideal environment 10 model many features of
economic behavior it also permits one to Incorporate the financial side of the economy If the
government is allowed to run deficits the question 01 deficll financing automatically follows If
investment is optimally dewmined and returns to capital are different aClOSS sectors problems of
investment financing arise If there are seeral final1cfal assets In the economymiddot government
4-
I i
ibull
I I I I
I
j
22
bonds privpte bonds and equity (or simply physical capital installed inmiddot different sectors) the
problem of allocation of saving among assets with potentially different returns arises
The papers surveyed var greatly with respect to l~e extent of their attention to the financial
side of the economy At Clne extreme are the models in ~anardmiddotFuliertonmiddotShoven-Whalley (1985)
Andersson (1987) Ballltld (1983) Ballard-Goulder (1985) Bovenberg (1985 1986) and
Erlich-Gjnsburgh-Heyden C19B which are devoted exclusively to the real side of the economy
In turn Auerbach-KotliKoff (1984 1987) Feltenstein (1984 1986) and Goulde (1985)
allow for government debt In these models saving finances changes in government debt and physical
capital Private and publlc 2ssets are perceived by the households as perfect substitutes The
allocation of saving merely adjusts to the relative demands for funds
Feltenstein (1984 1986) is the only model surveyed which introduces money Government deficits are financed by issuing money and bonds according to an exogenously given rule Money is
demanded by consumers for transaction motives and an exogenously given fraction as a store of value
On the other hand government bonds and physical capital are the vehicles for the intertemporal
transfer of wealth
Summers (1985) and GOlldermiddotSummers (1987) introduce a whole menu of financial assets shy
firm specific equity capital Different assets earn different rates of retum However such rates
are equal up to constant and exogenous sector speCific risk premia Therefore the introduction of
constant exogenous risk premia as helpful as it may be in the context of calibration does not solve
the main issue of the non-optimality of the allocation of saving Also talking about risk premia in a
deterministic context is somewhat unsatisfactory
Pereira (1986b 1987d) also introduces a whole menu of assets private and public bonds and
firm specific equity However all the assets are expected to yield the same rate of return and
therefore perc~jved as perfect substitutes The different asset types allow consideration of
23
exogenOU$~8tUequjty triO dividendlrc~enlion rufes and therefore seve-ral sources of investment
financing bond equity and retained earnings
The nonmiddotcptimalHy of the allocation of saving and the absence or exogenelty of corporate
financial rules is _a reftectlon of the limitations of the determInistic approach Either in a context of
perfect anticipation oi the future prices Of in general within the t$atm of point price
expectations all the papers follow a deterministic approach Under such circumstances consumers
either expect diHerent rales of return (inclusive of risk premium) across assets in which case
they wilt buy only one asset (that with highest rate) Of they expect equal rat~s of relurn in which
case Ihey are indifferent about the asset composition of Iheir portfoliO There is no way of
Iradingmiddotoff rales of return and risks to obtain an optimal interiof solution to the problem of the
allocation of saving
The most advanced contribution in the modelling of saving allocation in a CGE setting is due to
Slemrod (1980 1983) in the context of a static one-period model In his model consumers act
according to a two stage separable decision process They Hrsl decide on how much to save Then
they decide on the allocation of saving according to an Indirect uttlity function dependent on the rates
of ratum and variances offered by the different assets in the eoonomy The sourCe of riskiness in the
ampConory comes from an uncertain marginal product of capital On the other hand aside from
portfolio decisions the res of the economy is insulated from uncertainty
The most complete contribution in terms of the treatment of the corporate financial rules is
FulienonmiddotGordon (1983) In a variant 01 Ihe BailardmiddotFulienon-ShovenmiddotWhelley (1985) model
Fullerton and Gordon have capital intensity and optimal financial decisions Jointly deter~ined
through a two siage process The cosl of financing capital is minimized by trading off the tax
advantages of debt against the epected real bankrupcy coSls inherenl with high debVequity ratios
Given the optimal debtequity ratio the level of inveSlment Is chosen such that at Ihe margin the
24 l
bull
rerurn pn ~uity equas the return on bonds plus an exogenous risk premium
36 Market assumptions equilibrium and expectations
Virtually all of the papers arc characterized by Walrasian market clearing assumptions All
markets are perlect1y competi1ive Atomistic competition among agents 1s assumed even thoughmiddot only
a flnile number of agents is considered Virtually no mark~t disequilibria or price stickiness afe
considered The only exception is ErlichmiddotGinsburghmiddotHeyden (1987) bull In thei model of the Belgium
economy I the _wage rate is fixed in the short-run Therefore in the shOrlwrun desequilibrium in
the labor market will generate endogenous unemployment However in lh$ long-run all prices
Including the wage rata are Ilexible and accordingly ali markets ciear
Given the dynamic nature of behavlor in the economy markelmiddotcfearing prices in each period
depend on expectations of fture prices and on tax variables In the economy There are essentially
two ways at interpreting the economic equilibrium in such a dynamic conte~bull If future prices are
perfectly anticipated (io expectations are self-fulfilling) a perfect foresight equilibrium
prevaHs Then future actions are merely the implementation of current decisions for future
periods However jf ptica expecla110ns are not perfect (Le agents make mistakes w1th re$pect 10
future prices) then a temporary or shortrun equilibrium prevails Markets ciear and clearing
prices depend on fu1ure price expectations Current plans about the future are typically not
precisely implemented They will be revised as more or better information becomes available to the
economic agents See Grandmant (1982) for a comprehensive survey 01 the temporary equilibrium
literature
Wilh the exception of Gould (1985) and Pereira (19S6b 19S7d) all the models surveyed
adopt th~ concept of a perfect foresight equilibrium In turn BaliardmiddotGouldr (1965) considers a
J
2S
flexible am9unt of foresight In terms of the number of years over which price movements are ~
foreseen Pereiras model (1986b 1987b)) is flexible in a somewhat different way in that it can
indude any range of foresight from myopia to perfect foresight The choice between the perfct foresight and lemporary equilibrium is ultimately to be made on
philosophic grounds It can be argued that less than perfect expectations Imply that agents are
irralional in some way (see Auerbach-Kctiikoff (1987 p 10) However the reverse argument can
be made One can question wiether agents are reaHy rationar and perfectly knowledgable about
future prices
Recent evidene of Balird (1987) Ballard-Goulder (19851 Goulder (19851 and Pereira
(1986b 1987d) suggests that the choice in modemng expectations is an important one They show
that the degree of foresight inl0 thc future (ranging from perfect foresight to-myopic expectations)
may have dramatic impacts on the pollcy conclusions of Ihe model Accordingly the best research
strategy may b to design models which are flexible enough 10 allow for different rules regarding the
formation of expectations With the exception of the articles just mentioned sensitivity analysis
bullhave previously not been performed along this dimension
In terms of implementation the two concepts of equilibrium - perfect foresight and temporary
equilibrium have different implications The dimensionality of the equ1ibrium soll1ion algorIthm
is involved Suppose we have a model with ten markets to be run for a period of 50 years Aside
from normalization a perfect foresight model implies computing prices In 500 dimensions while a
-t~mporary equilibrium model requires sorving SO equilibria each in ten dimensions Given that
computational speed often varies with the cube of the number of dimensions the lemporary
equilibrium formulation is potentially strikingly more feasible However Ballard (1987) and
Goulder-Ballard (1985) have devoloped techniques to greatly speed the computation of a perfect
foresight equilibrium
26
j
I
4 Jrnplemenlalion and Issues on policy evaluallon
41 Calibration and equilibrium comparlslon
CGE models at typically parameterized by the use of a calibration procedure Some parameters
are exogenously given However some crucial parameters are determined in such a way that the
model replicates the data for a given base year See MansurmiddotWhalleY 11984) for an eXlensive
discussion of 1~i$ issue
CalibratIon in a dynamic context is generaliy interpreted as requiring two properties First
replication of base year data is required Second the model is parameterized to simulate an
intertemporal balanced growth path when Ihe base policy is maintaineq This Is the approach
followed by Ballard (1983) BallardmiddotGoulder (1985) Goulder (1985) and SummersmiddotGoulde
(1986) It follows the practice of BallardmiddotFuliertonmiddotShovenmiddotWhalley (1985) with their
sequential cquilibrium dynamics
Other authors Auerbach-Kotlikof (1983 1984 1987) Boyenberg (1984 1985 1986)
and Pereira (198Gb 1987d) follow whal we cali a qualitalive calibration The slructural
palametars are exogenously chosen so that lh economy follows a reasonable path into Ihe fulure
ThaI has 10 do wllh the fact thaI given the recursive nature 01 the dynamic economy and aside tom
such structural parameters only iniUa stock values are needed to run these models Given initial
conditions on the stocks of say private wealth capltal and government debt agents will optimize
and Ihereby generate a lirst round of net demands and equilibrium conditions In lurn the
equilibrium prices will determine the evolution of the stock variables into the next perIod
There are several potential problems with calibration in the conlext of dynamic models The
assumption of a steadymiddotstate growth ~ath in the base case can be questioned First while
27 I Ibullbull
steadymiddotstaU is a possibity it certainly is not the only meaningfull solution to dynamic models
Even in the case of a perfect foresight equilibrium the model implies an equilibrium path which i may or may not involve colanced growth The model not the modeller should dictate the nature 01 I themiddotbase case path Second in the context of a temporarY equilibrium path a steady state solutlcn is Inol a likely model QutcOn1tl In fact unlessmiddot expectations are stattc short run behavior consistent
with a steady-state evolution will in general not be generated On the other hand if static I expectations are self~fufjili[1g we have in fact a perfect foresight model Third even the base year
replication requirement may cause problems In Ihe cOntexl of temporary equilibrium Any
calibration parameter would be conditional on expectation rules which is probably an undesirable
feature
The nalur of the two-requirement calibration strategy is very muchin the spirit of the bull
traditional design of the comparision of alternative equilibria comparis_ion between a steady-state
base case on one hand and alternative paths including a transition period and a finat sleadyyenstate on
the other hand Pereira (198Gb 1987d) compare different (not necessarily steadystate)
equilibrium paths The arguments against such a procedure are based on the idea that the impact of
policy changes can be observed most easily since all departures from the steadymiddotstate can be
attributed to the altemalive polioy On the other hand the base case so dEfined as a steadymiddotstate is
consist~nl with previous work in a less dynamtc setting and Iherefore allows a common standard
for comparing model results
42 Computation algorithms
We are still al a stage in which basically each author uses a different computational technique
Th development of dynamic models - in parlicular wilh adjustment costs andior perfect foresight
bull
28
has corresponded with the decline in the use of fixed point algorithms In -fact given the relative
arge dimensions inevitably involved such algorithms lend to be very inefficient at the best and
chen prohibitively stow See Stone (1985) and Preckel (1965) for a comparative assessment of
different computation techniques Among the models surveyed only Ball1rdFulle(ton~
-ShavenmiddotWhalley (1985) and Feltensteln 1985 use Merrills variant of the fixed point
algorithm 1echnique
AuerbachmiddotKoHikoff (1963 1984 1967) follow a three stage procedure They first compute a
base case steady-state then a revised case steady state and finally transition path for the economy
berveen these two steady-states tn all stages a Gauss-Seidel iterative procedure is used
Ballard (1982) Ballard-Goulder (1965) Goulder (HiSS) and Summers-Goulder (1986)
~
use a method developed by Ballard and Goutder which is similar in many awects to the FairmiddotTaylQr bull
(1983) algorl1hm Short-run equilibria are calculated (using Merrills algorithm) parametric on
nrice expeC~lions The model is then iterated to generate self-fulfilling intertemporaf expectations
and the corresponding perfect foresight equilibrium In a relatively similar approach Andersson
1987 uses a simulation program SIMNON developed In the University of Lund This program
can handle two-paint boundary problems In a fashion consistent with the multiple shooting
algcrlthm (see Lipton-Poterba-Sachs-Eummers (1982))
Boyenbergs computational approach (1985 1985) differs from the other models surveyed in
that he relies heavily on analytical techniques Computations are done by using a dynamic version of
Johansons linearization method Being essentially determined by the continuous time nature of the
I I
f
model thiS linearization model has the disadvantage of confining the analysls to infinltesinal changes
around the base case equilibrium (see Bovenberg (1985) p 53) Pereira (1986b 1987d) uses an optimization algorithm NPSOL - developped by
rGillMurray-Saunders-Wrlght (1986)) This algorithm is used 10 compute the sequence of
I
29
Shortmiddotrun temporary equilibrium which makes up the intertemporal equilibrium path The
equilibrium conditions are seen as nonlinear equality constraints in the minimization of an artlcial
objective function The prices are normalized to the unit simplex by an additional linear equality I I 1 bull constraint
II ErlichmiddotGinsburgh-Heyden (1987) follow a unique approach in that they use a variant of the
optimization technique introduced by Negishl (1960) The economic equilibrium can be generated i
as a solution of a mathemallcal program Ihe objective function of which is a weighted sum of the
utility functions of the various agents while tho constraints set consists of the market dearing 1
conditions Ginsburgh-Heyden (1985) have extended Negishis resulllO tho case of downward price
lrigidities I
Finally the paper by Jorgonson-Yun (1984) is also unique in tIlat it is the only I
econometrically estimated model Different blocks for the consumption and production Side of the
-
model are separately estimated to provide the necessary structural parameters i
I The diversity of computation techniques is yet another indicator of the exploratory nature of the
body of literature surveyed in this article
43 Equal yield comparisions
The link between a base case and counteriactual simulations Is usually provided by the concept of
equal yield Shaven-Whaley (1977) discuss the meaning 01 equal yield in a general equilibrium
context when government is not allowed to run defICits Equal yield is interpreted to moan constant
public utility Government base case utility is maintained in the counferfactual experiments With
balanced budgets the concepl of equal yield is unambiguous The new equUibrium prices and the
balanced budge condition will del ermine the minimal expendilure and taxes needed to maintain base
30
case publi~ utility Accordingly in general equal yield is inconsistent with equal nominal tax
revenue Some change in tax revenue Is necessary Different tax replacement schemes ate
considered to assure that enough tax revenue is collected This is the approach essentially followed
by most 01 the papers surveyed Only Feltenstein (1985) Goulder (198) and Jorgenson-Yun
(1984) chose not to follow an equal yield strategy
The question is of how to Interprete lila concept of equal yield when the government is alowed to
fUn deflcLts The optimal leve of expenditure for base ease public trliUty can now be taAinanced
bond financed or financed by a mix of bonds and taxation We have several versions of equal yield In
particular equal yield may now be consistent with equal tax revenUE Furthermore some meaSure
ltgtf financial crowding out effects of government deficns can be inferred from the comparislon of the
several equal yield alternatives These issues are extensively discussed ~Pereira (1987b)
44 Price expectations and the dynamic generalization of compensation
Indicators
The sum of equivalent and compensation variations over households is the most widely used
aggregated measure of efficiency gains or losses In the context of $teady~state eomparisions andor
when complete future markets exist andlor perfect foresighl is assumed there are no difficulties
associated with the use of the standard Hicksian indicators In fact correct future prices are known
in these cases
If expectations are not selfmiddot fulfilling andlor future markels are not available a dynamic
generalization is necessary 8allardmiddotGoulder (1985) provide some steps in that direction by
defining an indicator that accomodates periods far into the future when households do not have
perfect foresight but a steady~state prevails On he other hand this issue is more fundamental in
i
[
I
i I I I I [
I I
bullI
31
the contex~ of a Gner1 lemporary equilibrium ftamawor~ In such Circumstances Pereira
(1986a) dev~lops a cnamlc generalization which is obtained as the present discounted value of a i i
sequence of short-fUn optimal expenditure functions consistent with a base case expected future I stream of IlWities I
32
5 Empirical evidenee from selected policy Issues
Dynamic tax models have generated several important results which escaped Sialic modellers
The following is a seleCpoundd set of issuos whIch stress the marginal benefits of dynam~ modelHng of
economic behavior in innovltlttive areas of analysis The discussion of a consumption tax emphasizes
the benefits of dynamic tusehold behavior and intergenerational aspects The study of the
flliminatlon or the re~intToduction of the investsnemt tax credits is made meaningful by the dynamic
modelling of production betlavor Modelling governmenl in a dynamic conlexts allows the modeller
to study the irrrpact of government deflcits Finally a dynamic framework lets us appreCiate the I
relevance of consumer expectations in terms of the evaluation of policy alternatives
i
i
51 Consumption Tax
t
By Its very nature a consumption tax can not be adequately investigated with a static model bull
Fullerion-Shoven-Whaicy (1983) use Ihe model of the US economy s described in
Ballard-Fullerton-Shoven-Whalley (1935) to evaluate the movement from the currenl US tax Isys1em to a progressive consumption tax Since their modeJ Incorporates a Jaborleisure choice
where leisure is an untaxed commodlty their results reflect the fact that both the consumption tax
and Ihe present system are dslortionary
Concentrating on the rntertemporaf distortions they show that sheltering more saving from lhe
current US income tax could improve econornic efficiency even if marginal tax rate increases are
necessary in order to maintain government fevenue~ At first you have a revenue shortfall
However the economy moves to a higher sustatned growth path and uitimatefy lower tax rates
cangenerate the same revenue path Also wages increase in the long run with this policy The
bull
bullbull ~
33 shyt
i
I
I
present value of welfare gi[lS for a potiey of complete savin9 deduction wilh marginal tate
adjustments (consumption lax) is simulated to be around $500 billion to $600 billion 01 1973
dollars
They also investiga1e th 1ngth of ttme it takes th~ economy to adjust to these policy changes
Roughly they estimate the long run to b thirty years although this figure Is very sensitive to the
specification of the savings elasticity - a crucial parameler in this model
52 Investment T~x Credit
Goulder-$ummers (1 S87) address the impact of eliminating the lnvestmenl Tax Cradit (ITC) on
intersectoral capital formation and on economic growth Their model isYa1ticularly adequate to
addies$ this issu~ in that is postulates a forward looking investment m9we with adjustment costs
They show that the eliminating the ITC causes a reduction in the rale of investment Inveslment is
estimated to fall by about 7 in the short-run and by about 12 in Ihe new long run steady state
In lurn a previous policy anncuncement lowers Ihe overall al1ractiveness of Investment and leads 10
a downward shift of the Investrrent profile On the other hand the combined effect of a revenue
neutral simultaneous efimiraron of the ITC and reduction of the corporate tax rates is a long run
reduction ef the capital Sieck by 35 This pattern suggests that a revenue neutral increament in
both Ihe corporale tax rates and the inveSlment lax credits would be preferable 10 the formula
adopted in Ihe US Tax Reform Act of 1986
Pereira (1987d) addresses the impact of the ITe from Ihe oland point of Ihe curTenllax syslem
under the Tax Reform Act of 1986 Given Ihe concern over the potential depressiVe effects upon
savings and investment of the new tax system in conjunction with deficit reduC1ion mechanisms of
the GrammmiddotRudman type a pertinent question is should Ih lTC be e-introduced In the context
bull
34
-of his mod~1 of the US economy Pereira focuses on the trade-off between the distortion in the
intersectoral allocation of capital induced by the lTC the lower tax revenues it generates and
potential financial crowding-out effects on one hand and the positive effects of lowering the relative
price of new capital goods on the other hand Preliminary results confirm the qualitative results in
Goulder-Summers (1987) suggestting a welfare gain of about 2 of the present value of GNP in
the best scenario of absence of replacement taxes Furthermore preliminary results show an
important time pattern to welfare gains with the average benefits increasing the further you look
into the future This pattern is due to the presence of constraints to intersectoral mobility of capital
and inertia in the adjustment towards the optimal capital levels as reflected by the presence of
adjustment costs
I
I
bullAuerbach-Katlikaff (1gB) in the context of their intergenerational model of the US economy
consider the impact on savings capital formation and interest rates of deficit policies and balanced
budget increases in government consumption They conclude that deficit finance and government
consumption can significa-ntly crowd out capital formation and lower the welfare of future
generations However crowding out from deficit finance is a very slow process because it results
from increased government spending over potentially long horizons Also deficit policies may
substantially influence the longmiddotrun interest rates while leaving the short-run rates essentially
unaffected Finally and in opposition to the central role adjustment costs seem to play in both
53 Financial crowding outeff~cts of government deficits
Goulder-Summers and Pereiras models Auerbach-Kotlikoff suggest that the time path of interest
rates induced by a policy of deficit financing seems to be insensitive to the presence of adjustment
costs
bull
35
54 The role 01 consumer expectatlons
The expectations analysis has uncovered one of the benefits of dynamic modelling
Ballard-Goulder (1985) find that the appeal of adopting a consumption tax depends on the level of
foresight possessed by the consumers Furthermore additional foresight may be welfare worsening
This is a second best type of result This tends 10 occur under policies that lead to capital deepening
and declining rate of return to capital over time To the extent that consumers have more foresight
they will be beller equippod to anticipate the fall in the rental price of capital Consequently if a
saving incentive is enacted people save more with myopia than with perfect foresight Given the
exi~tence of taxes on capita and the discrepancy between the private an~--sccial returns to capital
the greater savings with myopia is beller socially Ballard-Goulder find that the welfare gains
from a consumption tax are reduced by about 10 when we move from myopia to farleet foresight
Therefore the attractivenness of adopting a consumption tax seems fairly [obust across these
foresight specifications
I
bull
36
6 Tbe power and weaknesses 01 dynamic modelling the example 01 corporale
tax Integrallon
The Issue of corporate tax integration allows us to sh~w the stre-ngths and weaknesses or the
dynamic approach Empirical evidence using CGE techniques indicates that depending on the precise
scheme of integration and the 1ax replacement methods integratiOn may have substantial effects In
Ihe work of FuliertonmiddotKingmiddotShoven-Whalley (1980 1981) total integration was found to yield an bull
annual static eHielency gain of $4 to $8 billion in 1973 dollars Simulated dynamic gains may be as
large as $695 billion or abcu 14 of the present value of tulUre consumption and leisure in the
US economy These gains result primarily from interindustry reallocalions of Investment and an
improved inter~emporal allocation of consumption
Mora recent work emphasizes hOw consumers asset portfolio decisions and firms finanCial
decisions affect the efficiency gains from integration Slemrod (1980) focusing on consumers
asset ponfoio deCisions finds static efficiency gains which are about twice -as Iarge as those
repOrted iiy Fulierton-King-ShovenWhalley (1961) FuUrlon-Gurdon (1983) focus on the
firms financial decisions They roporl efficiency gains of 6 of GNP from the elimination of the
tax distor1ions favoring debt However when they eliminate the corporate tax and repiace it with
increaSed persona income taxes addlHonal dls10t1ions are created in the optimal labot~leisure
decisions These distortions tend to dominate the analysis slJ(h that the overall effects of complete
integration are very modest Galpr-LuckmiddotTodr (1986) address simultaneously consumers
asset pOrlfollo decisions and the firms financial decisions They also find very modest efficiency
gains from integralion
The above results are important but may be severely biased There are severa aspects of
economic behavior and modelling crucial for the study of Income tax Integration whiCh have no been
i 37
captured innny of the above papers One aspect is the absence of government deficitsmiddot a balanced
budget is assumed It is true Ihal resource crowding OUI is caplured In these mOdels but financial
crowding out induced by government spending is nOI Second investment is not derived from
optimtztlcn behavior Investment behavior passively accomodates endogenous saving decisions As
a consequence the differential impacts of policies in the incentives to save and to invest are not
captured Aso full capital mobility across secors is assumed with instantaneous capital
adjustmenlS towards optimal levels This assumption rules qut different costs of capital across
seclors and therefore differentiated reactions to tax policies changes Finallyhe modelling of both
gov9rnmerlt deficits and of endogenous real and financial investment decisions necessitates the
~nsideraiion of a dynamilt framework and the introduction of financiaJ assets govemment bonds and
iprivate financial assets A dynamic framework also highlights the efficiency effects of Integration on r
ithe ptimal intertemporal decisions The above models are either sIalic (Slemrods and
GalpermiddotltJckemiddotTode(sj or a dynamic sequ~nce of otherwise Slatic model bullbull I Pereir a (1986b) develops a dynamic applied general equilibrium model bull to study the tbull
efficiency effects of integration on the growth and allocation of investmenl across sectors Special
Iattention is paid to the structural affeelS of resource and financial crowding oul induced by
gove mect spending and deficits and to the real and financial Investment reactions across industries
to both tax integration and flnancial crowding out His mode departs from previous work on income
lax Integraticn and for that matler from most 01 the CGIE literalure for tax policy evaluation in
several direcllons It encompasses an endogenous sequential equilibrium struelure founded on
dynamic behavior with flexible expectations Government deficits are optlmaUy determined
Investment decisions are tward looking and the resull of optimizing behavior Several financial
assetsmiddot public and privatemiddot are considered
Simulation resulls Indicale that the welfare gains from Integration under large deficits are at
32
best lJlodest when measured in terms of the GNP The elimination of the corporate tax and its I replacemenl by increased income lax rates yields long run benefits which are never larger than
2 of the presenl yalue ct telr consumption and leisure onder the previous tax regime and 1
under the current tax regime However the short run w~lare effects of full integration tend 10 be
very low and in in some scenarios even negative This is a flew intertemporar pattern of efficiency Ieffects which reflects an adjJsiment lag in the interindustry investment decisions due to the
existence of costs of adjustment On the of her hand partial integration achieved by excluding I dividends from the corporale lax base syslematically yields negalive efficlencyeffeclS In Ihe long Inm these can be as high as 3 of present value of the future value of consumption and leisure This
is a new second best effect suggesllng Ihat less than complete integration may have perverse I efficiency effects
The dynamic slructure of PerclrBs model snowing for an improved freatment of real investment
declsions and government deficts provides a potential setting for a more accurate measure of the
costs and benefits of inlegration The simulation resullS suggest lower benefits and an intertempora Ibull
pattern of growing benefits
Simulation results also clearly suggest robust negative effects from partial integration How
reliable is this result If dJvidends were deductible from the corporate lax base (as are interest
payments) the preferenlial Ireatmant of debl over equily would be ellminaled Corporallons should
be expecled 10 react by decreasing the optimal deWeqully rallo However corporale financial
gecisions are exogenously set In Pereiras model as in aft of the other models surveyed that go as far
as dealing with the issue
Also withdrawing dividends from Ihe corporate tax base is a way of eliminating Ihe double
taxation of dividends at belh Ihe corporate and personnal Income levels How will the optimal
aliocellon of household saving accomodate that change in Ihe relalive return to corporale equity
39 I
t However hOusehold portfolio decisions are exogenouly set in Pereiras model as in all of the other
models surveyeltj go as far as dealing with the issue
IIt is sate to say that the evaluation 01 the benefits 01 partial Integration as defined above are
sevrey underestima1eO Meanwhile the distortionSoenerated by the increase in the marginal
personal tax rates designed to ralse the revenue foregone by the dividend exclusion are fully
aCCltlunted for A good measure of the costs of integration together With a poor evaluation of the
benefils may very well be the reason why partial integration is Simulated 10 yield neg alive
eHikiency gains
On a different vein as most of Ihe model surveyed here Pereiras is a closed economy model It
is legitimate to wonder how the resullS would change If international capital flows were anowed
IThis almost certainly would affect financial crowding out since a good d~al of the capital inflows
could be used to finance public debt
For these various reasons the results should be treated as preliminary but they strongly I suggosl that the dynamic structure provides valuable new insights into the corporate tax integration SSU2
40
7 Concluding remarks I I I i
What has been acccmplished Ihis far The success of Ihe CGE research in laelding the challenge
of dynamic modelling can not be denied
I iGreat progress in the modelling of household behavior has been achieved Promising
developm~nts in the modeUing of production and government behavior have also been accomplished
Interesting innovations in the concept of equilibrium and computation techniques were discussed
The policy analysis was greatly enriched by considering transitional effects together with longer
run steadymiddotstate equilibrium paths generalized equal yield strategies accomodating different
financial crowding out impacts and generalized dynamic policy evaluation indicators
An important set of issues have been addresed by the models surveyed Traditional issues
focusing on the effects of capital taxation and consumption taxes have been pursued In terms of
capital taxation for example the intertemporal nature of the issue was enhanced by allowing an
optimal evolution of capital stock in the economy and forward looking optimal investment decisions
In Andersson (1987) GouldermiddotSummers (1987) and Pereira (1986b 1987d) this is coupled
with a improved treatment of several tax provisions like investment tax credits and depreciation
allowances In terms of the consumption taxes developments in the specification of Intertemporal
household behavior the introduction of overlapping generations and the modelling of
intergenerational links via bequest motives as well as a closer attention to demographic evolution
provided a much improved economic setting for the understanding of the several aspects of the
problem
Furthermore dynamic modelling has permitted the addressing of a set of new issues Policy
Issues ranging from dynamic tax policy analysis to the evaluation of exogenous changes in
government expenditures to the impact of different methods of financing government expenditure to
4 1
the importance of tinanciai croHrii1g out to 1he relevance of consumorS expectations in tne policy
results have been addressed
Despile all of the progress and in part due to such progress several avenues are wide open for
much neded additional research The following seem to be the most Interesting and promising
areasmiddot First the specincation of liquidity constraints to household behavior may help to obtain a
better understanding of policy chenges that affect directly the interest rates in the economy Second
major attention needs to be paid to the incorporafion of a roraign sector (with open capital markets)
Third some efforts are needed in terms of finding adequate computation techniques now that
dynamic modelling eally makes the curse of dimensionality worse than eVOrbull
The modelling of financial markets is the single most unsatisfactory speet of the dynamic
models surveyed here~ The problems associated with provlding a mo~ adequate treatment of
financial markets and in general endogenous and optimal financial deci$ions bull both at the corporate
and at the personal levelsmiddot have 10 be addressed That necessarily requires Introducing uncertainty
into the models To b adequate this must go well beyond the mere assumption of point expeetalions
with the whole array of modelling implementalion and evaluation problems thai generates
bull I i
42
REFERENCES
Abel A and O Blanchard 1983 An Intertemporal Model of Saving and Interest Econometrica
51 pp 675-92
~ Andersson K 1987 Tar-ation of Capital in Sweden - A General Equilibrium Model Ph D
Dissenation UnIversity of Lund Sweder
Auerbach A L KotlikoH 1983 National Savings Economic Welfare and the Structure of
Taxation in M Feldstein ed Behavioral Simulation Methods in Tax Policy- Analysis Chicago
University of Chicago Press
Auerbach A l Kotlikoff 1984 Social Security and the Economics of Demographic-
Transition in H Aaron and G Burtless eds Retirement and Economic Behavior Washington DC
Brookings Institution
Auerbach A L Kotlikoff 1987 Dynamic Fiscal Policy Cambridge Cambridge University
Press
Auerbach A L Kotlikof and J Skinner 1983 The Efficiency Gains from Dynamic Tax
Reform Internatiooaj Economic Reyiew 24 pp 81middot100
Ballard C 1983 Evaluation of the Consumption Tax with Dynamic General Equilibrium
Models Ph D Dissertation Stanford University
Ballard C 1987 Tax Policy and Consumer Foresight A General Equilibrium Simulation
Study Economic Inquiry 25 pp 267-284
Ballard C and L Goulder 1985 Consumption Taxes ForeSight and Welfare a Computatonal
General Equilibrium Analysis in J Piggott and J Whalley eds New Deyelooments in Apolied
General Equilibrium Analysis Cambridge Cambridge University Press
Ballard C D Fullerton J Shoven and J Whalley 1985 A General Equilibrium Model for
t 43
Tax Eoliey Eyalua~iQO NBER University or Chicago Press
Borges A (1987) bull A Survey of Compula1ional General Equilibnum Models CECD Economic
Slud~ 8
Bovenbig L 19B3 imperfect Capital Mobility in a Perfect Foresight Herberger Model
Mimea UC Br~eley
Bovenberg L 1984 Capitel Accumulation and Capital Immobility Qmiddottheory In a Dynamic General Equilibrium Framework PhD dissertation University of California at Berkeley t
Bovenberg L 19850 Dynamic General Equilibrium Tax Models with Adjustment Costs in A iManne ed bull 1985 EtQnomrc ECUjilbrium Model Formulation and Solution Mathematical
lProgramming Siudy 23 Arnsterdam NorthmiddotHoliand
Bovenberg L t 1986 Capital Income Taxation in Growing Open Eeondmies Journa of Pubtic
Eamp(omlcs 3t pp 347middot376 IBovenberg L ald W Keller 1981 Dynamics and Nonlinearlties in Applied General
Equilibrium Models Internal Report Mlmiddot81middot208 Department for Statistical Methods CBS bull
Vooiburg The Netherlands
Charnley C 1981 The Welfare Cost of C~pltal income Taxation In a Growing Economy
IJournal of Political ECQnQm~ 89 3 pp 468middot496 I
Charnley C 19810 Efficient Tax Reform in a Dynamic Model of General Equilibrium I Mimeo Yale University I Charnley C 1982 On the Optimal Taxation In Stylized Models 01 Inlartemporal General
Equilibrium Mimeo Yale University I Decaluw B and A Manens 1985 Pays en Oeveloppement at Modeles Calculables OEquilibre
General - Une Revue de al litterature empirique Monographie Sene 9 Volume 3 Centre de
Recherche et Developpemont Economlque Unrslt de Montreal
44
bull
I I
I
I
1
Erlich $ V Ginsburgh and L van der Heyden 1987 Wher~ do Real Wage Policies Lead
Belgium bull A General Equilibrium Analysis European Economic Reyiew (forthcoming)
air R and J Taylor 1983 Solution and Maximum Likelihood Estimation of Dynamic
Nonlinear Ratioant Expeclations Models Ecooometrica 51 pp 1169middot1186
Feltenstein A 1984 Money and Bonds In a Disaggregated Ec~nonYin Searl 1-1 J Shoven
Eds Aoolied ~eoeral EqJiliDCiulD tioalYsis Cambridge University Press Cambridge
Feltenstein A 19B6 A Dynamic General Equllibrlu Analysis of Financial Crowding Out
Theory with an Application to Australia Journal of Public Economics 31 No1
Fullerlon D R Gorden 1983 A Reexamination of Tax Distorllons in General Equilibrium
Models in M Feldstein Ed Behayora Simulation Methods in Tax pOlicy Analysis Chicago
University of Chipgo Press
Fvllerlon D J Shaven and J Whalley 1983 Replacing the US Income Tax with a
Progressive Consumption Tax JoulOal of Public ECQnomics 20 pp3middot23
Fullerton D Y Henderson J Shaven 1984 A COmparision of Methodologies in Empirical
General Equilibrium Models of Taxation in Searl H J Shoven Eds ~pQlied Geoaral Jguilibrium
lIoalyss Cambridge University Press Cambridge
Fullerton 0 A King J Shaven and J Whaney 1981 Corporate Tax Integration in the United
Statesmiddot a General Equiiibdufi1 Approach American Economic Review 71 pp~ 677~691
Galper H R Lucke E Tader 1985 Taxation portlor Choice and the Allocation of Capital A
General Equilibrium Approach Mimeo
Gill p W Murray M Saunders M Wright 1986 Users Guide for NPSOL (Version 40) A
Fortran Package for Nonlinear Programming T echnicsl Report SOL 86middot2 Deparlment of
Operations Research Stanford University
Glnsbur~h V and L van der Heyden 1985 General Equilibrium with Wage Rigidities An
45
AppJication to Belgium in A Manne ed t Economic EQumprit1m~ Made Formulation and $oJioo
Mathematical Programming Study 23 Amsterdam NorthmiddotHolland ~
Goulet J 1968 1 Adjustrent CO$1S in the Theory o(the Firm Reyiew of EconomiC Studies 35 bull
pp47middot55
Goulder L 1985 Tax-Financed versus BondmiddotFinancedGhanges in Governent Spending
bull Mimeo Harva(d UniversilYmiddot
Goulder l J Shaven and J Whalley 1983 Domestic Tax Policy and the Foreign Sector The
importance of altemative foreign policy formulations to results from a general equilibrium tax
analysis model In ~n methods in Ibe antico of income from caoia1 ed Martin
Feldstein Chicago University of Chicago riSS
Goulder l L Summers 1987 Tax Policy Asset Prices and 3rowtly ~
A General Equilibrium Analysis NBER Working Pper No 212B
Grandmont J 1982 Temporary General Equilibrium Theory Ch 20 in Handbook 01
Mathemalical Eccnornics Vol II NorthmiddotHoliand
Harberger A 1959 The Corporate Income Tax An empirical Appraisal In lax Revision
Comoendjum Voll House Cornmitee on Ways and Means Washington DC Govemment Printing
Office
Harberger A 1962 The Incidence of the Corporate Income Tax JOULn1 of Political
Economy 70 pp 215middot240
Harberger A 1964 Taxation Resource Aliocalrlon and Welfare fn lb Role of Dir bullbulll and
Indjresl Taxes io tbe Fede[al Reserve s~I~Ill Princeton Princeton Unlveresity Press
James J 1985 New Developments in the Applications of General Equilibrium models to
Economic History in Piggott aand Whalley Eds
Jorgenson D and K Yun 1984 Tax Policy and Capital Allocation Harvard Inslitute of
bull
I
I f
Economic Research WP 1107
Judd K 1965 ShonmiddotRun Analysis of Fiscal Policy in a Simple Perfect Foresight Model
Journal of poJiHcal 5CCJiW1l pp 298middot319 r
Upton D J Potbc J Sachsa nd L Summers 1982 Multiple Shooting in Rational
Expectations Models fuoometrlc 50 1329middot1333
Manne A ed 1985 fconQrnjc Eoullibrium Model FOrIDujaljQo and Solution Mathematical
Pogrammlng Study 23 Amsterdam NorthmiddotHolland
Manne A 1985 Or the Formulation and Solution 01 Economic Equilibrium Models In A
Manne ed ECQOQlli-Jr~ibmiddotum ~tQdel EQlJulailon and Solution Malhematical Programming
Study 23 Amsterdam NorthmiddotHolland
Mansur A and J Whalley 1984 Numerical Specification of Applied General Equilibrium
Models in Scarf H J Shoven Eds Aoplied General Eouillbrium AnaiysectIsect Cambridge Universily
Press Cambridge
Merrill 0 1972 Applications and Extensions of an Algorithm thaI Computes FixedmiddotPoints of
Cenain Upper Semimiddotcominuou Point to Set Mappings PhD Disserlation Unlvel$ity 01 Michigan
Negishi T 1960 Welfar Economics and Exislence I an Equnlbrlum for a Competitive
Economy MelroeCOOQ11ka 12 92middot97
Pereira A 1985 On the Existence and Uniqueness of Optimal Ou1put Path lor CRTS Firms in
Adjustment Costs Technologies Mimeo Stanford University
Pereira A 1985a Consumer Expectations and Ihe COSI 01 UllUty In an Intertemporat
Framework Mlmeo Stanford University
Pereira A 1985b Corporale Tax Integrallon and Government Deficits A Dynamic General
Equilibrium Anaiysls Mlmeo Stanford University
Pereir A 1987 On Ihe Computation of the Users Costs of Stocks Ecooomi (forthcoming)
47
Pereira A bull 1987b Equal Yield Alternatives and Government Oeficlts Mimeof Stanford
University
Pereira A 1987c A Dynamic Applied General ~uHlbrium Model for Tax Policy Evaluation
Comlletamp Docurnent2tion Mimeo University of California San Diego (in progress)
Pereira A 1987d Should the investment Tax Credit be Rei(ltroduced Mlmeo University
of California San Diego (ir progreSs)
Piggott J and J Whalley eds 1985 tJew QeyeQpments in Applied General EQuilibrium
bullenalysjs Cambridge Cambridge University Press
Preckel Pbull 1985 Alternative algorithms for computing economic equilibria In A Manne Ed
Radnor R 19amp3 Equllibrium under Uncertainty In K Arrow M Intrilligaor ods
I ibull
Robinson S 1 g86 Mullisectorai Models of Oeveloping CountrieS a Survey Oivision of
Agricullure and Resources UC Borkeley WP 401
Scarf H 1967 The approximation of fixed points of a continuous mapping ~MM Jouwal of
epplid Mathornallcs 15 pp 1328middot43
Scarf H with T Hansen 1973 00 the Computation of EconomiC Eguilibria New Haven Yale
University Press
Scarf H J Shoven Eds 1984 61l1llied Genera Elujlibrium 6nalyssect Cambridge University
~ Press Cambridge
Shoven J 1976 The Incidence and Efficiency Effects of Taxe on Income from Capital
Jurnal of ramie1 Economy 84 pp 1261-83
Shoven J 1983 Applied General Equilibrium Tax Modemng IMF Slaff PaoerS 30
Shoven J and LB Simon 1986 Share Repurchases and Acquisitions An Analysis of which
Firms Participate NBER Working Paper No 2243
bull
48
IShaven J and J Whalley 1972 ft General Equmbrium Calculation of lhe Effects of
Differential Taxation of Incoma from Capital in the US Jauroe of public Econgmics 1
pp281middot321 (
Shoven J bull and J Whalley 1973 General Equilibrium with Tax$s a computation procedure
and an existence pr00f ReviEhi of Economic Studies 40 pp 475middot490
Shoven J J Whalley 1977 Equal Yield Tax Alternatives A General Equilibrium
Computa~ui0nal Tfchnique J~HJrnal of PUblic EconomIcs 8 pp 211~224 bull
Shaven J J Whalley 1984 Applied GeneralmiddotEquilibrlum MOdels of Taxalion and
International Tr8de An Introduction and Survey Journal of Economic literature 23t pp
1007middot1051
Slemrod J 1980 A General Equilibrium Model of Capital Income Taxation Ph D bullDissertation Harvard Univnrsity
Slemrod J 1983 A General Equilibrium Model of Taxation with Endogenous Financial
Behavior in M Feldstein edbull Behavioral Simulation Methods In Tax Policy Analysis Chicago
UniVerSity of Chicago Press
Stone J 1985 Sequential Optimization and Complementarily Techniques for computing
Economic equilibrium in A Manne Ed
Summers L 1981 Taxation and Corporate Investment a q-Theory Approach BroQkinas
Summers L 1985 Taxation and the Size and ComposHion 01 the Capital Stock An Asset Price
Approach Mlmeo Harvard University
Whalley J 1975 A Geneal Equiloibrium Assessment of the 1973 United Kingdom tax
reform Economica 42 pp 139middot61
1JLTIllOS JIQll~)ltG PAPERS PUBLICADOS bull
I
I I
j
112 70 - LUCENA iJiCIIJo ~ IIEnvironrilent Honitoring and Orqanization Structure T (Setellibro 1987)
nI 71 - L~ENA Diogo A Note on the Representation of Information Stxueture (SeteJObro 1987)
nQ 72 - LUCENA D1CgO Envuorment 111)nitoring and Organization
Structure II (Setelilhro 1987)
ril 73 - AlmmE3 Antonio Pais e CAsPAR Yitor bull Tributacao Inoentivos e InvesUlllentos Analise Qualitativa noYellibro 19S7)
n2 74 - liAlltOSA rose Pedro Opt1lllal lIage R1gidity A SUggestedtlthodology to Teet tb Theory and an Application (Outubro 1987)
nI 75 - CABRAL L 11 B Three lIotes on SYJIIlletric Gemes witt itsYJIIlletric Equilibri (NoveJObro 1987)
n2 76 - LUCENlt D10gOTo Search or Not to Search (Fevereiro1967)
nQ 17 - SA Jorge asconcelos euroI A Theory of SynergyB (Fevereiro1987)
n2 76 - VILARES tanuei Jose Os Be IntenaMios IlIIportados ColIIO Factor de Produao (Julho 1987)
nQ 79 - SA J~rge VasCQncel03 e HoY to Copete and Co~unicate in tature Industrial Products_ (aio 1988)
n2 80 - ROIl Rafael Learning and Capacity Expansion in a New larket Under Uncertainty (Fevereiro 1988)
n2 81 PEREIRA Alfredo llarvilo Survey of Dynamic Computational General EquilibriUlll Hodel for Tax Policy Evaluation (Outubro 1987)
Qualquer in ormacao sabre o 1IIorlltlng Papers ja publlcados sera prestada pelo Secretariado de Apoio aos Docentes podendo os ~e3OSf
ser adquirido5 ne Secyo de Vlndas da Faculdade de EconollLia UNL na Travessa Esteviio Pinto Cltpolide - 1000 LISllOA
--
- - - --------
--------- - - - --------
---
1nltJ1 bull I ~ 1J11ltI1 v I ~
-
Sa Ita rd-Futlertonmiddot intargration consumption lax ~StHlven~Whall~y VAT capila~ or labor tax
Ch~mqflS1985 Andersson
consumpion tax neutral corporals lax system 1987
it ue rb a c h~ i(at II Sltotf COIlSUft1ptian lax (poundpird ItS hibor Mx~on ufocto of taxation on capital fOfrnJilti1
dvnamio fiscal policv effects af OQvernmanl defioit (3tlard-Goulder
1883 1987
allocts 01 the adoption of a 1ge3 1985 pure consumpHon lax
im~rtmce of cons fCreSClht -Bovenborg analytically orianiad approach
1985 1986 effects ot pure conumption tax capllal income taxation in Opo economies
Erllch-Glnburghshy-Heyden reat wage policios in Belgium
1987 Felter1steln 1984 - theorotical
1985 ~ linnncial crcwdlng out1984 1986
Goulder effects of tax- vs bond-financed 1985 chanQes In pubtlc expondllufe
~~--
Gouidr~SUmmampfI torpora1$ lax cvts reduced 1987 ITO increase 10 gasolioe taxes
Jorgenson-Yun
effects ot several programs of tax reform on the allocation of ~~ptai
1984
-- Purelra
corporate la)( Integration effects cf romiddotinlroducinQ imreslamnl tax credits
1986 1967
10 bull
bull -~- ~- --- _ - ---____--_ - -----_-_ --~ ---- ~
1 1
2 Brief overview of non-dynamic CGE models
The origin of most or the models we survey goes back to the work of Scarf (1967 1973) whlgt
first developed a reliablo algorithm to compute equilibrium price for a ArrowmiddotDebreu economy
HiS algorithm used simpicial Bubdivision techniques and can -be shown to be the computational analog
of the fixed point theorems previously used to prove the existence of equilibrium His technique
could soW a model with an arbitrary number of consumers and commodities as long as all agents
were price takers consumers were subject to budget constraints demands were continuous andmiddot
production did not display increasing returns to scale The alQorithm while guaranteed to converge I
I I I
was relatively slow for prcb1ems involving more than S1If 20 dimensions A maior improvement in
computational speed Woo offered by Merrills algorithm (1972) which used tpe same fundamental
ideagt as Scarfs procedure
Scarfs model (as is true for the standard Arrow-Debreu model) does not include a government
sector ~ neither taxes nor public goocs At one of the mosi promising appHciUions of the new
computational techniqu w~s in Iha area of tax policy evaluation Shoven and Whalley (1973)
exlended the general equilibrium model and computational approach to include a wide rry of taxes
and a governl11ampnt spending plan The original ShavenmiddotWhalley model was static (although the
different commodi~ies CQuid be considered simitar goods available at different datest as In
Arrow-Debrau) and government was aumad to run a balanced budget The data are arranged as a
social accounling matrix The models spacification and calibration is checked by solving it in the
presence of the base set of laxes The result should ba exactly tha initial social accounting matrix
After having pasd this replication check the model is solved for a counterlactua equilibrium In
the presence of a new tux design The result is once again a social accounting matrix The two
equilibria are compared in order to tSS($S the impact of the new tax plan The first uses ot this
bull
1 Z
model for lax policy evauajon were ShovenWhalley (1972) Whalley (1975) and Shoven
(1916)
Completely static models as Shaven-Whalleys are unsatisfactory for many tax reform ISsues
These include corporate lax Integration effects of investment tax credits effects of accelerated
depreciations consumer Or e)penditure tzxes~ importance of saving subsidies like IRAs etc These
are essentialy dynamiC ksues They involve not only the allocation of capital across sectQfsbut
perhaps more Important the capital intensity of the economy But the capital accumulation and
capital reallocation take time and may involve adjustment costs Because of these issues Ballardmiddot
Fulierton~Shoven~Whaloy took the first steps towards developing a dynamic model
In this conlexl the ShovenWhalley model has been extended and implemented for Ihe US
-economy In BuIIBrdmiddotFulienonmiddotShovenmiddotWhalley (1985) While Ihis book completely documents the
mo~el it was used in several publicalions beginning in 1978 Their mode) consists of nineteem
production sectors twelvo households and fifteen consumer goods It includes a very detailed set of
taxes including the federal and state personallncome taxes federal and state corporate income taxes
SOCial Security taxes pro1J0rty taxes unemployment Insurance excise taxes sales taxes etc It
has been calibl att-d to reproduce the 1973 US economy Recentty a version corresponding to the
1983 JS economy has been developed
wThe 8alard-Fullenon-Shoven Whalley model is dynamic In the sense that consumers face a
choice between current consumption and leisure versus future consumption (which can be
purchased via s wlngs) The consumer classes act as if they were maximizing a nested CES utility
function over the domain of a CES aggregate of contempo-raneous consumer goods and leisure and
future consumption subject to their income constraint The parameters of those funcHons
determine the shares of income devo1ed to each commodity to saving and to the purchase of leisure
They also determine the two key elasticities in the models ~ the elasticity of labor supply with
bull
13
respecl to tgte rbullbull1 aoer lax wage rae and the elasticity of saving with respecl 10 the real after lax- rate of retum to capital
In the model consumerS have myopic expectations regarding lulure prices and in particular
regarding the future rale of return 10 capilal Ballard (1983) and BallardmiddotGoulder (1985)
inco~ora1e both perfect foresight and limited foresight into this model Future consumption is
acquired by buying a fixed composition portfolio of real investments that offer an infinite annuity
of returns
The production side of Ihe model is compielely static The mod1 Incorporates a constantmiddot
olasicily 01 substitution betweeen prlmary inputs in production (capital and labor) and fixed I coefficients for intermediate inputs The model distinguIshes between Industrial outputs and
eonsumer goods for the simple reason Ihat tile data are classified differenUy Industrial sectol$ I I involve such categories as forestry and fisherle metal mino9 and publishing and printing while
consumers purchas-3 fUfHi1ure 2ulomoblles and books This fact is recognized in the model by Ibull
incorporatirg a secord stage Of production which converts industrial outputs intO consumer goods
This technology is uStally mod9-led as a flxed~coefficient conversion matrix
The BallardmiddotFINton-ShovenWhalley model assumeS that the private sector finance
marginai ioveslment with the same composition of debt and equity as currently exists in each sector
This is the same arsumrtion that Herberger originally used Investors all hold debt and equity in tile
same proportion Thampreiore this ownership can be aggregated into simply capital ownership For
tax purposes however the separate treatment of debt and equity is taken into account at both the
corporate and persofa Io=vel SimllarlYJ the dividend poIicies on corporale equitY are established
exogenously There are no government bonds in the model since there are no government deficits
The modol is solved for a sequence (as many as 100) of temporary equilibria wltll consumers
allocating income between present and future conumption at each point in time The palh for 1he
14
economy ~ a sel of connected equilibria The connectIon is pra~ided by capital accumulation
Capital accumulation is endogenous and determined by saving The model stans with a sociagt
accOunting mi1rix In the bzse case the economy is assumed to be in a steaqy state growth -path (along
which all relative prices ~fe constant) The madel solves for both the naw steady state growth path
and the transltion 10 it zf~er a policy intervention rhe authors have frequently addressed the
question-of howong rt tKes to effectively settle un10 a new steady S1ate growth path
The dynamics of the model are limited however in that future consumption is collapsed into a
composite commodity Aso the absence of government deficits and the lack ofproductlon dynamics
limits the realism of the BallardmiddotFuliertonmiddotShoven-Whatley model for the analysis of dynamic
policy issues (such as the adoption 01 a consumption tax or Ihe elimination of the investment tax
credit) The models we survey try to Improve upon Ihe Ballard-FlIiertonmiddotShaven-Whalley
dynamic modelling The models Jn~orporate more forward lookTng behavior They improve the
dynamics on the conampumrtkm side of the economy and some of them introduce true 1ynamlc behavior
on the part oI the producers We turn now to the issues inY9lv_ed in developing such dynamic models
15 I I ~3 ~ssues in the modelling Of dynamic behavior
We will 100k firt at the speclficalion of economic behavior of consume(s~ producers
goyemment and Ihe r~sl of Ihe world The modemog of corporale and household financial decisions
I is then addr5sed Finally Ihe eencept of equilibrium is discussed
I
i
I 1
31 Consumers behavior
Early efforts 10 build dynamic features into the economic behavior 01 consumers are due to
Ballard (1963) and Auerbach-Kotllkoff (1983 1984) Their work has been closely followed by
several subsequent authors In faet dynamic household behavior Is the single most pervasive aspect
amohg the models surveyed
All the models Inc-O1orale some loro of lila-cycle behavior Household behavior is dewmloed
by the maximization of an addWVely separable time invariant intertemporal utillty function The
utility lune1ion is defined oyer the domain of the consumption goods In the economy In most of the
models leisure is also an argument in utility so that labor supply is optimally determined
Utility maximi4ation is subject to a lifetime intertemporal budget constraint which equalizes
the presen1 value of consumers income and expenditure More reeentiy in Andersson (1987) bull
Bovenberg (1985) and Pereira (1986b 1987d) the constraint i$ defined as a sequence 01
recursive equations of motion on wealth This has the potential advantage 01 accomodating liquidity
constraints However It should be recognized that in the absence 01 liquidity constraints (ie when
consumers are free to completely borrow against future income) the two specificallons of the
househofd ccnstraint are essentiay equivatent Furthermore tn both versions saving is optimally
determined as a way of translaring wealth intertemporally
1 6
Some lodels now have a sophisticated dynamic specification of hoosehokl In Ballard (1983) -
BallardmiddotGoulder (1985) and Auerbach-Kotlikof (1983 1984 1987) consumers behavio is
embedded in an iterQEneraional setting Each of them has S5 age cohorts simulataneously alive
Som-e measure of jntergener1ional altruism is considered in the form of bequest motives In Ballard
(1983) and Ballard-Gouldr (1985) consumers derive utility directly from bequthing parl of
their wealth Therefore bequests assume 1he form of a scrap value of terminal wealth of a
generation In AuerbachmiddotKotlikof (1983 1984 1987) each generation empathyz bullbull with Ihe
utility of future generations ovar bequethd wealth In addition Ballard (1ge3) includes detailed
demographic poiections wfill the intent of incorporating in the policy analysis the effects of the
bull Post World War tI baby-boom
In the real world household decisions also include the optimat allo~atiOn of saving among
shyalternative physical or financial assets Theories of household portfolio behavior are relatively
Iwen~establshed However they involve uncertainty as a crucial element Accordingly they have
middotnot been incorpora~ed in the context of the determinIstic dynamic models we sutyey here We will 1 Come backt-O this Issue below
32 Producer behavior
The efforts to build dynamic features into the economic behavior of producers are more recellt
and less widely adopted The first attempts are due to Bovnberg (1984 1985) and Summers
(1965) Dynamic behavior has been more fully incorporated In the recent models of Andersson
(1987) Auerbach-Kotlikoff (19B7) GouldermiddotSummers (1987) and Pereira (1986b 1987d)
Part of the reason why production side dinamlcs has been more slowty adopted is the weak
supply of accepted theories referring to the dynamic behavior of the firms In the models referred
I 17
to above gynamlc producion and investment b~havior are middotinduced by the existence of capital
adjustment costs and linked to Tobins q theory Adjuslment costs are designed to capture both the
incomplete mobility of capital aoross industries and installation costs ~e the costs of adjusting
capital towards its cpHmai level)
Adjllstment costs can be conceived as internal to the firm and measured in terms of foregone
outpul along the lines of Lucas (1967) Alternatively adjustment costs can be viewed as actual
external cests incurred together with Ihe purchase costs along the Unes of Gould (1969) With the
exception of Pereira (198Gb 1937d) all of the CGE authors follow the former approach
The firms in the economy maximize their market value as the present discountad value of the
future stream of dividends In Andersson (1981) and Pereira (198Gb 1987d) firms are seen as
maximizing the present discounted value of net cash flow Maximization is constrained by Ihe
adjustmenl cosl lechnology and ~n equation of motion describing Ihe evolution of the capital stock
It should be noted that undereogenous divldendrelention rules the two problems are equivalent II
should also be nOled Ihat a salisfectory economic rational for th bullbullistoo of dMdonds with the
present tax code is missing in the profession (Shoven (1986))
The models with static formulation of producers behavior are characterized by passive
investment behavior Investment merely accomodates to saving in the eccnomy With a dynamic
formulation induced b adjustment costs real investmen decisions are forward looking Investmenl
Is endogenously and optimally determined by the firms A fundamental difference between the
shonmiddotrun in which capital stock is given and longmiddotrun in which Ihe level of capital is allowed to be
optimally determined is emphasized
This extra richness of produc1ion dynamics is not without costs A careful look at the summary
tables will clearly show an inverse relation between the adoptlon of dynamic features in production
and the level of disagreggation of the production side of the economy In fact with production
bull
j
1 6
dynamics $e dimension of the problems is immensly increased Let IS be more specific
In a static framework with constant returns to scale production technology the output level ismiddot
indeterminate The optimal aliocation of inputs can be obtained by cost minimization with any
feasible output level generatillg zero profits Such zero profit conditions are used to solve for the
output prices in terms of the factor prices and hencemiddot to reduce the dlmensionality of the problem
from the -number of commodities and factors to the number of factors Thus even if the model deals
with 30 production sectors the computation of an economic equilibrium can take place using only
the dimensionality of the primary inputs in the economy
Now under certain regularity conditions on the production and adjustment costs technologies
leading to enough concavity of the optimality objective the intertemporal output path for the firm is
endogenously optimally and uniquely determined even with constant relums to scale technologies
(see Pereira (1985 1987a) Accordin~y with adjustment costs in the model optimal profits
will in general be non-zero This result has crucial implications for the computatiQn of equilibrium
in the model~ with production dynamics With adjustment costs no reduction of dimensionality is
possible The curse of dimensionality returns as a binding constraint
The introduction of adjustment costs adds another significant complication to the model With
capital being less that perfectly mobile in the economy different rates of return on capital will
exist in different sectors This is a difficult problem to tackle conceptually in the absence of
uncertainty Also in the real world producer maximization choices include also its choice of
financial ratios (debVequity) and payout rates (dividendretained earnings) These financial
subjects are difficult and much Clf this behavior is still taken as exogenous by the modellers We
will come back to these issues below
middot i
19
bull 33 government behavior
middot
Two issues dominate the modelling of government behavIor First govemment behavior has
Ibeen typically seen in Ihe CGE liIerature for lax poliCY evaluation as constrained by yearly balanced
I budsets (see for example 8allard-Fullrton-S~ovn-Whafley 098Sraquo The analysis of
govemment defidlS and public debl in a CelE context requires a dynamic selling Second the level of
government expenditures is either exegenouty given as in Auerbach-KoUikofi (1984 1987)
6ovenbrg (1984 1985) Feltenstein (1984 1986) and Jorgenson-Yun (1984) or
endogenously (but no optimally) determined by the balanced budget conditions as in
6allard-Fullerton-Shoven-Whalley (1985) Andersson (1987) Erllch-Ginshurgh-Heyden
~
(1987) Gouder (19aS) and Gouder-Summers (1987) In the second case the composition of
bull public expenditures is often optimally determined However the leve~ of government expendi1ures
can only 00 endogenously and optimally determined if the government is seen as an optimizing agent
and is allowed to run deficits
The first attempts to deal wiU the government defiCits in CGE tax models are due to
Auerbach-Kollilltoff (1984) and Feensain (1984) In Auerbach-Kotlikof (1984) government
expenditures afe exogenous They grow at the rate of growth of population However given 1he tax
structure and tax revenues yearly deficits and surpluses are atTowed subject to an in1ertempond
constraint that the present value of future tax revenues equals the present value of future
expenditures
In Fellenstein (1984 1986) government expenditures are also exogenously given He also
allows the government to run deficits to finance expenditures in excess to 1ax revenues However
surpluses are returned to CQnsumars in the form cf transfers Accordingly government ts not
subject to any constraint regarding the future repayment of public debt
In GouJder (19851 the 9Vernmenl maximizes a static social welfare ~nctio subject to a-
bull
20 I i
I
I I I
I bull
balanced budget conSirtlL This follows the optimal alocatkm of government expenditures alpng
the lines of Baliard-FHeiton-Shoven-Whalley (1985) Ttle model Is generallzea to allow for
exogenous changes in HIe time path of government expenditures Two financing alternatives are
consjdeft~d and contr8s1ed EldditionaJ tax revenues and bond issuance
Pereira (198Gb 1137d) auempts to address both the incorporation of deficits and the
determination of govement expenditures The path of government expenditures and the path of
delicilSlsurpluses (ano therefore the path lor debt) are endogenously and optimally determined The
government is seen as mltJximizing n intertemporal social welfare function given the 13x structure
Optimization is subject to a sequence of recursive equations of motion reflecting the evolution of the
ptlbllc debt allowing fDr government budget imbalances It should be stressathat this specification
01 the government cOMtcalnt is equivalent to the specificaUon in Auerbaeh-Kotlikoff (1983 1987)
when no liquidity constraints affect -gove~nment behavior
The extra richness in the treatment of government behavior allows tfte examination of
government debt poices in a truly dynamic setting Also financial crowding out effects induced by
government deficits can be analyzed (see Auerbach-Kottlikoff (1987) Feltenstein (1986) and
Perer (198Gb 1987e))
34 Foreign sector
Most of the open economy models follow the assumptions of balanced trade with import and
export net demands characterized by constant elasticities along the line of Ballardmiddot
-Fullerton-Shaven-Whalley (1965) Such is the case of Salfard (1963) Ballard-Goulder
(19S5) and Goulder-Sllmmers (1987)_ In Feltenstin (1985) the rest of the world is treated as
21
an addjlion~ consumer group
Bovcoberg (1986) develops a model In which two economies lIr considered each following
intertempoml perlect foresight paths These economies meet in the international forum Their
trade relaroMhips are characterized by yearly balanced trade accounts
None of the new generation of dynamic cae tax models has yet IncorPorated the lnternational
capital flows as done in aoulder-Shoven-Whalley (1983) for the earlier Ballarrshy
-Fullertonmiddot Shaven-Whalley (1985) model This is unfortunate as there is an important
Intertmporal aspect to international lending The first al1empts along Ihese lines are due to
Andersson (1987) and Erlich-Glnsburgh-Heyden (1987) Andersson (1987) in his model of the
Swedish economy adopts a closed economy appraoch in which rates of return in the domesJic
economy are largely determined by the international capital markets Fidinterest rate induce
Intlnatlonal capital flows which determine and finance the International lrade imbalance In
lurn In ErlichGinsburghHeydenmiddot (1987) foreign trade Is generaled according to an
intertemporal Irade welfare lunction with constant import and export elasticities In the shortmiddotrun
they allow inlernational trade inbalances which generate capilal ftows to the domestic households In
the long run however trade balance is assumed
35 The need for financial markets
A dynamic economic Wucture not only provides the ideal environment 10 model many features of
economic behavior it also permits one to Incorporate the financial side of the economy If the
government is allowed to run deficits the question 01 deficll financing automatically follows If
investment is optimally dewmined and returns to capital are different aClOSS sectors problems of
investment financing arise If there are seeral final1cfal assets In the economymiddot government
4-
I i
ibull
I I I I
I
j
22
bonds privpte bonds and equity (or simply physical capital installed inmiddot different sectors) the
problem of allocation of saving among assets with potentially different returns arises
The papers surveyed var greatly with respect to l~e extent of their attention to the financial
side of the economy At Clne extreme are the models in ~anardmiddotFuliertonmiddotShoven-Whalley (1985)
Andersson (1987) Ballltld (1983) Ballard-Goulder (1985) Bovenberg (1985 1986) and
Erlich-Gjnsburgh-Heyden C19B which are devoted exclusively to the real side of the economy
In turn Auerbach-KotliKoff (1984 1987) Feltenstein (1984 1986) and Goulde (1985)
allow for government debt In these models saving finances changes in government debt and physical
capital Private and publlc 2ssets are perceived by the households as perfect substitutes The
allocation of saving merely adjusts to the relative demands for funds
Feltenstein (1984 1986) is the only model surveyed which introduces money Government deficits are financed by issuing money and bonds according to an exogenously given rule Money is
demanded by consumers for transaction motives and an exogenously given fraction as a store of value
On the other hand government bonds and physical capital are the vehicles for the intertemporal
transfer of wealth
Summers (1985) and GOlldermiddotSummers (1987) introduce a whole menu of financial assets shy
firm specific equity capital Different assets earn different rates of retum However such rates
are equal up to constant and exogenous sector speCific risk premia Therefore the introduction of
constant exogenous risk premia as helpful as it may be in the context of calibration does not solve
the main issue of the non-optimality of the allocation of saving Also talking about risk premia in a
deterministic context is somewhat unsatisfactory
Pereira (1986b 1987d) also introduces a whole menu of assets private and public bonds and
firm specific equity However all the assets are expected to yield the same rate of return and
therefore perc~jved as perfect substitutes The different asset types allow consideration of
23
exogenOU$~8tUequjty triO dividendlrc~enlion rufes and therefore seve-ral sources of investment
financing bond equity and retained earnings
The nonmiddotcptimalHy of the allocation of saving and the absence or exogenelty of corporate
financial rules is _a reftectlon of the limitations of the determInistic approach Either in a context of
perfect anticipation oi the future prices Of in general within the t$atm of point price
expectations all the papers follow a deterministic approach Under such circumstances consumers
either expect diHerent rales of return (inclusive of risk premium) across assets in which case
they wilt buy only one asset (that with highest rate) Of they expect equal rat~s of relurn in which
case Ihey are indifferent about the asset composition of Iheir portfoliO There is no way of
Iradingmiddotoff rales of return and risks to obtain an optimal interiof solution to the problem of the
allocation of saving
The most advanced contribution in the modelling of saving allocation in a CGE setting is due to
Slemrod (1980 1983) in the context of a static one-period model In his model consumers act
according to a two stage separable decision process They Hrsl decide on how much to save Then
they decide on the allocation of saving according to an Indirect uttlity function dependent on the rates
of ratum and variances offered by the different assets in the eoonomy The sourCe of riskiness in the
ampConory comes from an uncertain marginal product of capital On the other hand aside from
portfolio decisions the res of the economy is insulated from uncertainty
The most complete contribution in terms of the treatment of the corporate financial rules is
FulienonmiddotGordon (1983) In a variant 01 Ihe BailardmiddotFulienon-ShovenmiddotWhelley (1985) model
Fullerton and Gordon have capital intensity and optimal financial decisions Jointly deter~ined
through a two siage process The cosl of financing capital is minimized by trading off the tax
advantages of debt against the epected real bankrupcy coSls inherenl with high debVequity ratios
Given the optimal debtequity ratio the level of inveSlment Is chosen such that at Ihe margin the
24 l
bull
rerurn pn ~uity equas the return on bonds plus an exogenous risk premium
36 Market assumptions equilibrium and expectations
Virtually all of the papers arc characterized by Walrasian market clearing assumptions All
markets are perlect1y competi1ive Atomistic competition among agents 1s assumed even thoughmiddot only
a flnile number of agents is considered Virtually no mark~t disequilibria or price stickiness afe
considered The only exception is ErlichmiddotGinsburghmiddotHeyden (1987) bull In thei model of the Belgium
economy I the _wage rate is fixed in the short-run Therefore in the shOrlwrun desequilibrium in
the labor market will generate endogenous unemployment However in lh$ long-run all prices
Including the wage rata are Ilexible and accordingly ali markets ciear
Given the dynamic nature of behavlor in the economy markelmiddotcfearing prices in each period
depend on expectations of fture prices and on tax variables In the economy There are essentially
two ways at interpreting the economic equilibrium in such a dynamic conte~bull If future prices are
perfectly anticipated (io expectations are self-fulfilling) a perfect foresight equilibrium
prevaHs Then future actions are merely the implementation of current decisions for future
periods However jf ptica expecla110ns are not perfect (Le agents make mistakes w1th re$pect 10
future prices) then a temporary or shortrun equilibrium prevails Markets ciear and clearing
prices depend on fu1ure price expectations Current plans about the future are typically not
precisely implemented They will be revised as more or better information becomes available to the
economic agents See Grandmant (1982) for a comprehensive survey 01 the temporary equilibrium
literature
Wilh the exception of Gould (1985) and Pereira (19S6b 19S7d) all the models surveyed
adopt th~ concept of a perfect foresight equilibrium In turn BaliardmiddotGouldr (1965) considers a
J
2S
flexible am9unt of foresight In terms of the number of years over which price movements are ~
foreseen Pereiras model (1986b 1987b)) is flexible in a somewhat different way in that it can
indude any range of foresight from myopia to perfect foresight The choice between the perfct foresight and lemporary equilibrium is ultimately to be made on
philosophic grounds It can be argued that less than perfect expectations Imply that agents are
irralional in some way (see Auerbach-Kctiikoff (1987 p 10) However the reverse argument can
be made One can question wiether agents are reaHy rationar and perfectly knowledgable about
future prices
Recent evidene of Balird (1987) Ballard-Goulder (19851 Goulder (19851 and Pereira
(1986b 1987d) suggests that the choice in modemng expectations is an important one They show
that the degree of foresight inl0 thc future (ranging from perfect foresight to-myopic expectations)
may have dramatic impacts on the pollcy conclusions of Ihe model Accordingly the best research
strategy may b to design models which are flexible enough 10 allow for different rules regarding the
formation of expectations With the exception of the articles just mentioned sensitivity analysis
bullhave previously not been performed along this dimension
In terms of implementation the two concepts of equilibrium - perfect foresight and temporary
equilibrium have different implications The dimensionality of the equ1ibrium soll1ion algorIthm
is involved Suppose we have a model with ten markets to be run for a period of 50 years Aside
from normalization a perfect foresight model implies computing prices In 500 dimensions while a
-t~mporary equilibrium model requires sorving SO equilibria each in ten dimensions Given that
computational speed often varies with the cube of the number of dimensions the lemporary
equilibrium formulation is potentially strikingly more feasible However Ballard (1987) and
Goulder-Ballard (1985) have devoloped techniques to greatly speed the computation of a perfect
foresight equilibrium
26
j
I
4 Jrnplemenlalion and Issues on policy evaluallon
41 Calibration and equilibrium comparlslon
CGE models at typically parameterized by the use of a calibration procedure Some parameters
are exogenously given However some crucial parameters are determined in such a way that the
model replicates the data for a given base year See MansurmiddotWhalleY 11984) for an eXlensive
discussion of 1~i$ issue
CalibratIon in a dynamic context is generaliy interpreted as requiring two properties First
replication of base year data is required Second the model is parameterized to simulate an
intertemporal balanced growth path when Ihe base policy is maintaineq This Is the approach
followed by Ballard (1983) BallardmiddotGoulder (1985) Goulder (1985) and SummersmiddotGoulde
(1986) It follows the practice of BallardmiddotFuliertonmiddotShovenmiddotWhalley (1985) with their
sequential cquilibrium dynamics
Other authors Auerbach-Kotlikof (1983 1984 1987) Boyenberg (1984 1985 1986)
and Pereira (198Gb 1987d) follow whal we cali a qualitalive calibration The slructural
palametars are exogenously chosen so that lh economy follows a reasonable path into Ihe fulure
ThaI has 10 do wllh the fact thaI given the recursive nature 01 the dynamic economy and aside tom
such structural parameters only iniUa stock values are needed to run these models Given initial
conditions on the stocks of say private wealth capltal and government debt agents will optimize
and Ihereby generate a lirst round of net demands and equilibrium conditions In lurn the
equilibrium prices will determine the evolution of the stock variables into the next perIod
There are several potential problems with calibration in the conlext of dynamic models The
assumption of a steadymiddotstate growth ~ath in the base case can be questioned First while
27 I Ibullbull
steadymiddotstaU is a possibity it certainly is not the only meaningfull solution to dynamic models
Even in the case of a perfect foresight equilibrium the model implies an equilibrium path which i may or may not involve colanced growth The model not the modeller should dictate the nature 01 I themiddotbase case path Second in the context of a temporarY equilibrium path a steady state solutlcn is Inol a likely model QutcOn1tl In fact unlessmiddot expectations are stattc short run behavior consistent
with a steady-state evolution will in general not be generated On the other hand if static I expectations are self~fufjili[1g we have in fact a perfect foresight model Third even the base year
replication requirement may cause problems In Ihe cOntexl of temporary equilibrium Any
calibration parameter would be conditional on expectation rules which is probably an undesirable
feature
The nalur of the two-requirement calibration strategy is very muchin the spirit of the bull
traditional design of the comparision of alternative equilibria comparis_ion between a steady-state
base case on one hand and alternative paths including a transition period and a finat sleadyyenstate on
the other hand Pereira (198Gb 1987d) compare different (not necessarily steadystate)
equilibrium paths The arguments against such a procedure are based on the idea that the impact of
policy changes can be observed most easily since all departures from the steadymiddotstate can be
attributed to the altemalive polioy On the other hand the base case so dEfined as a steadymiddotstate is
consist~nl with previous work in a less dynamtc setting and Iherefore allows a common standard
for comparing model results
42 Computation algorithms
We are still al a stage in which basically each author uses a different computational technique
Th development of dynamic models - in parlicular wilh adjustment costs andior perfect foresight
bull
28
has corresponded with the decline in the use of fixed point algorithms In -fact given the relative
arge dimensions inevitably involved such algorithms lend to be very inefficient at the best and
chen prohibitively stow See Stone (1985) and Preckel (1965) for a comparative assessment of
different computation techniques Among the models surveyed only Ball1rdFulle(ton~
-ShavenmiddotWhalley (1985) and Feltensteln 1985 use Merrills variant of the fixed point
algorithm 1echnique
AuerbachmiddotKoHikoff (1963 1984 1967) follow a three stage procedure They first compute a
base case steady-state then a revised case steady state and finally transition path for the economy
berveen these two steady-states tn all stages a Gauss-Seidel iterative procedure is used
Ballard (1982) Ballard-Goulder (1965) Goulder (HiSS) and Summers-Goulder (1986)
~
use a method developed by Ballard and Goutder which is similar in many awects to the FairmiddotTaylQr bull
(1983) algorl1hm Short-run equilibria are calculated (using Merrills algorithm) parametric on
nrice expeC~lions The model is then iterated to generate self-fulfilling intertemporaf expectations
and the corresponding perfect foresight equilibrium In a relatively similar approach Andersson
1987 uses a simulation program SIMNON developed In the University of Lund This program
can handle two-paint boundary problems In a fashion consistent with the multiple shooting
algcrlthm (see Lipton-Poterba-Sachs-Eummers (1982))
Boyenbergs computational approach (1985 1985) differs from the other models surveyed in
that he relies heavily on analytical techniques Computations are done by using a dynamic version of
Johansons linearization method Being essentially determined by the continuous time nature of the
I I
f
model thiS linearization model has the disadvantage of confining the analysls to infinltesinal changes
around the base case equilibrium (see Bovenberg (1985) p 53) Pereira (1986b 1987d) uses an optimization algorithm NPSOL - developped by
rGillMurray-Saunders-Wrlght (1986)) This algorithm is used 10 compute the sequence of
I
29
Shortmiddotrun temporary equilibrium which makes up the intertemporal equilibrium path The
equilibrium conditions are seen as nonlinear equality constraints in the minimization of an artlcial
objective function The prices are normalized to the unit simplex by an additional linear equality I I 1 bull constraint
II ErlichmiddotGinsburgh-Heyden (1987) follow a unique approach in that they use a variant of the
optimization technique introduced by Negishl (1960) The economic equilibrium can be generated i
as a solution of a mathemallcal program Ihe objective function of which is a weighted sum of the
utility functions of the various agents while tho constraints set consists of the market dearing 1
conditions Ginsburgh-Heyden (1985) have extended Negishis resulllO tho case of downward price
lrigidities I
Finally the paper by Jorgonson-Yun (1984) is also unique in tIlat it is the only I
econometrically estimated model Different blocks for the consumption and production Side of the
-
model are separately estimated to provide the necessary structural parameters i
I The diversity of computation techniques is yet another indicator of the exploratory nature of the
body of literature surveyed in this article
43 Equal yield comparisions
The link between a base case and counteriactual simulations Is usually provided by the concept of
equal yield Shaven-Whaley (1977) discuss the meaning 01 equal yield in a general equilibrium
context when government is not allowed to run defICits Equal yield is interpreted to moan constant
public utility Government base case utility is maintained in the counferfactual experiments With
balanced budgets the concepl of equal yield is unambiguous The new equUibrium prices and the
balanced budge condition will del ermine the minimal expendilure and taxes needed to maintain base
30
case publi~ utility Accordingly in general equal yield is inconsistent with equal nominal tax
revenue Some change in tax revenue Is necessary Different tax replacement schemes ate
considered to assure that enough tax revenue is collected This is the approach essentially followed
by most 01 the papers surveyed Only Feltenstein (1985) Goulder (198) and Jorgenson-Yun
(1984) chose not to follow an equal yield strategy
The question is of how to Interprete lila concept of equal yield when the government is alowed to
fUn deflcLts The optimal leve of expenditure for base ease public trliUty can now be taAinanced
bond financed or financed by a mix of bonds and taxation We have several versions of equal yield In
particular equal yield may now be consistent with equal tax revenUE Furthermore some meaSure
ltgtf financial crowding out effects of government deficns can be inferred from the comparislon of the
several equal yield alternatives These issues are extensively discussed ~Pereira (1987b)
44 Price expectations and the dynamic generalization of compensation
Indicators
The sum of equivalent and compensation variations over households is the most widely used
aggregated measure of efficiency gains or losses In the context of $teady~state eomparisions andor
when complete future markets exist andlor perfect foresighl is assumed there are no difficulties
associated with the use of the standard Hicksian indicators In fact correct future prices are known
in these cases
If expectations are not selfmiddot fulfilling andlor future markels are not available a dynamic
generalization is necessary 8allardmiddotGoulder (1985) provide some steps in that direction by
defining an indicator that accomodates periods far into the future when households do not have
perfect foresight but a steady~state prevails On he other hand this issue is more fundamental in
i
[
I
i I I I I [
I I
bullI
31
the contex~ of a Gner1 lemporary equilibrium ftamawor~ In such Circumstances Pereira
(1986a) dev~lops a cnamlc generalization which is obtained as the present discounted value of a i i
sequence of short-fUn optimal expenditure functions consistent with a base case expected future I stream of IlWities I
32
5 Empirical evidenee from selected policy Issues
Dynamic tax models have generated several important results which escaped Sialic modellers
The following is a seleCpoundd set of issuos whIch stress the marginal benefits of dynam~ modelHng of
economic behavior in innovltlttive areas of analysis The discussion of a consumption tax emphasizes
the benefits of dynamic tusehold behavior and intergenerational aspects The study of the
flliminatlon or the re~intToduction of the investsnemt tax credits is made meaningful by the dynamic
modelling of production betlavor Modelling governmenl in a dynamic conlexts allows the modeller
to study the irrrpact of government deflcits Finally a dynamic framework lets us appreCiate the I
relevance of consumer expectations in terms of the evaluation of policy alternatives
i
i
51 Consumption Tax
t
By Its very nature a consumption tax can not be adequately investigated with a static model bull
Fullerion-Shoven-Whaicy (1983) use Ihe model of the US economy s described in
Ballard-Fullerton-Shoven-Whalley (1935) to evaluate the movement from the currenl US tax Isys1em to a progressive consumption tax Since their modeJ Incorporates a Jaborleisure choice
where leisure is an untaxed commodlty their results reflect the fact that both the consumption tax
and Ihe present system are dslortionary
Concentrating on the rntertemporaf distortions they show that sheltering more saving from lhe
current US income tax could improve econornic efficiency even if marginal tax rate increases are
necessary in order to maintain government fevenue~ At first you have a revenue shortfall
However the economy moves to a higher sustatned growth path and uitimatefy lower tax rates
cangenerate the same revenue path Also wages increase in the long run with this policy The
bull
bullbull ~
33 shyt
i
I
I
present value of welfare gi[lS for a potiey of complete savin9 deduction wilh marginal tate
adjustments (consumption lax) is simulated to be around $500 billion to $600 billion 01 1973
dollars
They also investiga1e th 1ngth of ttme it takes th~ economy to adjust to these policy changes
Roughly they estimate the long run to b thirty years although this figure Is very sensitive to the
specification of the savings elasticity - a crucial parameler in this model
52 Investment T~x Credit
Goulder-$ummers (1 S87) address the impact of eliminating the lnvestmenl Tax Cradit (ITC) on
intersectoral capital formation and on economic growth Their model isYa1ticularly adequate to
addies$ this issu~ in that is postulates a forward looking investment m9we with adjustment costs
They show that the eliminating the ITC causes a reduction in the rale of investment Inveslment is
estimated to fall by about 7 in the short-run and by about 12 in Ihe new long run steady state
In lurn a previous policy anncuncement lowers Ihe overall al1ractiveness of Investment and leads 10
a downward shift of the Investrrent profile On the other hand the combined effect of a revenue
neutral simultaneous efimiraron of the ITC and reduction of the corporate tax rates is a long run
reduction ef the capital Sieck by 35 This pattern suggests that a revenue neutral increament in
both Ihe corporale tax rates and the inveSlment lax credits would be preferable 10 the formula
adopted in Ihe US Tax Reform Act of 1986
Pereira (1987d) addresses the impact of the ITe from Ihe oland point of Ihe curTenllax syslem
under the Tax Reform Act of 1986 Given Ihe concern over the potential depressiVe effects upon
savings and investment of the new tax system in conjunction with deficit reduC1ion mechanisms of
the GrammmiddotRudman type a pertinent question is should Ih lTC be e-introduced In the context
bull
34
-of his mod~1 of the US economy Pereira focuses on the trade-off between the distortion in the
intersectoral allocation of capital induced by the lTC the lower tax revenues it generates and
potential financial crowding-out effects on one hand and the positive effects of lowering the relative
price of new capital goods on the other hand Preliminary results confirm the qualitative results in
Goulder-Summers (1987) suggestting a welfare gain of about 2 of the present value of GNP in
the best scenario of absence of replacement taxes Furthermore preliminary results show an
important time pattern to welfare gains with the average benefits increasing the further you look
into the future This pattern is due to the presence of constraints to intersectoral mobility of capital
and inertia in the adjustment towards the optimal capital levels as reflected by the presence of
adjustment costs
I
I
bullAuerbach-Katlikaff (1gB) in the context of their intergenerational model of the US economy
consider the impact on savings capital formation and interest rates of deficit policies and balanced
budget increases in government consumption They conclude that deficit finance and government
consumption can significa-ntly crowd out capital formation and lower the welfare of future
generations However crowding out from deficit finance is a very slow process because it results
from increased government spending over potentially long horizons Also deficit policies may
substantially influence the longmiddotrun interest rates while leaving the short-run rates essentially
unaffected Finally and in opposition to the central role adjustment costs seem to play in both
53 Financial crowding outeff~cts of government deficits
Goulder-Summers and Pereiras models Auerbach-Kotlikoff suggest that the time path of interest
rates induced by a policy of deficit financing seems to be insensitive to the presence of adjustment
costs
bull
35
54 The role 01 consumer expectatlons
The expectations analysis has uncovered one of the benefits of dynamic modelling
Ballard-Goulder (1985) find that the appeal of adopting a consumption tax depends on the level of
foresight possessed by the consumers Furthermore additional foresight may be welfare worsening
This is a second best type of result This tends 10 occur under policies that lead to capital deepening
and declining rate of return to capital over time To the extent that consumers have more foresight
they will be beller equippod to anticipate the fall in the rental price of capital Consequently if a
saving incentive is enacted people save more with myopia than with perfect foresight Given the
exi~tence of taxes on capita and the discrepancy between the private an~--sccial returns to capital
the greater savings with myopia is beller socially Ballard-Goulder find that the welfare gains
from a consumption tax are reduced by about 10 when we move from myopia to farleet foresight
Therefore the attractivenness of adopting a consumption tax seems fairly [obust across these
foresight specifications
I
bull
36
6 Tbe power and weaknesses 01 dynamic modelling the example 01 corporale
tax Integrallon
The Issue of corporate tax integration allows us to sh~w the stre-ngths and weaknesses or the
dynamic approach Empirical evidence using CGE techniques indicates that depending on the precise
scheme of integration and the 1ax replacement methods integratiOn may have substantial effects In
Ihe work of FuliertonmiddotKingmiddotShoven-Whalley (1980 1981) total integration was found to yield an bull
annual static eHielency gain of $4 to $8 billion in 1973 dollars Simulated dynamic gains may be as
large as $695 billion or abcu 14 of the present value of tulUre consumption and leisure in the
US economy These gains result primarily from interindustry reallocalions of Investment and an
improved inter~emporal allocation of consumption
Mora recent work emphasizes hOw consumers asset portfolio decisions and firms finanCial
decisions affect the efficiency gains from integration Slemrod (1980) focusing on consumers
asset ponfoio deCisions finds static efficiency gains which are about twice -as Iarge as those
repOrted iiy Fulierton-King-ShovenWhalley (1961) FuUrlon-Gurdon (1983) focus on the
firms financial decisions They roporl efficiency gains of 6 of GNP from the elimination of the
tax distor1ions favoring debt However when they eliminate the corporate tax and repiace it with
increaSed persona income taxes addlHonal dls10t1ions are created in the optimal labot~leisure
decisions These distortions tend to dominate the analysis slJ(h that the overall effects of complete
integration are very modest Galpr-LuckmiddotTodr (1986) address simultaneously consumers
asset pOrlfollo decisions and the firms financial decisions They also find very modest efficiency
gains from integralion
The above results are important but may be severely biased There are severa aspects of
economic behavior and modelling crucial for the study of Income tax Integration whiCh have no been
i 37
captured innny of the above papers One aspect is the absence of government deficitsmiddot a balanced
budget is assumed It is true Ihal resource crowding OUI is caplured In these mOdels but financial
crowding out induced by government spending is nOI Second investment is not derived from
optimtztlcn behavior Investment behavior passively accomodates endogenous saving decisions As
a consequence the differential impacts of policies in the incentives to save and to invest are not
captured Aso full capital mobility across secors is assumed with instantaneous capital
adjustmenlS towards optimal levels This assumption rules qut different costs of capital across
seclors and therefore differentiated reactions to tax policies changes Finallyhe modelling of both
gov9rnmerlt deficits and of endogenous real and financial investment decisions necessitates the
~nsideraiion of a dynamilt framework and the introduction of financiaJ assets govemment bonds and
iprivate financial assets A dynamic framework also highlights the efficiency effects of Integration on r
ithe ptimal intertemporal decisions The above models are either sIalic (Slemrods and
GalpermiddotltJckemiddotTode(sj or a dynamic sequ~nce of otherwise Slatic model bullbull I Pereir a (1986b) develops a dynamic applied general equilibrium model bull to study the tbull
efficiency effects of integration on the growth and allocation of investmenl across sectors Special
Iattention is paid to the structural affeelS of resource and financial crowding oul induced by
gove mect spending and deficits and to the real and financial Investment reactions across industries
to both tax integration and flnancial crowding out His mode departs from previous work on income
lax Integraticn and for that matler from most 01 the CGIE literalure for tax policy evaluation in
several direcllons It encompasses an endogenous sequential equilibrium struelure founded on
dynamic behavior with flexible expectations Government deficits are optlmaUy determined
Investment decisions are tward looking and the resull of optimizing behavior Several financial
assetsmiddot public and privatemiddot are considered
Simulation resulls Indicale that the welfare gains from Integration under large deficits are at
32
best lJlodest when measured in terms of the GNP The elimination of the corporate tax and its I replacemenl by increased income lax rates yields long run benefits which are never larger than
2 of the presenl yalue ct telr consumption and leisure onder the previous tax regime and 1
under the current tax regime However the short run w~lare effects of full integration tend 10 be
very low and in in some scenarios even negative This is a flew intertemporar pattern of efficiency Ieffects which reflects an adjJsiment lag in the interindustry investment decisions due to the
existence of costs of adjustment On the of her hand partial integration achieved by excluding I dividends from the corporale lax base syslematically yields negalive efficlencyeffeclS In Ihe long Inm these can be as high as 3 of present value of the future value of consumption and leisure This
is a new second best effect suggesllng Ihat less than complete integration may have perverse I efficiency effects
The dynamic slructure of PerclrBs model snowing for an improved freatment of real investment
declsions and government deficts provides a potential setting for a more accurate measure of the
costs and benefits of inlegration The simulation resullS suggest lower benefits and an intertempora Ibull
pattern of growing benefits
Simulation results also clearly suggest robust negative effects from partial integration How
reliable is this result If dJvidends were deductible from the corporate lax base (as are interest
payments) the preferenlial Ireatmant of debl over equily would be ellminaled Corporallons should
be expecled 10 react by decreasing the optimal deWeqully rallo However corporale financial
gecisions are exogenously set In Pereiras model as in aft of the other models surveyed that go as far
as dealing with the issue
Also withdrawing dividends from Ihe corporate tax base is a way of eliminating Ihe double
taxation of dividends at belh Ihe corporate and personnal Income levels How will the optimal
aliocellon of household saving accomodate that change in Ihe relalive return to corporale equity
39 I
t However hOusehold portfolio decisions are exogenouly set in Pereiras model as in all of the other
models surveyeltj go as far as dealing with the issue
IIt is sate to say that the evaluation 01 the benefits 01 partial Integration as defined above are
sevrey underestima1eO Meanwhile the distortionSoenerated by the increase in the marginal
personal tax rates designed to ralse the revenue foregone by the dividend exclusion are fully
aCCltlunted for A good measure of the costs of integration together With a poor evaluation of the
benefils may very well be the reason why partial integration is Simulated 10 yield neg alive
eHikiency gains
On a different vein as most of Ihe model surveyed here Pereiras is a closed economy model It
is legitimate to wonder how the resullS would change If international capital flows were anowed
IThis almost certainly would affect financial crowding out since a good d~al of the capital inflows
could be used to finance public debt
For these various reasons the results should be treated as preliminary but they strongly I suggosl that the dynamic structure provides valuable new insights into the corporate tax integration SSU2
40
7 Concluding remarks I I I i
What has been acccmplished Ihis far The success of Ihe CGE research in laelding the challenge
of dynamic modelling can not be denied
I iGreat progress in the modelling of household behavior has been achieved Promising
developm~nts in the modeUing of production and government behavior have also been accomplished
Interesting innovations in the concept of equilibrium and computation techniques were discussed
The policy analysis was greatly enriched by considering transitional effects together with longer
run steadymiddotstate equilibrium paths generalized equal yield strategies accomodating different
financial crowding out impacts and generalized dynamic policy evaluation indicators
An important set of issues have been addresed by the models surveyed Traditional issues
focusing on the effects of capital taxation and consumption taxes have been pursued In terms of
capital taxation for example the intertemporal nature of the issue was enhanced by allowing an
optimal evolution of capital stock in the economy and forward looking optimal investment decisions
In Andersson (1987) GouldermiddotSummers (1987) and Pereira (1986b 1987d) this is coupled
with a improved treatment of several tax provisions like investment tax credits and depreciation
allowances In terms of the consumption taxes developments in the specification of Intertemporal
household behavior the introduction of overlapping generations and the modelling of
intergenerational links via bequest motives as well as a closer attention to demographic evolution
provided a much improved economic setting for the understanding of the several aspects of the
problem
Furthermore dynamic modelling has permitted the addressing of a set of new issues Policy
Issues ranging from dynamic tax policy analysis to the evaluation of exogenous changes in
government expenditures to the impact of different methods of financing government expenditure to
4 1
the importance of tinanciai croHrii1g out to 1he relevance of consumorS expectations in tne policy
results have been addressed
Despile all of the progress and in part due to such progress several avenues are wide open for
much neded additional research The following seem to be the most Interesting and promising
areasmiddot First the specincation of liquidity constraints to household behavior may help to obtain a
better understanding of policy chenges that affect directly the interest rates in the economy Second
major attention needs to be paid to the incorporafion of a roraign sector (with open capital markets)
Third some efforts are needed in terms of finding adequate computation techniques now that
dynamic modelling eally makes the curse of dimensionality worse than eVOrbull
The modelling of financial markets is the single most unsatisfactory speet of the dynamic
models surveyed here~ The problems associated with provlding a mo~ adequate treatment of
financial markets and in general endogenous and optimal financial deci$ions bull both at the corporate
and at the personal levelsmiddot have 10 be addressed That necessarily requires Introducing uncertainty
into the models To b adequate this must go well beyond the mere assumption of point expeetalions
with the whole array of modelling implementalion and evaluation problems thai generates
bull I i
42
REFERENCES
Abel A and O Blanchard 1983 An Intertemporal Model of Saving and Interest Econometrica
51 pp 675-92
~ Andersson K 1987 Tar-ation of Capital in Sweden - A General Equilibrium Model Ph D
Dissenation UnIversity of Lund Sweder
Auerbach A L KotlikoH 1983 National Savings Economic Welfare and the Structure of
Taxation in M Feldstein ed Behavioral Simulation Methods in Tax Policy- Analysis Chicago
University of Chicago Press
Auerbach A l Kotlikoff 1984 Social Security and the Economics of Demographic-
Transition in H Aaron and G Burtless eds Retirement and Economic Behavior Washington DC
Brookings Institution
Auerbach A L Kotlikoff 1987 Dynamic Fiscal Policy Cambridge Cambridge University
Press
Auerbach A L Kotlikof and J Skinner 1983 The Efficiency Gains from Dynamic Tax
Reform Internatiooaj Economic Reyiew 24 pp 81middot100
Ballard C 1983 Evaluation of the Consumption Tax with Dynamic General Equilibrium
Models Ph D Dissertation Stanford University
Ballard C 1987 Tax Policy and Consumer Foresight A General Equilibrium Simulation
Study Economic Inquiry 25 pp 267-284
Ballard C and L Goulder 1985 Consumption Taxes ForeSight and Welfare a Computatonal
General Equilibrium Analysis in J Piggott and J Whalley eds New Deyelooments in Apolied
General Equilibrium Analysis Cambridge Cambridge University Press
Ballard C D Fullerton J Shoven and J Whalley 1985 A General Equilibrium Model for
t 43
Tax Eoliey Eyalua~iQO NBER University or Chicago Press
Borges A (1987) bull A Survey of Compula1ional General Equilibnum Models CECD Economic
Slud~ 8
Bovenbig L 19B3 imperfect Capital Mobility in a Perfect Foresight Herberger Model
Mimea UC Br~eley
Bovenberg L 1984 Capitel Accumulation and Capital Immobility Qmiddottheory In a Dynamic General Equilibrium Framework PhD dissertation University of California at Berkeley t
Bovenberg L 19850 Dynamic General Equilibrium Tax Models with Adjustment Costs in A iManne ed bull 1985 EtQnomrc ECUjilbrium Model Formulation and Solution Mathematical
lProgramming Siudy 23 Arnsterdam NorthmiddotHoliand
Bovenberg L t 1986 Capital Income Taxation in Growing Open Eeondmies Journa of Pubtic
Eamp(omlcs 3t pp 347middot376 IBovenberg L ald W Keller 1981 Dynamics and Nonlinearlties in Applied General
Equilibrium Models Internal Report Mlmiddot81middot208 Department for Statistical Methods CBS bull
Vooiburg The Netherlands
Charnley C 1981 The Welfare Cost of C~pltal income Taxation In a Growing Economy
IJournal of Political ECQnQm~ 89 3 pp 468middot496 I
Charnley C 19810 Efficient Tax Reform in a Dynamic Model of General Equilibrium I Mimeo Yale University I Charnley C 1982 On the Optimal Taxation In Stylized Models 01 Inlartemporal General
Equilibrium Mimeo Yale University I Decaluw B and A Manens 1985 Pays en Oeveloppement at Modeles Calculables OEquilibre
General - Une Revue de al litterature empirique Monographie Sene 9 Volume 3 Centre de
Recherche et Developpemont Economlque Unrslt de Montreal
44
bull
I I
I
I
1
Erlich $ V Ginsburgh and L van der Heyden 1987 Wher~ do Real Wage Policies Lead
Belgium bull A General Equilibrium Analysis European Economic Reyiew (forthcoming)
air R and J Taylor 1983 Solution and Maximum Likelihood Estimation of Dynamic
Nonlinear Ratioant Expeclations Models Ecooometrica 51 pp 1169middot1186
Feltenstein A 1984 Money and Bonds In a Disaggregated Ec~nonYin Searl 1-1 J Shoven
Eds Aoolied ~eoeral EqJiliDCiulD tioalYsis Cambridge University Press Cambridge
Feltenstein A 19B6 A Dynamic General Equllibrlu Analysis of Financial Crowding Out
Theory with an Application to Australia Journal of Public Economics 31 No1
Fullerlon D R Gorden 1983 A Reexamination of Tax Distorllons in General Equilibrium
Models in M Feldstein Ed Behayora Simulation Methods in Tax pOlicy Analysis Chicago
University of Chipgo Press
Fvllerlon D J Shaven and J Whalley 1983 Replacing the US Income Tax with a
Progressive Consumption Tax JoulOal of Public ECQnomics 20 pp3middot23
Fullerton D Y Henderson J Shaven 1984 A COmparision of Methodologies in Empirical
General Equilibrium Models of Taxation in Searl H J Shoven Eds ~pQlied Geoaral Jguilibrium
lIoalyss Cambridge University Press Cambridge
Fullerton 0 A King J Shaven and J Whaney 1981 Corporate Tax Integration in the United
Statesmiddot a General Equiiibdufi1 Approach American Economic Review 71 pp~ 677~691
Galper H R Lucke E Tader 1985 Taxation portlor Choice and the Allocation of Capital A
General Equilibrium Approach Mimeo
Gill p W Murray M Saunders M Wright 1986 Users Guide for NPSOL (Version 40) A
Fortran Package for Nonlinear Programming T echnicsl Report SOL 86middot2 Deparlment of
Operations Research Stanford University
Glnsbur~h V and L van der Heyden 1985 General Equilibrium with Wage Rigidities An
45
AppJication to Belgium in A Manne ed t Economic EQumprit1m~ Made Formulation and $oJioo
Mathematical Programming Study 23 Amsterdam NorthmiddotHolland ~
Goulet J 1968 1 Adjustrent CO$1S in the Theory o(the Firm Reyiew of EconomiC Studies 35 bull
pp47middot55
Goulder L 1985 Tax-Financed versus BondmiddotFinancedGhanges in Governent Spending
bull Mimeo Harva(d UniversilYmiddot
Goulder l J Shaven and J Whalley 1983 Domestic Tax Policy and the Foreign Sector The
importance of altemative foreign policy formulations to results from a general equilibrium tax
analysis model In ~n methods in Ibe antico of income from caoia1 ed Martin
Feldstein Chicago University of Chicago riSS
Goulder l L Summers 1987 Tax Policy Asset Prices and 3rowtly ~
A General Equilibrium Analysis NBER Working Pper No 212B
Grandmont J 1982 Temporary General Equilibrium Theory Ch 20 in Handbook 01
Mathemalical Eccnornics Vol II NorthmiddotHoliand
Harberger A 1959 The Corporate Income Tax An empirical Appraisal In lax Revision
Comoendjum Voll House Cornmitee on Ways and Means Washington DC Govemment Printing
Office
Harberger A 1962 The Incidence of the Corporate Income Tax JOULn1 of Political
Economy 70 pp 215middot240
Harberger A 1964 Taxation Resource Aliocalrlon and Welfare fn lb Role of Dir bullbulll and
Indjresl Taxes io tbe Fede[al Reserve s~I~Ill Princeton Princeton Unlveresity Press
James J 1985 New Developments in the Applications of General Equilibrium models to
Economic History in Piggott aand Whalley Eds
Jorgenson D and K Yun 1984 Tax Policy and Capital Allocation Harvard Inslitute of
bull
I
I f
Economic Research WP 1107
Judd K 1965 ShonmiddotRun Analysis of Fiscal Policy in a Simple Perfect Foresight Model
Journal of poJiHcal 5CCJiW1l pp 298middot319 r
Upton D J Potbc J Sachsa nd L Summers 1982 Multiple Shooting in Rational
Expectations Models fuoometrlc 50 1329middot1333
Manne A ed 1985 fconQrnjc Eoullibrium Model FOrIDujaljQo and Solution Mathematical
Pogrammlng Study 23 Amsterdam NorthmiddotHolland
Manne A 1985 Or the Formulation and Solution 01 Economic Equilibrium Models In A
Manne ed ECQOQlli-Jr~ibmiddotum ~tQdel EQlJulailon and Solution Malhematical Programming
Study 23 Amsterdam NorthmiddotHolland
Mansur A and J Whalley 1984 Numerical Specification of Applied General Equilibrium
Models in Scarf H J Shoven Eds Aoplied General Eouillbrium AnaiysectIsect Cambridge Universily
Press Cambridge
Merrill 0 1972 Applications and Extensions of an Algorithm thaI Computes FixedmiddotPoints of
Cenain Upper Semimiddotcominuou Point to Set Mappings PhD Disserlation Unlvel$ity 01 Michigan
Negishi T 1960 Welfar Economics and Exislence I an Equnlbrlum for a Competitive
Economy MelroeCOOQ11ka 12 92middot97
Pereira A 1985 On the Existence and Uniqueness of Optimal Ou1put Path lor CRTS Firms in
Adjustment Costs Technologies Mimeo Stanford University
Pereira A 1985a Consumer Expectations and Ihe COSI 01 UllUty In an Intertemporat
Framework Mlmeo Stanford University
Pereira A 1985b Corporale Tax Integrallon and Government Deficits A Dynamic General
Equilibrium Anaiysls Mlmeo Stanford University
Pereir A 1987 On Ihe Computation of the Users Costs of Stocks Ecooomi (forthcoming)
47
Pereira A bull 1987b Equal Yield Alternatives and Government Oeficlts Mimeof Stanford
University
Pereira A 1987c A Dynamic Applied General ~uHlbrium Model for Tax Policy Evaluation
Comlletamp Docurnent2tion Mimeo University of California San Diego (in progress)
Pereira A 1987d Should the investment Tax Credit be Rei(ltroduced Mlmeo University
of California San Diego (ir progreSs)
Piggott J and J Whalley eds 1985 tJew QeyeQpments in Applied General EQuilibrium
bullenalysjs Cambridge Cambridge University Press
Preckel Pbull 1985 Alternative algorithms for computing economic equilibria In A Manne Ed
Radnor R 19amp3 Equllibrium under Uncertainty In K Arrow M Intrilligaor ods
I ibull
Robinson S 1 g86 Mullisectorai Models of Oeveloping CountrieS a Survey Oivision of
Agricullure and Resources UC Borkeley WP 401
Scarf H 1967 The approximation of fixed points of a continuous mapping ~MM Jouwal of
epplid Mathornallcs 15 pp 1328middot43
Scarf H with T Hansen 1973 00 the Computation of EconomiC Eguilibria New Haven Yale
University Press
Scarf H J Shoven Eds 1984 61l1llied Genera Elujlibrium 6nalyssect Cambridge University
~ Press Cambridge
Shoven J 1976 The Incidence and Efficiency Effects of Taxe on Income from Capital
Jurnal of ramie1 Economy 84 pp 1261-83
Shoven J 1983 Applied General Equilibrium Tax Modemng IMF Slaff PaoerS 30
Shoven J and LB Simon 1986 Share Repurchases and Acquisitions An Analysis of which
Firms Participate NBER Working Paper No 2243
bull
48
IShaven J and J Whalley 1972 ft General Equmbrium Calculation of lhe Effects of
Differential Taxation of Incoma from Capital in the US Jauroe of public Econgmics 1
pp281middot321 (
Shoven J bull and J Whalley 1973 General Equilibrium with Tax$s a computation procedure
and an existence pr00f ReviEhi of Economic Studies 40 pp 475middot490
Shoven J J Whalley 1977 Equal Yield Tax Alternatives A General Equilibrium
Computa~ui0nal Tfchnique J~HJrnal of PUblic EconomIcs 8 pp 211~224 bull
Shaven J J Whalley 1984 Applied GeneralmiddotEquilibrlum MOdels of Taxalion and
International Tr8de An Introduction and Survey Journal of Economic literature 23t pp
1007middot1051
Slemrod J 1980 A General Equilibrium Model of Capital Income Taxation Ph D bullDissertation Harvard Univnrsity
Slemrod J 1983 A General Equilibrium Model of Taxation with Endogenous Financial
Behavior in M Feldstein edbull Behavioral Simulation Methods In Tax Policy Analysis Chicago
UniVerSity of Chicago Press
Stone J 1985 Sequential Optimization and Complementarily Techniques for computing
Economic equilibrium in A Manne Ed
Summers L 1981 Taxation and Corporate Investment a q-Theory Approach BroQkinas
Summers L 1985 Taxation and the Size and ComposHion 01 the Capital Stock An Asset Price
Approach Mlmeo Harvard University
Whalley J 1975 A Geneal Equiloibrium Assessment of the 1973 United Kingdom tax
reform Economica 42 pp 139middot61
1JLTIllOS JIQll~)ltG PAPERS PUBLICADOS bull
I
I I
j
112 70 - LUCENA iJiCIIJo ~ IIEnvironrilent Honitoring and Orqanization Structure T (Setellibro 1987)
nI 71 - L~ENA Diogo A Note on the Representation of Information Stxueture (SeteJObro 1987)
nQ 72 - LUCENA D1CgO Envuorment 111)nitoring and Organization
Structure II (Setelilhro 1987)
ril 73 - AlmmE3 Antonio Pais e CAsPAR Yitor bull Tributacao Inoentivos e InvesUlllentos Analise Qualitativa noYellibro 19S7)
n2 74 - liAlltOSA rose Pedro Opt1lllal lIage R1gidity A SUggestedtlthodology to Teet tb Theory and an Application (Outubro 1987)
nI 75 - CABRAL L 11 B Three lIotes on SYJIIlletric Gemes witt itsYJIIlletric Equilibri (NoveJObro 1987)
n2 76 - LUCENlt D10gOTo Search or Not to Search (Fevereiro1967)
nQ 17 - SA Jorge asconcelos euroI A Theory of SynergyB (Fevereiro1987)
n2 76 - VILARES tanuei Jose Os Be IntenaMios IlIIportados ColIIO Factor de Produao (Julho 1987)
nQ 79 - SA J~rge VasCQncel03 e HoY to Copete and Co~unicate in tature Industrial Products_ (aio 1988)
n2 80 - ROIl Rafael Learning and Capacity Expansion in a New larket Under Uncertainty (Fevereiro 1988)
n2 81 PEREIRA Alfredo llarvilo Survey of Dynamic Computational General EquilibriUlll Hodel for Tax Policy Evaluation (Outubro 1987)
Qualquer in ormacao sabre o 1IIorlltlng Papers ja publlcados sera prestada pelo Secretariado de Apoio aos Docentes podendo os ~e3OSf
ser adquirido5 ne Secyo de Vlndas da Faculdade de EconollLia UNL na Travessa Esteviio Pinto Cltpolide - 1000 LISllOA
1 1
2 Brief overview of non-dynamic CGE models
The origin of most or the models we survey goes back to the work of Scarf (1967 1973) whlgt
first developed a reliablo algorithm to compute equilibrium price for a ArrowmiddotDebreu economy
HiS algorithm used simpicial Bubdivision techniques and can -be shown to be the computational analog
of the fixed point theorems previously used to prove the existence of equilibrium His technique
could soW a model with an arbitrary number of consumers and commodities as long as all agents
were price takers consumers were subject to budget constraints demands were continuous andmiddot
production did not display increasing returns to scale The alQorithm while guaranteed to converge I
I I I
was relatively slow for prcb1ems involving more than S1If 20 dimensions A maior improvement in
computational speed Woo offered by Merrills algorithm (1972) which used tpe same fundamental
ideagt as Scarfs procedure
Scarfs model (as is true for the standard Arrow-Debreu model) does not include a government
sector ~ neither taxes nor public goocs At one of the mosi promising appHciUions of the new
computational techniqu w~s in Iha area of tax policy evaluation Shoven and Whalley (1973)
exlended the general equilibrium model and computational approach to include a wide rry of taxes
and a governl11ampnt spending plan The original ShavenmiddotWhalley model was static (although the
different commodi~ies CQuid be considered simitar goods available at different datest as In
Arrow-Debrau) and government was aumad to run a balanced budget The data are arranged as a
social accounling matrix The models spacification and calibration is checked by solving it in the
presence of the base set of laxes The result should ba exactly tha initial social accounting matrix
After having pasd this replication check the model is solved for a counterlactua equilibrium In
the presence of a new tux design The result is once again a social accounting matrix The two
equilibria are compared in order to tSS($S the impact of the new tax plan The first uses ot this
bull
1 Z
model for lax policy evauajon were ShovenWhalley (1972) Whalley (1975) and Shoven
(1916)
Completely static models as Shaven-Whalleys are unsatisfactory for many tax reform ISsues
These include corporate lax Integration effects of investment tax credits effects of accelerated
depreciations consumer Or e)penditure tzxes~ importance of saving subsidies like IRAs etc These
are essentialy dynamiC ksues They involve not only the allocation of capital across sectQfsbut
perhaps more Important the capital intensity of the economy But the capital accumulation and
capital reallocation take time and may involve adjustment costs Because of these issues Ballardmiddot
Fulierton~Shoven~Whaloy took the first steps towards developing a dynamic model
In this conlexl the ShovenWhalley model has been extended and implemented for Ihe US
-economy In BuIIBrdmiddotFulienonmiddotShovenmiddotWhalley (1985) While Ihis book completely documents the
mo~el it was used in several publicalions beginning in 1978 Their mode) consists of nineteem
production sectors twelvo households and fifteen consumer goods It includes a very detailed set of
taxes including the federal and state personallncome taxes federal and state corporate income taxes
SOCial Security taxes pro1J0rty taxes unemployment Insurance excise taxes sales taxes etc It
has been calibl att-d to reproduce the 1973 US economy Recentty a version corresponding to the
1983 JS economy has been developed
wThe 8alard-Fullenon-Shoven Whalley model is dynamic In the sense that consumers face a
choice between current consumption and leisure versus future consumption (which can be
purchased via s wlngs) The consumer classes act as if they were maximizing a nested CES utility
function over the domain of a CES aggregate of contempo-raneous consumer goods and leisure and
future consumption subject to their income constraint The parameters of those funcHons
determine the shares of income devo1ed to each commodity to saving and to the purchase of leisure
They also determine the two key elasticities in the models ~ the elasticity of labor supply with
bull
13
respecl to tgte rbullbull1 aoer lax wage rae and the elasticity of saving with respecl 10 the real after lax- rate of retum to capital
In the model consumerS have myopic expectations regarding lulure prices and in particular
regarding the future rale of return 10 capilal Ballard (1983) and BallardmiddotGoulder (1985)
inco~ora1e both perfect foresight and limited foresight into this model Future consumption is
acquired by buying a fixed composition portfolio of real investments that offer an infinite annuity
of returns
The production side of Ihe model is compielely static The mod1 Incorporates a constantmiddot
olasicily 01 substitution betweeen prlmary inputs in production (capital and labor) and fixed I coefficients for intermediate inputs The model distinguIshes between Industrial outputs and
eonsumer goods for the simple reason Ihat tile data are classified differenUy Industrial sectol$ I I involve such categories as forestry and fisherle metal mino9 and publishing and printing while
consumers purchas-3 fUfHi1ure 2ulomoblles and books This fact is recognized in the model by Ibull
incorporatirg a secord stage Of production which converts industrial outputs intO consumer goods
This technology is uStally mod9-led as a flxed~coefficient conversion matrix
The BallardmiddotFINton-ShovenWhalley model assumeS that the private sector finance
marginai ioveslment with the same composition of debt and equity as currently exists in each sector
This is the same arsumrtion that Herberger originally used Investors all hold debt and equity in tile
same proportion Thampreiore this ownership can be aggregated into simply capital ownership For
tax purposes however the separate treatment of debt and equity is taken into account at both the
corporate and persofa Io=vel SimllarlYJ the dividend poIicies on corporale equitY are established
exogenously There are no government bonds in the model since there are no government deficits
The modol is solved for a sequence (as many as 100) of temporary equilibria wltll consumers
allocating income between present and future conumption at each point in time The palh for 1he
14
economy ~ a sel of connected equilibria The connectIon is pra~ided by capital accumulation
Capital accumulation is endogenous and determined by saving The model stans with a sociagt
accOunting mi1rix In the bzse case the economy is assumed to be in a steaqy state growth -path (along
which all relative prices ~fe constant) The madel solves for both the naw steady state growth path
and the transltion 10 it zf~er a policy intervention rhe authors have frequently addressed the
question-of howong rt tKes to effectively settle un10 a new steady S1ate growth path
The dynamics of the model are limited however in that future consumption is collapsed into a
composite commodity Aso the absence of government deficits and the lack ofproductlon dynamics
limits the realism of the BallardmiddotFuliertonmiddotShoven-Whatley model for the analysis of dynamic
policy issues (such as the adoption 01 a consumption tax or Ihe elimination of the investment tax
credit) The models we survey try to Improve upon Ihe Ballard-FlIiertonmiddotShaven-Whalley
dynamic modelling The models Jn~orporate more forward lookTng behavior They improve the
dynamics on the conampumrtkm side of the economy and some of them introduce true 1ynamlc behavior
on the part oI the producers We turn now to the issues inY9lv_ed in developing such dynamic models
15 I I ~3 ~ssues in the modelling Of dynamic behavior
We will 100k firt at the speclficalion of economic behavior of consume(s~ producers
goyemment and Ihe r~sl of Ihe world The modemog of corporale and household financial decisions
I is then addr5sed Finally Ihe eencept of equilibrium is discussed
I
i
I 1
31 Consumers behavior
Early efforts 10 build dynamic features into the economic behavior 01 consumers are due to
Ballard (1963) and Auerbach-Kotllkoff (1983 1984) Their work has been closely followed by
several subsequent authors In faet dynamic household behavior Is the single most pervasive aspect
amohg the models surveyed
All the models Inc-O1orale some loro of lila-cycle behavior Household behavior is dewmloed
by the maximization of an addWVely separable time invariant intertemporal utillty function The
utility lune1ion is defined oyer the domain of the consumption goods In the economy In most of the
models leisure is also an argument in utility so that labor supply is optimally determined
Utility maximi4ation is subject to a lifetime intertemporal budget constraint which equalizes
the presen1 value of consumers income and expenditure More reeentiy in Andersson (1987) bull
Bovenberg (1985) and Pereira (1986b 1987d) the constraint i$ defined as a sequence 01
recursive equations of motion on wealth This has the potential advantage 01 accomodating liquidity
constraints However It should be recognized that in the absence 01 liquidity constraints (ie when
consumers are free to completely borrow against future income) the two specificallons of the
househofd ccnstraint are essentiay equivatent Furthermore tn both versions saving is optimally
determined as a way of translaring wealth intertemporally
1 6
Some lodels now have a sophisticated dynamic specification of hoosehokl In Ballard (1983) -
BallardmiddotGoulder (1985) and Auerbach-Kotlikof (1983 1984 1987) consumers behavio is
embedded in an iterQEneraional setting Each of them has S5 age cohorts simulataneously alive
Som-e measure of jntergener1ional altruism is considered in the form of bequest motives In Ballard
(1983) and Ballard-Gouldr (1985) consumers derive utility directly from bequthing parl of
their wealth Therefore bequests assume 1he form of a scrap value of terminal wealth of a
generation In AuerbachmiddotKotlikof (1983 1984 1987) each generation empathyz bullbull with Ihe
utility of future generations ovar bequethd wealth In addition Ballard (1ge3) includes detailed
demographic poiections wfill the intent of incorporating in the policy analysis the effects of the
bull Post World War tI baby-boom
In the real world household decisions also include the optimat allo~atiOn of saving among
shyalternative physical or financial assets Theories of household portfolio behavior are relatively
Iwen~establshed However they involve uncertainty as a crucial element Accordingly they have
middotnot been incorpora~ed in the context of the determinIstic dynamic models we sutyey here We will 1 Come backt-O this Issue below
32 Producer behavior
The efforts to build dynamic features into the economic behavior of producers are more recellt
and less widely adopted The first attempts are due to Bovnberg (1984 1985) and Summers
(1965) Dynamic behavior has been more fully incorporated In the recent models of Andersson
(1987) Auerbach-Kotlikoff (19B7) GouldermiddotSummers (1987) and Pereira (1986b 1987d)
Part of the reason why production side dinamlcs has been more slowty adopted is the weak
supply of accepted theories referring to the dynamic behavior of the firms In the models referred
I 17
to above gynamlc producion and investment b~havior are middotinduced by the existence of capital
adjustment costs and linked to Tobins q theory Adjuslment costs are designed to capture both the
incomplete mobility of capital aoross industries and installation costs ~e the costs of adjusting
capital towards its cpHmai level)
Adjllstment costs can be conceived as internal to the firm and measured in terms of foregone
outpul along the lines of Lucas (1967) Alternatively adjustment costs can be viewed as actual
external cests incurred together with Ihe purchase costs along the Unes of Gould (1969) With the
exception of Pereira (198Gb 1937d) all of the CGE authors follow the former approach
The firms in the economy maximize their market value as the present discountad value of the
future stream of dividends In Andersson (1981) and Pereira (198Gb 1987d) firms are seen as
maximizing the present discounted value of net cash flow Maximization is constrained by Ihe
adjustmenl cosl lechnology and ~n equation of motion describing Ihe evolution of the capital stock
It should be noted that undereogenous divldendrelention rules the two problems are equivalent II
should also be nOled Ihat a salisfectory economic rational for th bullbullistoo of dMdonds with the
present tax code is missing in the profession (Shoven (1986))
The models with static formulation of producers behavior are characterized by passive
investment behavior Investment merely accomodates to saving in the eccnomy With a dynamic
formulation induced b adjustment costs real investmen decisions are forward looking Investmenl
Is endogenously and optimally determined by the firms A fundamental difference between the
shonmiddotrun in which capital stock is given and longmiddotrun in which Ihe level of capital is allowed to be
optimally determined is emphasized
This extra richness of produc1ion dynamics is not without costs A careful look at the summary
tables will clearly show an inverse relation between the adoptlon of dynamic features in production
and the level of disagreggation of the production side of the economy In fact with production
bull
j
1 6
dynamics $e dimension of the problems is immensly increased Let IS be more specific
In a static framework with constant returns to scale production technology the output level ismiddot
indeterminate The optimal aliocation of inputs can be obtained by cost minimization with any
feasible output level generatillg zero profits Such zero profit conditions are used to solve for the
output prices in terms of the factor prices and hencemiddot to reduce the dlmensionality of the problem
from the -number of commodities and factors to the number of factors Thus even if the model deals
with 30 production sectors the computation of an economic equilibrium can take place using only
the dimensionality of the primary inputs in the economy
Now under certain regularity conditions on the production and adjustment costs technologies
leading to enough concavity of the optimality objective the intertemporal output path for the firm is
endogenously optimally and uniquely determined even with constant relums to scale technologies
(see Pereira (1985 1987a) Accordin~y with adjustment costs in the model optimal profits
will in general be non-zero This result has crucial implications for the computatiQn of equilibrium
in the model~ with production dynamics With adjustment costs no reduction of dimensionality is
possible The curse of dimensionality returns as a binding constraint
The introduction of adjustment costs adds another significant complication to the model With
capital being less that perfectly mobile in the economy different rates of return on capital will
exist in different sectors This is a difficult problem to tackle conceptually in the absence of
uncertainty Also in the real world producer maximization choices include also its choice of
financial ratios (debVequity) and payout rates (dividendretained earnings) These financial
subjects are difficult and much Clf this behavior is still taken as exogenous by the modellers We
will come back to these issues below
middot i
19
bull 33 government behavior
middot
Two issues dominate the modelling of government behavIor First govemment behavior has
Ibeen typically seen in Ihe CGE liIerature for lax poliCY evaluation as constrained by yearly balanced
I budsets (see for example 8allard-Fullrton-S~ovn-Whafley 098Sraquo The analysis of
govemment defidlS and public debl in a CelE context requires a dynamic selling Second the level of
government expenditures is either exegenouty given as in Auerbach-KoUikofi (1984 1987)
6ovenbrg (1984 1985) Feltenstein (1984 1986) and Jorgenson-Yun (1984) or
endogenously (but no optimally) determined by the balanced budget conditions as in
6allard-Fullerton-Shoven-Whalley (1985) Andersson (1987) Erllch-Ginshurgh-Heyden
~
(1987) Gouder (19aS) and Gouder-Summers (1987) In the second case the composition of
bull public expenditures is often optimally determined However the leve~ of government expendi1ures
can only 00 endogenously and optimally determined if the government is seen as an optimizing agent
and is allowed to run deficits
The first attempts to deal wiU the government defiCits in CGE tax models are due to
Auerbach-Kollilltoff (1984) and Feensain (1984) In Auerbach-Kotlikof (1984) government
expenditures afe exogenous They grow at the rate of growth of population However given 1he tax
structure and tax revenues yearly deficits and surpluses are atTowed subject to an in1ertempond
constraint that the present value of future tax revenues equals the present value of future
expenditures
In Fellenstein (1984 1986) government expenditures are also exogenously given He also
allows the government to run deficits to finance expenditures in excess to 1ax revenues However
surpluses are returned to CQnsumars in the form cf transfers Accordingly government ts not
subject to any constraint regarding the future repayment of public debt
In GouJder (19851 the 9Vernmenl maximizes a static social welfare ~nctio subject to a-
bull
20 I i
I
I I I
I bull
balanced budget conSirtlL This follows the optimal alocatkm of government expenditures alpng
the lines of Baliard-FHeiton-Shoven-Whalley (1985) Ttle model Is generallzea to allow for
exogenous changes in HIe time path of government expenditures Two financing alternatives are
consjdeft~d and contr8s1ed EldditionaJ tax revenues and bond issuance
Pereira (198Gb 1137d) auempts to address both the incorporation of deficits and the
determination of govement expenditures The path of government expenditures and the path of
delicilSlsurpluses (ano therefore the path lor debt) are endogenously and optimally determined The
government is seen as mltJximizing n intertemporal social welfare function given the 13x structure
Optimization is subject to a sequence of recursive equations of motion reflecting the evolution of the
ptlbllc debt allowing fDr government budget imbalances It should be stressathat this specification
01 the government cOMtcalnt is equivalent to the specificaUon in Auerbaeh-Kotlikoff (1983 1987)
when no liquidity constraints affect -gove~nment behavior
The extra richness in the treatment of government behavior allows tfte examination of
government debt poices in a truly dynamic setting Also financial crowding out effects induced by
government deficits can be analyzed (see Auerbach-Kottlikoff (1987) Feltenstein (1986) and
Perer (198Gb 1987e))
34 Foreign sector
Most of the open economy models follow the assumptions of balanced trade with import and
export net demands characterized by constant elasticities along the line of Ballardmiddot
-Fullerton-Shaven-Whalley (1965) Such is the case of Salfard (1963) Ballard-Goulder
(19S5) and Goulder-Sllmmers (1987)_ In Feltenstin (1985) the rest of the world is treated as
21
an addjlion~ consumer group
Bovcoberg (1986) develops a model In which two economies lIr considered each following
intertempoml perlect foresight paths These economies meet in the international forum Their
trade relaroMhips are characterized by yearly balanced trade accounts
None of the new generation of dynamic cae tax models has yet IncorPorated the lnternational
capital flows as done in aoulder-Shoven-Whalley (1983) for the earlier Ballarrshy
-Fullertonmiddot Shaven-Whalley (1985) model This is unfortunate as there is an important
Intertmporal aspect to international lending The first al1empts along Ihese lines are due to
Andersson (1987) and Erlich-Glnsburgh-Heyden (1987) Andersson (1987) in his model of the
Swedish economy adopts a closed economy appraoch in which rates of return in the domesJic
economy are largely determined by the international capital markets Fidinterest rate induce
Intlnatlonal capital flows which determine and finance the International lrade imbalance In
lurn In ErlichGinsburghHeydenmiddot (1987) foreign trade Is generaled according to an
intertemporal Irade welfare lunction with constant import and export elasticities In the shortmiddotrun
they allow inlernational trade inbalances which generate capilal ftows to the domestic households In
the long run however trade balance is assumed
35 The need for financial markets
A dynamic economic Wucture not only provides the ideal environment 10 model many features of
economic behavior it also permits one to Incorporate the financial side of the economy If the
government is allowed to run deficits the question 01 deficll financing automatically follows If
investment is optimally dewmined and returns to capital are different aClOSS sectors problems of
investment financing arise If there are seeral final1cfal assets In the economymiddot government
4-
I i
ibull
I I I I
I
j
22
bonds privpte bonds and equity (or simply physical capital installed inmiddot different sectors) the
problem of allocation of saving among assets with potentially different returns arises
The papers surveyed var greatly with respect to l~e extent of their attention to the financial
side of the economy At Clne extreme are the models in ~anardmiddotFuliertonmiddotShoven-Whalley (1985)
Andersson (1987) Ballltld (1983) Ballard-Goulder (1985) Bovenberg (1985 1986) and
Erlich-Gjnsburgh-Heyden C19B which are devoted exclusively to the real side of the economy
In turn Auerbach-KotliKoff (1984 1987) Feltenstein (1984 1986) and Goulde (1985)
allow for government debt In these models saving finances changes in government debt and physical
capital Private and publlc 2ssets are perceived by the households as perfect substitutes The
allocation of saving merely adjusts to the relative demands for funds
Feltenstein (1984 1986) is the only model surveyed which introduces money Government deficits are financed by issuing money and bonds according to an exogenously given rule Money is
demanded by consumers for transaction motives and an exogenously given fraction as a store of value
On the other hand government bonds and physical capital are the vehicles for the intertemporal
transfer of wealth
Summers (1985) and GOlldermiddotSummers (1987) introduce a whole menu of financial assets shy
firm specific equity capital Different assets earn different rates of retum However such rates
are equal up to constant and exogenous sector speCific risk premia Therefore the introduction of
constant exogenous risk premia as helpful as it may be in the context of calibration does not solve
the main issue of the non-optimality of the allocation of saving Also talking about risk premia in a
deterministic context is somewhat unsatisfactory
Pereira (1986b 1987d) also introduces a whole menu of assets private and public bonds and
firm specific equity However all the assets are expected to yield the same rate of return and
therefore perc~jved as perfect substitutes The different asset types allow consideration of
23
exogenOU$~8tUequjty triO dividendlrc~enlion rufes and therefore seve-ral sources of investment
financing bond equity and retained earnings
The nonmiddotcptimalHy of the allocation of saving and the absence or exogenelty of corporate
financial rules is _a reftectlon of the limitations of the determInistic approach Either in a context of
perfect anticipation oi the future prices Of in general within the t$atm of point price
expectations all the papers follow a deterministic approach Under such circumstances consumers
either expect diHerent rales of return (inclusive of risk premium) across assets in which case
they wilt buy only one asset (that with highest rate) Of they expect equal rat~s of relurn in which
case Ihey are indifferent about the asset composition of Iheir portfoliO There is no way of
Iradingmiddotoff rales of return and risks to obtain an optimal interiof solution to the problem of the
allocation of saving
The most advanced contribution in the modelling of saving allocation in a CGE setting is due to
Slemrod (1980 1983) in the context of a static one-period model In his model consumers act
according to a two stage separable decision process They Hrsl decide on how much to save Then
they decide on the allocation of saving according to an Indirect uttlity function dependent on the rates
of ratum and variances offered by the different assets in the eoonomy The sourCe of riskiness in the
ampConory comes from an uncertain marginal product of capital On the other hand aside from
portfolio decisions the res of the economy is insulated from uncertainty
The most complete contribution in terms of the treatment of the corporate financial rules is
FulienonmiddotGordon (1983) In a variant 01 Ihe BailardmiddotFulienon-ShovenmiddotWhelley (1985) model
Fullerton and Gordon have capital intensity and optimal financial decisions Jointly deter~ined
through a two siage process The cosl of financing capital is minimized by trading off the tax
advantages of debt against the epected real bankrupcy coSls inherenl with high debVequity ratios
Given the optimal debtequity ratio the level of inveSlment Is chosen such that at Ihe margin the
24 l
bull
rerurn pn ~uity equas the return on bonds plus an exogenous risk premium
36 Market assumptions equilibrium and expectations
Virtually all of the papers arc characterized by Walrasian market clearing assumptions All
markets are perlect1y competi1ive Atomistic competition among agents 1s assumed even thoughmiddot only
a flnile number of agents is considered Virtually no mark~t disequilibria or price stickiness afe
considered The only exception is ErlichmiddotGinsburghmiddotHeyden (1987) bull In thei model of the Belgium
economy I the _wage rate is fixed in the short-run Therefore in the shOrlwrun desequilibrium in
the labor market will generate endogenous unemployment However in lh$ long-run all prices
Including the wage rata are Ilexible and accordingly ali markets ciear
Given the dynamic nature of behavlor in the economy markelmiddotcfearing prices in each period
depend on expectations of fture prices and on tax variables In the economy There are essentially
two ways at interpreting the economic equilibrium in such a dynamic conte~bull If future prices are
perfectly anticipated (io expectations are self-fulfilling) a perfect foresight equilibrium
prevaHs Then future actions are merely the implementation of current decisions for future
periods However jf ptica expecla110ns are not perfect (Le agents make mistakes w1th re$pect 10
future prices) then a temporary or shortrun equilibrium prevails Markets ciear and clearing
prices depend on fu1ure price expectations Current plans about the future are typically not
precisely implemented They will be revised as more or better information becomes available to the
economic agents See Grandmant (1982) for a comprehensive survey 01 the temporary equilibrium
literature
Wilh the exception of Gould (1985) and Pereira (19S6b 19S7d) all the models surveyed
adopt th~ concept of a perfect foresight equilibrium In turn BaliardmiddotGouldr (1965) considers a
J
2S
flexible am9unt of foresight In terms of the number of years over which price movements are ~
foreseen Pereiras model (1986b 1987b)) is flexible in a somewhat different way in that it can
indude any range of foresight from myopia to perfect foresight The choice between the perfct foresight and lemporary equilibrium is ultimately to be made on
philosophic grounds It can be argued that less than perfect expectations Imply that agents are
irralional in some way (see Auerbach-Kctiikoff (1987 p 10) However the reverse argument can
be made One can question wiether agents are reaHy rationar and perfectly knowledgable about
future prices
Recent evidene of Balird (1987) Ballard-Goulder (19851 Goulder (19851 and Pereira
(1986b 1987d) suggests that the choice in modemng expectations is an important one They show
that the degree of foresight inl0 thc future (ranging from perfect foresight to-myopic expectations)
may have dramatic impacts on the pollcy conclusions of Ihe model Accordingly the best research
strategy may b to design models which are flexible enough 10 allow for different rules regarding the
formation of expectations With the exception of the articles just mentioned sensitivity analysis
bullhave previously not been performed along this dimension
In terms of implementation the two concepts of equilibrium - perfect foresight and temporary
equilibrium have different implications The dimensionality of the equ1ibrium soll1ion algorIthm
is involved Suppose we have a model with ten markets to be run for a period of 50 years Aside
from normalization a perfect foresight model implies computing prices In 500 dimensions while a
-t~mporary equilibrium model requires sorving SO equilibria each in ten dimensions Given that
computational speed often varies with the cube of the number of dimensions the lemporary
equilibrium formulation is potentially strikingly more feasible However Ballard (1987) and
Goulder-Ballard (1985) have devoloped techniques to greatly speed the computation of a perfect
foresight equilibrium
26
j
I
4 Jrnplemenlalion and Issues on policy evaluallon
41 Calibration and equilibrium comparlslon
CGE models at typically parameterized by the use of a calibration procedure Some parameters
are exogenously given However some crucial parameters are determined in such a way that the
model replicates the data for a given base year See MansurmiddotWhalleY 11984) for an eXlensive
discussion of 1~i$ issue
CalibratIon in a dynamic context is generaliy interpreted as requiring two properties First
replication of base year data is required Second the model is parameterized to simulate an
intertemporal balanced growth path when Ihe base policy is maintaineq This Is the approach
followed by Ballard (1983) BallardmiddotGoulder (1985) Goulder (1985) and SummersmiddotGoulde
(1986) It follows the practice of BallardmiddotFuliertonmiddotShovenmiddotWhalley (1985) with their
sequential cquilibrium dynamics
Other authors Auerbach-Kotlikof (1983 1984 1987) Boyenberg (1984 1985 1986)
and Pereira (198Gb 1987d) follow whal we cali a qualitalive calibration The slructural
palametars are exogenously chosen so that lh economy follows a reasonable path into Ihe fulure
ThaI has 10 do wllh the fact thaI given the recursive nature 01 the dynamic economy and aside tom
such structural parameters only iniUa stock values are needed to run these models Given initial
conditions on the stocks of say private wealth capltal and government debt agents will optimize
and Ihereby generate a lirst round of net demands and equilibrium conditions In lurn the
equilibrium prices will determine the evolution of the stock variables into the next perIod
There are several potential problems with calibration in the conlext of dynamic models The
assumption of a steadymiddotstate growth ~ath in the base case can be questioned First while
27 I Ibullbull
steadymiddotstaU is a possibity it certainly is not the only meaningfull solution to dynamic models
Even in the case of a perfect foresight equilibrium the model implies an equilibrium path which i may or may not involve colanced growth The model not the modeller should dictate the nature 01 I themiddotbase case path Second in the context of a temporarY equilibrium path a steady state solutlcn is Inol a likely model QutcOn1tl In fact unlessmiddot expectations are stattc short run behavior consistent
with a steady-state evolution will in general not be generated On the other hand if static I expectations are self~fufjili[1g we have in fact a perfect foresight model Third even the base year
replication requirement may cause problems In Ihe cOntexl of temporary equilibrium Any
calibration parameter would be conditional on expectation rules which is probably an undesirable
feature
The nalur of the two-requirement calibration strategy is very muchin the spirit of the bull
traditional design of the comparision of alternative equilibria comparis_ion between a steady-state
base case on one hand and alternative paths including a transition period and a finat sleadyyenstate on
the other hand Pereira (198Gb 1987d) compare different (not necessarily steadystate)
equilibrium paths The arguments against such a procedure are based on the idea that the impact of
policy changes can be observed most easily since all departures from the steadymiddotstate can be
attributed to the altemalive polioy On the other hand the base case so dEfined as a steadymiddotstate is
consist~nl with previous work in a less dynamtc setting and Iherefore allows a common standard
for comparing model results
42 Computation algorithms
We are still al a stage in which basically each author uses a different computational technique
Th development of dynamic models - in parlicular wilh adjustment costs andior perfect foresight
bull
28
has corresponded with the decline in the use of fixed point algorithms In -fact given the relative
arge dimensions inevitably involved such algorithms lend to be very inefficient at the best and
chen prohibitively stow See Stone (1985) and Preckel (1965) for a comparative assessment of
different computation techniques Among the models surveyed only Ball1rdFulle(ton~
-ShavenmiddotWhalley (1985) and Feltensteln 1985 use Merrills variant of the fixed point
algorithm 1echnique
AuerbachmiddotKoHikoff (1963 1984 1967) follow a three stage procedure They first compute a
base case steady-state then a revised case steady state and finally transition path for the economy
berveen these two steady-states tn all stages a Gauss-Seidel iterative procedure is used
Ballard (1982) Ballard-Goulder (1965) Goulder (HiSS) and Summers-Goulder (1986)
~
use a method developed by Ballard and Goutder which is similar in many awects to the FairmiddotTaylQr bull
(1983) algorl1hm Short-run equilibria are calculated (using Merrills algorithm) parametric on
nrice expeC~lions The model is then iterated to generate self-fulfilling intertemporaf expectations
and the corresponding perfect foresight equilibrium In a relatively similar approach Andersson
1987 uses a simulation program SIMNON developed In the University of Lund This program
can handle two-paint boundary problems In a fashion consistent with the multiple shooting
algcrlthm (see Lipton-Poterba-Sachs-Eummers (1982))
Boyenbergs computational approach (1985 1985) differs from the other models surveyed in
that he relies heavily on analytical techniques Computations are done by using a dynamic version of
Johansons linearization method Being essentially determined by the continuous time nature of the
I I
f
model thiS linearization model has the disadvantage of confining the analysls to infinltesinal changes
around the base case equilibrium (see Bovenberg (1985) p 53) Pereira (1986b 1987d) uses an optimization algorithm NPSOL - developped by
rGillMurray-Saunders-Wrlght (1986)) This algorithm is used 10 compute the sequence of
I
29
Shortmiddotrun temporary equilibrium which makes up the intertemporal equilibrium path The
equilibrium conditions are seen as nonlinear equality constraints in the minimization of an artlcial
objective function The prices are normalized to the unit simplex by an additional linear equality I I 1 bull constraint
II ErlichmiddotGinsburgh-Heyden (1987) follow a unique approach in that they use a variant of the
optimization technique introduced by Negishl (1960) The economic equilibrium can be generated i
as a solution of a mathemallcal program Ihe objective function of which is a weighted sum of the
utility functions of the various agents while tho constraints set consists of the market dearing 1
conditions Ginsburgh-Heyden (1985) have extended Negishis resulllO tho case of downward price
lrigidities I
Finally the paper by Jorgonson-Yun (1984) is also unique in tIlat it is the only I
econometrically estimated model Different blocks for the consumption and production Side of the
-
model are separately estimated to provide the necessary structural parameters i
I The diversity of computation techniques is yet another indicator of the exploratory nature of the
body of literature surveyed in this article
43 Equal yield comparisions
The link between a base case and counteriactual simulations Is usually provided by the concept of
equal yield Shaven-Whaley (1977) discuss the meaning 01 equal yield in a general equilibrium
context when government is not allowed to run defICits Equal yield is interpreted to moan constant
public utility Government base case utility is maintained in the counferfactual experiments With
balanced budgets the concepl of equal yield is unambiguous The new equUibrium prices and the
balanced budge condition will del ermine the minimal expendilure and taxes needed to maintain base
30
case publi~ utility Accordingly in general equal yield is inconsistent with equal nominal tax
revenue Some change in tax revenue Is necessary Different tax replacement schemes ate
considered to assure that enough tax revenue is collected This is the approach essentially followed
by most 01 the papers surveyed Only Feltenstein (1985) Goulder (198) and Jorgenson-Yun
(1984) chose not to follow an equal yield strategy
The question is of how to Interprete lila concept of equal yield when the government is alowed to
fUn deflcLts The optimal leve of expenditure for base ease public trliUty can now be taAinanced
bond financed or financed by a mix of bonds and taxation We have several versions of equal yield In
particular equal yield may now be consistent with equal tax revenUE Furthermore some meaSure
ltgtf financial crowding out effects of government deficns can be inferred from the comparislon of the
several equal yield alternatives These issues are extensively discussed ~Pereira (1987b)
44 Price expectations and the dynamic generalization of compensation
Indicators
The sum of equivalent and compensation variations over households is the most widely used
aggregated measure of efficiency gains or losses In the context of $teady~state eomparisions andor
when complete future markets exist andlor perfect foresighl is assumed there are no difficulties
associated with the use of the standard Hicksian indicators In fact correct future prices are known
in these cases
If expectations are not selfmiddot fulfilling andlor future markels are not available a dynamic
generalization is necessary 8allardmiddotGoulder (1985) provide some steps in that direction by
defining an indicator that accomodates periods far into the future when households do not have
perfect foresight but a steady~state prevails On he other hand this issue is more fundamental in
i
[
I
i I I I I [
I I
bullI
31
the contex~ of a Gner1 lemporary equilibrium ftamawor~ In such Circumstances Pereira
(1986a) dev~lops a cnamlc generalization which is obtained as the present discounted value of a i i
sequence of short-fUn optimal expenditure functions consistent with a base case expected future I stream of IlWities I
32
5 Empirical evidenee from selected policy Issues
Dynamic tax models have generated several important results which escaped Sialic modellers
The following is a seleCpoundd set of issuos whIch stress the marginal benefits of dynam~ modelHng of
economic behavior in innovltlttive areas of analysis The discussion of a consumption tax emphasizes
the benefits of dynamic tusehold behavior and intergenerational aspects The study of the
flliminatlon or the re~intToduction of the investsnemt tax credits is made meaningful by the dynamic
modelling of production betlavor Modelling governmenl in a dynamic conlexts allows the modeller
to study the irrrpact of government deflcits Finally a dynamic framework lets us appreCiate the I
relevance of consumer expectations in terms of the evaluation of policy alternatives
i
i
51 Consumption Tax
t
By Its very nature a consumption tax can not be adequately investigated with a static model bull
Fullerion-Shoven-Whaicy (1983) use Ihe model of the US economy s described in
Ballard-Fullerton-Shoven-Whalley (1935) to evaluate the movement from the currenl US tax Isys1em to a progressive consumption tax Since their modeJ Incorporates a Jaborleisure choice
where leisure is an untaxed commodlty their results reflect the fact that both the consumption tax
and Ihe present system are dslortionary
Concentrating on the rntertemporaf distortions they show that sheltering more saving from lhe
current US income tax could improve econornic efficiency even if marginal tax rate increases are
necessary in order to maintain government fevenue~ At first you have a revenue shortfall
However the economy moves to a higher sustatned growth path and uitimatefy lower tax rates
cangenerate the same revenue path Also wages increase in the long run with this policy The
bull
bullbull ~
33 shyt
i
I
I
present value of welfare gi[lS for a potiey of complete savin9 deduction wilh marginal tate
adjustments (consumption lax) is simulated to be around $500 billion to $600 billion 01 1973
dollars
They also investiga1e th 1ngth of ttme it takes th~ economy to adjust to these policy changes
Roughly they estimate the long run to b thirty years although this figure Is very sensitive to the
specification of the savings elasticity - a crucial parameler in this model
52 Investment T~x Credit
Goulder-$ummers (1 S87) address the impact of eliminating the lnvestmenl Tax Cradit (ITC) on
intersectoral capital formation and on economic growth Their model isYa1ticularly adequate to
addies$ this issu~ in that is postulates a forward looking investment m9we with adjustment costs
They show that the eliminating the ITC causes a reduction in the rale of investment Inveslment is
estimated to fall by about 7 in the short-run and by about 12 in Ihe new long run steady state
In lurn a previous policy anncuncement lowers Ihe overall al1ractiveness of Investment and leads 10
a downward shift of the Investrrent profile On the other hand the combined effect of a revenue
neutral simultaneous efimiraron of the ITC and reduction of the corporate tax rates is a long run
reduction ef the capital Sieck by 35 This pattern suggests that a revenue neutral increament in
both Ihe corporale tax rates and the inveSlment lax credits would be preferable 10 the formula
adopted in Ihe US Tax Reform Act of 1986
Pereira (1987d) addresses the impact of the ITe from Ihe oland point of Ihe curTenllax syslem
under the Tax Reform Act of 1986 Given Ihe concern over the potential depressiVe effects upon
savings and investment of the new tax system in conjunction with deficit reduC1ion mechanisms of
the GrammmiddotRudman type a pertinent question is should Ih lTC be e-introduced In the context
bull
34
-of his mod~1 of the US economy Pereira focuses on the trade-off between the distortion in the
intersectoral allocation of capital induced by the lTC the lower tax revenues it generates and
potential financial crowding-out effects on one hand and the positive effects of lowering the relative
price of new capital goods on the other hand Preliminary results confirm the qualitative results in
Goulder-Summers (1987) suggestting a welfare gain of about 2 of the present value of GNP in
the best scenario of absence of replacement taxes Furthermore preliminary results show an
important time pattern to welfare gains with the average benefits increasing the further you look
into the future This pattern is due to the presence of constraints to intersectoral mobility of capital
and inertia in the adjustment towards the optimal capital levels as reflected by the presence of
adjustment costs
I
I
bullAuerbach-Katlikaff (1gB) in the context of their intergenerational model of the US economy
consider the impact on savings capital formation and interest rates of deficit policies and balanced
budget increases in government consumption They conclude that deficit finance and government
consumption can significa-ntly crowd out capital formation and lower the welfare of future
generations However crowding out from deficit finance is a very slow process because it results
from increased government spending over potentially long horizons Also deficit policies may
substantially influence the longmiddotrun interest rates while leaving the short-run rates essentially
unaffected Finally and in opposition to the central role adjustment costs seem to play in both
53 Financial crowding outeff~cts of government deficits
Goulder-Summers and Pereiras models Auerbach-Kotlikoff suggest that the time path of interest
rates induced by a policy of deficit financing seems to be insensitive to the presence of adjustment
costs
bull
35
54 The role 01 consumer expectatlons
The expectations analysis has uncovered one of the benefits of dynamic modelling
Ballard-Goulder (1985) find that the appeal of adopting a consumption tax depends on the level of
foresight possessed by the consumers Furthermore additional foresight may be welfare worsening
This is a second best type of result This tends 10 occur under policies that lead to capital deepening
and declining rate of return to capital over time To the extent that consumers have more foresight
they will be beller equippod to anticipate the fall in the rental price of capital Consequently if a
saving incentive is enacted people save more with myopia than with perfect foresight Given the
exi~tence of taxes on capita and the discrepancy between the private an~--sccial returns to capital
the greater savings with myopia is beller socially Ballard-Goulder find that the welfare gains
from a consumption tax are reduced by about 10 when we move from myopia to farleet foresight
Therefore the attractivenness of adopting a consumption tax seems fairly [obust across these
foresight specifications
I
bull
36
6 Tbe power and weaknesses 01 dynamic modelling the example 01 corporale
tax Integrallon
The Issue of corporate tax integration allows us to sh~w the stre-ngths and weaknesses or the
dynamic approach Empirical evidence using CGE techniques indicates that depending on the precise
scheme of integration and the 1ax replacement methods integratiOn may have substantial effects In
Ihe work of FuliertonmiddotKingmiddotShoven-Whalley (1980 1981) total integration was found to yield an bull
annual static eHielency gain of $4 to $8 billion in 1973 dollars Simulated dynamic gains may be as
large as $695 billion or abcu 14 of the present value of tulUre consumption and leisure in the
US economy These gains result primarily from interindustry reallocalions of Investment and an
improved inter~emporal allocation of consumption
Mora recent work emphasizes hOw consumers asset portfolio decisions and firms finanCial
decisions affect the efficiency gains from integration Slemrod (1980) focusing on consumers
asset ponfoio deCisions finds static efficiency gains which are about twice -as Iarge as those
repOrted iiy Fulierton-King-ShovenWhalley (1961) FuUrlon-Gurdon (1983) focus on the
firms financial decisions They roporl efficiency gains of 6 of GNP from the elimination of the
tax distor1ions favoring debt However when they eliminate the corporate tax and repiace it with
increaSed persona income taxes addlHonal dls10t1ions are created in the optimal labot~leisure
decisions These distortions tend to dominate the analysis slJ(h that the overall effects of complete
integration are very modest Galpr-LuckmiddotTodr (1986) address simultaneously consumers
asset pOrlfollo decisions and the firms financial decisions They also find very modest efficiency
gains from integralion
The above results are important but may be severely biased There are severa aspects of
economic behavior and modelling crucial for the study of Income tax Integration whiCh have no been
i 37
captured innny of the above papers One aspect is the absence of government deficitsmiddot a balanced
budget is assumed It is true Ihal resource crowding OUI is caplured In these mOdels but financial
crowding out induced by government spending is nOI Second investment is not derived from
optimtztlcn behavior Investment behavior passively accomodates endogenous saving decisions As
a consequence the differential impacts of policies in the incentives to save and to invest are not
captured Aso full capital mobility across secors is assumed with instantaneous capital
adjustmenlS towards optimal levels This assumption rules qut different costs of capital across
seclors and therefore differentiated reactions to tax policies changes Finallyhe modelling of both
gov9rnmerlt deficits and of endogenous real and financial investment decisions necessitates the
~nsideraiion of a dynamilt framework and the introduction of financiaJ assets govemment bonds and
iprivate financial assets A dynamic framework also highlights the efficiency effects of Integration on r
ithe ptimal intertemporal decisions The above models are either sIalic (Slemrods and
GalpermiddotltJckemiddotTode(sj or a dynamic sequ~nce of otherwise Slatic model bullbull I Pereir a (1986b) develops a dynamic applied general equilibrium model bull to study the tbull
efficiency effects of integration on the growth and allocation of investmenl across sectors Special
Iattention is paid to the structural affeelS of resource and financial crowding oul induced by
gove mect spending and deficits and to the real and financial Investment reactions across industries
to both tax integration and flnancial crowding out His mode departs from previous work on income
lax Integraticn and for that matler from most 01 the CGIE literalure for tax policy evaluation in
several direcllons It encompasses an endogenous sequential equilibrium struelure founded on
dynamic behavior with flexible expectations Government deficits are optlmaUy determined
Investment decisions are tward looking and the resull of optimizing behavior Several financial
assetsmiddot public and privatemiddot are considered
Simulation resulls Indicale that the welfare gains from Integration under large deficits are at
32
best lJlodest when measured in terms of the GNP The elimination of the corporate tax and its I replacemenl by increased income lax rates yields long run benefits which are never larger than
2 of the presenl yalue ct telr consumption and leisure onder the previous tax regime and 1
under the current tax regime However the short run w~lare effects of full integration tend 10 be
very low and in in some scenarios even negative This is a flew intertemporar pattern of efficiency Ieffects which reflects an adjJsiment lag in the interindustry investment decisions due to the
existence of costs of adjustment On the of her hand partial integration achieved by excluding I dividends from the corporale lax base syslematically yields negalive efficlencyeffeclS In Ihe long Inm these can be as high as 3 of present value of the future value of consumption and leisure This
is a new second best effect suggesllng Ihat less than complete integration may have perverse I efficiency effects
The dynamic slructure of PerclrBs model snowing for an improved freatment of real investment
declsions and government deficts provides a potential setting for a more accurate measure of the
costs and benefits of inlegration The simulation resullS suggest lower benefits and an intertempora Ibull
pattern of growing benefits
Simulation results also clearly suggest robust negative effects from partial integration How
reliable is this result If dJvidends were deductible from the corporate lax base (as are interest
payments) the preferenlial Ireatmant of debl over equily would be ellminaled Corporallons should
be expecled 10 react by decreasing the optimal deWeqully rallo However corporale financial
gecisions are exogenously set In Pereiras model as in aft of the other models surveyed that go as far
as dealing with the issue
Also withdrawing dividends from Ihe corporate tax base is a way of eliminating Ihe double
taxation of dividends at belh Ihe corporate and personnal Income levels How will the optimal
aliocellon of household saving accomodate that change in Ihe relalive return to corporale equity
39 I
t However hOusehold portfolio decisions are exogenouly set in Pereiras model as in all of the other
models surveyeltj go as far as dealing with the issue
IIt is sate to say that the evaluation 01 the benefits 01 partial Integration as defined above are
sevrey underestima1eO Meanwhile the distortionSoenerated by the increase in the marginal
personal tax rates designed to ralse the revenue foregone by the dividend exclusion are fully
aCCltlunted for A good measure of the costs of integration together With a poor evaluation of the
benefils may very well be the reason why partial integration is Simulated 10 yield neg alive
eHikiency gains
On a different vein as most of Ihe model surveyed here Pereiras is a closed economy model It
is legitimate to wonder how the resullS would change If international capital flows were anowed
IThis almost certainly would affect financial crowding out since a good d~al of the capital inflows
could be used to finance public debt
For these various reasons the results should be treated as preliminary but they strongly I suggosl that the dynamic structure provides valuable new insights into the corporate tax integration SSU2
40
7 Concluding remarks I I I i
What has been acccmplished Ihis far The success of Ihe CGE research in laelding the challenge
of dynamic modelling can not be denied
I iGreat progress in the modelling of household behavior has been achieved Promising
developm~nts in the modeUing of production and government behavior have also been accomplished
Interesting innovations in the concept of equilibrium and computation techniques were discussed
The policy analysis was greatly enriched by considering transitional effects together with longer
run steadymiddotstate equilibrium paths generalized equal yield strategies accomodating different
financial crowding out impacts and generalized dynamic policy evaluation indicators
An important set of issues have been addresed by the models surveyed Traditional issues
focusing on the effects of capital taxation and consumption taxes have been pursued In terms of
capital taxation for example the intertemporal nature of the issue was enhanced by allowing an
optimal evolution of capital stock in the economy and forward looking optimal investment decisions
In Andersson (1987) GouldermiddotSummers (1987) and Pereira (1986b 1987d) this is coupled
with a improved treatment of several tax provisions like investment tax credits and depreciation
allowances In terms of the consumption taxes developments in the specification of Intertemporal
household behavior the introduction of overlapping generations and the modelling of
intergenerational links via bequest motives as well as a closer attention to demographic evolution
provided a much improved economic setting for the understanding of the several aspects of the
problem
Furthermore dynamic modelling has permitted the addressing of a set of new issues Policy
Issues ranging from dynamic tax policy analysis to the evaluation of exogenous changes in
government expenditures to the impact of different methods of financing government expenditure to
4 1
the importance of tinanciai croHrii1g out to 1he relevance of consumorS expectations in tne policy
results have been addressed
Despile all of the progress and in part due to such progress several avenues are wide open for
much neded additional research The following seem to be the most Interesting and promising
areasmiddot First the specincation of liquidity constraints to household behavior may help to obtain a
better understanding of policy chenges that affect directly the interest rates in the economy Second
major attention needs to be paid to the incorporafion of a roraign sector (with open capital markets)
Third some efforts are needed in terms of finding adequate computation techniques now that
dynamic modelling eally makes the curse of dimensionality worse than eVOrbull
The modelling of financial markets is the single most unsatisfactory speet of the dynamic
models surveyed here~ The problems associated with provlding a mo~ adequate treatment of
financial markets and in general endogenous and optimal financial deci$ions bull both at the corporate
and at the personal levelsmiddot have 10 be addressed That necessarily requires Introducing uncertainty
into the models To b adequate this must go well beyond the mere assumption of point expeetalions
with the whole array of modelling implementalion and evaluation problems thai generates
bull I i
42
REFERENCES
Abel A and O Blanchard 1983 An Intertemporal Model of Saving and Interest Econometrica
51 pp 675-92
~ Andersson K 1987 Tar-ation of Capital in Sweden - A General Equilibrium Model Ph D
Dissenation UnIversity of Lund Sweder
Auerbach A L KotlikoH 1983 National Savings Economic Welfare and the Structure of
Taxation in M Feldstein ed Behavioral Simulation Methods in Tax Policy- Analysis Chicago
University of Chicago Press
Auerbach A l Kotlikoff 1984 Social Security and the Economics of Demographic-
Transition in H Aaron and G Burtless eds Retirement and Economic Behavior Washington DC
Brookings Institution
Auerbach A L Kotlikoff 1987 Dynamic Fiscal Policy Cambridge Cambridge University
Press
Auerbach A L Kotlikof and J Skinner 1983 The Efficiency Gains from Dynamic Tax
Reform Internatiooaj Economic Reyiew 24 pp 81middot100
Ballard C 1983 Evaluation of the Consumption Tax with Dynamic General Equilibrium
Models Ph D Dissertation Stanford University
Ballard C 1987 Tax Policy and Consumer Foresight A General Equilibrium Simulation
Study Economic Inquiry 25 pp 267-284
Ballard C and L Goulder 1985 Consumption Taxes ForeSight and Welfare a Computatonal
General Equilibrium Analysis in J Piggott and J Whalley eds New Deyelooments in Apolied
General Equilibrium Analysis Cambridge Cambridge University Press
Ballard C D Fullerton J Shoven and J Whalley 1985 A General Equilibrium Model for
t 43
Tax Eoliey Eyalua~iQO NBER University or Chicago Press
Borges A (1987) bull A Survey of Compula1ional General Equilibnum Models CECD Economic
Slud~ 8
Bovenbig L 19B3 imperfect Capital Mobility in a Perfect Foresight Herberger Model
Mimea UC Br~eley
Bovenberg L 1984 Capitel Accumulation and Capital Immobility Qmiddottheory In a Dynamic General Equilibrium Framework PhD dissertation University of California at Berkeley t
Bovenberg L 19850 Dynamic General Equilibrium Tax Models with Adjustment Costs in A iManne ed bull 1985 EtQnomrc ECUjilbrium Model Formulation and Solution Mathematical
lProgramming Siudy 23 Arnsterdam NorthmiddotHoliand
Bovenberg L t 1986 Capital Income Taxation in Growing Open Eeondmies Journa of Pubtic
Eamp(omlcs 3t pp 347middot376 IBovenberg L ald W Keller 1981 Dynamics and Nonlinearlties in Applied General
Equilibrium Models Internal Report Mlmiddot81middot208 Department for Statistical Methods CBS bull
Vooiburg The Netherlands
Charnley C 1981 The Welfare Cost of C~pltal income Taxation In a Growing Economy
IJournal of Political ECQnQm~ 89 3 pp 468middot496 I
Charnley C 19810 Efficient Tax Reform in a Dynamic Model of General Equilibrium I Mimeo Yale University I Charnley C 1982 On the Optimal Taxation In Stylized Models 01 Inlartemporal General
Equilibrium Mimeo Yale University I Decaluw B and A Manens 1985 Pays en Oeveloppement at Modeles Calculables OEquilibre
General - Une Revue de al litterature empirique Monographie Sene 9 Volume 3 Centre de
Recherche et Developpemont Economlque Unrslt de Montreal
44
bull
I I
I
I
1
Erlich $ V Ginsburgh and L van der Heyden 1987 Wher~ do Real Wage Policies Lead
Belgium bull A General Equilibrium Analysis European Economic Reyiew (forthcoming)
air R and J Taylor 1983 Solution and Maximum Likelihood Estimation of Dynamic
Nonlinear Ratioant Expeclations Models Ecooometrica 51 pp 1169middot1186
Feltenstein A 1984 Money and Bonds In a Disaggregated Ec~nonYin Searl 1-1 J Shoven
Eds Aoolied ~eoeral EqJiliDCiulD tioalYsis Cambridge University Press Cambridge
Feltenstein A 19B6 A Dynamic General Equllibrlu Analysis of Financial Crowding Out
Theory with an Application to Australia Journal of Public Economics 31 No1
Fullerlon D R Gorden 1983 A Reexamination of Tax Distorllons in General Equilibrium
Models in M Feldstein Ed Behayora Simulation Methods in Tax pOlicy Analysis Chicago
University of Chipgo Press
Fvllerlon D J Shaven and J Whalley 1983 Replacing the US Income Tax with a
Progressive Consumption Tax JoulOal of Public ECQnomics 20 pp3middot23
Fullerton D Y Henderson J Shaven 1984 A COmparision of Methodologies in Empirical
General Equilibrium Models of Taxation in Searl H J Shoven Eds ~pQlied Geoaral Jguilibrium
lIoalyss Cambridge University Press Cambridge
Fullerton 0 A King J Shaven and J Whaney 1981 Corporate Tax Integration in the United
Statesmiddot a General Equiiibdufi1 Approach American Economic Review 71 pp~ 677~691
Galper H R Lucke E Tader 1985 Taxation portlor Choice and the Allocation of Capital A
General Equilibrium Approach Mimeo
Gill p W Murray M Saunders M Wright 1986 Users Guide for NPSOL (Version 40) A
Fortran Package for Nonlinear Programming T echnicsl Report SOL 86middot2 Deparlment of
Operations Research Stanford University
Glnsbur~h V and L van der Heyden 1985 General Equilibrium with Wage Rigidities An
45
AppJication to Belgium in A Manne ed t Economic EQumprit1m~ Made Formulation and $oJioo
Mathematical Programming Study 23 Amsterdam NorthmiddotHolland ~
Goulet J 1968 1 Adjustrent CO$1S in the Theory o(the Firm Reyiew of EconomiC Studies 35 bull
pp47middot55
Goulder L 1985 Tax-Financed versus BondmiddotFinancedGhanges in Governent Spending
bull Mimeo Harva(d UniversilYmiddot
Goulder l J Shaven and J Whalley 1983 Domestic Tax Policy and the Foreign Sector The
importance of altemative foreign policy formulations to results from a general equilibrium tax
analysis model In ~n methods in Ibe antico of income from caoia1 ed Martin
Feldstein Chicago University of Chicago riSS
Goulder l L Summers 1987 Tax Policy Asset Prices and 3rowtly ~
A General Equilibrium Analysis NBER Working Pper No 212B
Grandmont J 1982 Temporary General Equilibrium Theory Ch 20 in Handbook 01
Mathemalical Eccnornics Vol II NorthmiddotHoliand
Harberger A 1959 The Corporate Income Tax An empirical Appraisal In lax Revision
Comoendjum Voll House Cornmitee on Ways and Means Washington DC Govemment Printing
Office
Harberger A 1962 The Incidence of the Corporate Income Tax JOULn1 of Political
Economy 70 pp 215middot240
Harberger A 1964 Taxation Resource Aliocalrlon and Welfare fn lb Role of Dir bullbulll and
Indjresl Taxes io tbe Fede[al Reserve s~I~Ill Princeton Princeton Unlveresity Press
James J 1985 New Developments in the Applications of General Equilibrium models to
Economic History in Piggott aand Whalley Eds
Jorgenson D and K Yun 1984 Tax Policy and Capital Allocation Harvard Inslitute of
bull
I
I f
Economic Research WP 1107
Judd K 1965 ShonmiddotRun Analysis of Fiscal Policy in a Simple Perfect Foresight Model
Journal of poJiHcal 5CCJiW1l pp 298middot319 r
Upton D J Potbc J Sachsa nd L Summers 1982 Multiple Shooting in Rational
Expectations Models fuoometrlc 50 1329middot1333
Manne A ed 1985 fconQrnjc Eoullibrium Model FOrIDujaljQo and Solution Mathematical
Pogrammlng Study 23 Amsterdam NorthmiddotHolland
Manne A 1985 Or the Formulation and Solution 01 Economic Equilibrium Models In A
Manne ed ECQOQlli-Jr~ibmiddotum ~tQdel EQlJulailon and Solution Malhematical Programming
Study 23 Amsterdam NorthmiddotHolland
Mansur A and J Whalley 1984 Numerical Specification of Applied General Equilibrium
Models in Scarf H J Shoven Eds Aoplied General Eouillbrium AnaiysectIsect Cambridge Universily
Press Cambridge
Merrill 0 1972 Applications and Extensions of an Algorithm thaI Computes FixedmiddotPoints of
Cenain Upper Semimiddotcominuou Point to Set Mappings PhD Disserlation Unlvel$ity 01 Michigan
Negishi T 1960 Welfar Economics and Exislence I an Equnlbrlum for a Competitive
Economy MelroeCOOQ11ka 12 92middot97
Pereira A 1985 On the Existence and Uniqueness of Optimal Ou1put Path lor CRTS Firms in
Adjustment Costs Technologies Mimeo Stanford University
Pereira A 1985a Consumer Expectations and Ihe COSI 01 UllUty In an Intertemporat
Framework Mlmeo Stanford University
Pereira A 1985b Corporale Tax Integrallon and Government Deficits A Dynamic General
Equilibrium Anaiysls Mlmeo Stanford University
Pereir A 1987 On Ihe Computation of the Users Costs of Stocks Ecooomi (forthcoming)
47
Pereira A bull 1987b Equal Yield Alternatives and Government Oeficlts Mimeof Stanford
University
Pereira A 1987c A Dynamic Applied General ~uHlbrium Model for Tax Policy Evaluation
Comlletamp Docurnent2tion Mimeo University of California San Diego (in progress)
Pereira A 1987d Should the investment Tax Credit be Rei(ltroduced Mlmeo University
of California San Diego (ir progreSs)
Piggott J and J Whalley eds 1985 tJew QeyeQpments in Applied General EQuilibrium
bullenalysjs Cambridge Cambridge University Press
Preckel Pbull 1985 Alternative algorithms for computing economic equilibria In A Manne Ed
Radnor R 19amp3 Equllibrium under Uncertainty In K Arrow M Intrilligaor ods
I ibull
Robinson S 1 g86 Mullisectorai Models of Oeveloping CountrieS a Survey Oivision of
Agricullure and Resources UC Borkeley WP 401
Scarf H 1967 The approximation of fixed points of a continuous mapping ~MM Jouwal of
epplid Mathornallcs 15 pp 1328middot43
Scarf H with T Hansen 1973 00 the Computation of EconomiC Eguilibria New Haven Yale
University Press
Scarf H J Shoven Eds 1984 61l1llied Genera Elujlibrium 6nalyssect Cambridge University
~ Press Cambridge
Shoven J 1976 The Incidence and Efficiency Effects of Taxe on Income from Capital
Jurnal of ramie1 Economy 84 pp 1261-83
Shoven J 1983 Applied General Equilibrium Tax Modemng IMF Slaff PaoerS 30
Shoven J and LB Simon 1986 Share Repurchases and Acquisitions An Analysis of which
Firms Participate NBER Working Paper No 2243
bull
48
IShaven J and J Whalley 1972 ft General Equmbrium Calculation of lhe Effects of
Differential Taxation of Incoma from Capital in the US Jauroe of public Econgmics 1
pp281middot321 (
Shoven J bull and J Whalley 1973 General Equilibrium with Tax$s a computation procedure
and an existence pr00f ReviEhi of Economic Studies 40 pp 475middot490
Shoven J J Whalley 1977 Equal Yield Tax Alternatives A General Equilibrium
Computa~ui0nal Tfchnique J~HJrnal of PUblic EconomIcs 8 pp 211~224 bull
Shaven J J Whalley 1984 Applied GeneralmiddotEquilibrlum MOdels of Taxalion and
International Tr8de An Introduction and Survey Journal of Economic literature 23t pp
1007middot1051
Slemrod J 1980 A General Equilibrium Model of Capital Income Taxation Ph D bullDissertation Harvard Univnrsity
Slemrod J 1983 A General Equilibrium Model of Taxation with Endogenous Financial
Behavior in M Feldstein edbull Behavioral Simulation Methods In Tax Policy Analysis Chicago
UniVerSity of Chicago Press
Stone J 1985 Sequential Optimization and Complementarily Techniques for computing
Economic equilibrium in A Manne Ed
Summers L 1981 Taxation and Corporate Investment a q-Theory Approach BroQkinas
Summers L 1985 Taxation and the Size and ComposHion 01 the Capital Stock An Asset Price
Approach Mlmeo Harvard University
Whalley J 1975 A Geneal Equiloibrium Assessment of the 1973 United Kingdom tax
reform Economica 42 pp 139middot61
1JLTIllOS JIQll~)ltG PAPERS PUBLICADOS bull
I
I I
j
112 70 - LUCENA iJiCIIJo ~ IIEnvironrilent Honitoring and Orqanization Structure T (Setellibro 1987)
nI 71 - L~ENA Diogo A Note on the Representation of Information Stxueture (SeteJObro 1987)
nQ 72 - LUCENA D1CgO Envuorment 111)nitoring and Organization
Structure II (Setelilhro 1987)
ril 73 - AlmmE3 Antonio Pais e CAsPAR Yitor bull Tributacao Inoentivos e InvesUlllentos Analise Qualitativa noYellibro 19S7)
n2 74 - liAlltOSA rose Pedro Opt1lllal lIage R1gidity A SUggestedtlthodology to Teet tb Theory and an Application (Outubro 1987)
nI 75 - CABRAL L 11 B Three lIotes on SYJIIlletric Gemes witt itsYJIIlletric Equilibri (NoveJObro 1987)
n2 76 - LUCENlt D10gOTo Search or Not to Search (Fevereiro1967)
nQ 17 - SA Jorge asconcelos euroI A Theory of SynergyB (Fevereiro1987)
n2 76 - VILARES tanuei Jose Os Be IntenaMios IlIIportados ColIIO Factor de Produao (Julho 1987)
nQ 79 - SA J~rge VasCQncel03 e HoY to Copete and Co~unicate in tature Industrial Products_ (aio 1988)
n2 80 - ROIl Rafael Learning and Capacity Expansion in a New larket Under Uncertainty (Fevereiro 1988)
n2 81 PEREIRA Alfredo llarvilo Survey of Dynamic Computational General EquilibriUlll Hodel for Tax Policy Evaluation (Outubro 1987)
Qualquer in ormacao sabre o 1IIorlltlng Papers ja publlcados sera prestada pelo Secretariado de Apoio aos Docentes podendo os ~e3OSf
ser adquirido5 ne Secyo de Vlndas da Faculdade de EconollLia UNL na Travessa Esteviio Pinto Cltpolide - 1000 LISllOA
bull
1 Z
model for lax policy evauajon were ShovenWhalley (1972) Whalley (1975) and Shoven
(1916)
Completely static models as Shaven-Whalleys are unsatisfactory for many tax reform ISsues
These include corporate lax Integration effects of investment tax credits effects of accelerated
depreciations consumer Or e)penditure tzxes~ importance of saving subsidies like IRAs etc These
are essentialy dynamiC ksues They involve not only the allocation of capital across sectQfsbut
perhaps more Important the capital intensity of the economy But the capital accumulation and
capital reallocation take time and may involve adjustment costs Because of these issues Ballardmiddot
Fulierton~Shoven~Whaloy took the first steps towards developing a dynamic model
In this conlexl the ShovenWhalley model has been extended and implemented for Ihe US
-economy In BuIIBrdmiddotFulienonmiddotShovenmiddotWhalley (1985) While Ihis book completely documents the
mo~el it was used in several publicalions beginning in 1978 Their mode) consists of nineteem
production sectors twelvo households and fifteen consumer goods It includes a very detailed set of
taxes including the federal and state personallncome taxes federal and state corporate income taxes
SOCial Security taxes pro1J0rty taxes unemployment Insurance excise taxes sales taxes etc It
has been calibl att-d to reproduce the 1973 US economy Recentty a version corresponding to the
1983 JS economy has been developed
wThe 8alard-Fullenon-Shoven Whalley model is dynamic In the sense that consumers face a
choice between current consumption and leisure versus future consumption (which can be
purchased via s wlngs) The consumer classes act as if they were maximizing a nested CES utility
function over the domain of a CES aggregate of contempo-raneous consumer goods and leisure and
future consumption subject to their income constraint The parameters of those funcHons
determine the shares of income devo1ed to each commodity to saving and to the purchase of leisure
They also determine the two key elasticities in the models ~ the elasticity of labor supply with
bull
13
respecl to tgte rbullbull1 aoer lax wage rae and the elasticity of saving with respecl 10 the real after lax- rate of retum to capital
In the model consumerS have myopic expectations regarding lulure prices and in particular
regarding the future rale of return 10 capilal Ballard (1983) and BallardmiddotGoulder (1985)
inco~ora1e both perfect foresight and limited foresight into this model Future consumption is
acquired by buying a fixed composition portfolio of real investments that offer an infinite annuity
of returns
The production side of Ihe model is compielely static The mod1 Incorporates a constantmiddot
olasicily 01 substitution betweeen prlmary inputs in production (capital and labor) and fixed I coefficients for intermediate inputs The model distinguIshes between Industrial outputs and
eonsumer goods for the simple reason Ihat tile data are classified differenUy Industrial sectol$ I I involve such categories as forestry and fisherle metal mino9 and publishing and printing while
consumers purchas-3 fUfHi1ure 2ulomoblles and books This fact is recognized in the model by Ibull
incorporatirg a secord stage Of production which converts industrial outputs intO consumer goods
This technology is uStally mod9-led as a flxed~coefficient conversion matrix
The BallardmiddotFINton-ShovenWhalley model assumeS that the private sector finance
marginai ioveslment with the same composition of debt and equity as currently exists in each sector
This is the same arsumrtion that Herberger originally used Investors all hold debt and equity in tile
same proportion Thampreiore this ownership can be aggregated into simply capital ownership For
tax purposes however the separate treatment of debt and equity is taken into account at both the
corporate and persofa Io=vel SimllarlYJ the dividend poIicies on corporale equitY are established
exogenously There are no government bonds in the model since there are no government deficits
The modol is solved for a sequence (as many as 100) of temporary equilibria wltll consumers
allocating income between present and future conumption at each point in time The palh for 1he
14
economy ~ a sel of connected equilibria The connectIon is pra~ided by capital accumulation
Capital accumulation is endogenous and determined by saving The model stans with a sociagt
accOunting mi1rix In the bzse case the economy is assumed to be in a steaqy state growth -path (along
which all relative prices ~fe constant) The madel solves for both the naw steady state growth path
and the transltion 10 it zf~er a policy intervention rhe authors have frequently addressed the
question-of howong rt tKes to effectively settle un10 a new steady S1ate growth path
The dynamics of the model are limited however in that future consumption is collapsed into a
composite commodity Aso the absence of government deficits and the lack ofproductlon dynamics
limits the realism of the BallardmiddotFuliertonmiddotShoven-Whatley model for the analysis of dynamic
policy issues (such as the adoption 01 a consumption tax or Ihe elimination of the investment tax
credit) The models we survey try to Improve upon Ihe Ballard-FlIiertonmiddotShaven-Whalley
dynamic modelling The models Jn~orporate more forward lookTng behavior They improve the
dynamics on the conampumrtkm side of the economy and some of them introduce true 1ynamlc behavior
on the part oI the producers We turn now to the issues inY9lv_ed in developing such dynamic models
15 I I ~3 ~ssues in the modelling Of dynamic behavior
We will 100k firt at the speclficalion of economic behavior of consume(s~ producers
goyemment and Ihe r~sl of Ihe world The modemog of corporale and household financial decisions
I is then addr5sed Finally Ihe eencept of equilibrium is discussed
I
i
I 1
31 Consumers behavior
Early efforts 10 build dynamic features into the economic behavior 01 consumers are due to
Ballard (1963) and Auerbach-Kotllkoff (1983 1984) Their work has been closely followed by
several subsequent authors In faet dynamic household behavior Is the single most pervasive aspect
amohg the models surveyed
All the models Inc-O1orale some loro of lila-cycle behavior Household behavior is dewmloed
by the maximization of an addWVely separable time invariant intertemporal utillty function The
utility lune1ion is defined oyer the domain of the consumption goods In the economy In most of the
models leisure is also an argument in utility so that labor supply is optimally determined
Utility maximi4ation is subject to a lifetime intertemporal budget constraint which equalizes
the presen1 value of consumers income and expenditure More reeentiy in Andersson (1987) bull
Bovenberg (1985) and Pereira (1986b 1987d) the constraint i$ defined as a sequence 01
recursive equations of motion on wealth This has the potential advantage 01 accomodating liquidity
constraints However It should be recognized that in the absence 01 liquidity constraints (ie when
consumers are free to completely borrow against future income) the two specificallons of the
househofd ccnstraint are essentiay equivatent Furthermore tn both versions saving is optimally
determined as a way of translaring wealth intertemporally
1 6
Some lodels now have a sophisticated dynamic specification of hoosehokl In Ballard (1983) -
BallardmiddotGoulder (1985) and Auerbach-Kotlikof (1983 1984 1987) consumers behavio is
embedded in an iterQEneraional setting Each of them has S5 age cohorts simulataneously alive
Som-e measure of jntergener1ional altruism is considered in the form of bequest motives In Ballard
(1983) and Ballard-Gouldr (1985) consumers derive utility directly from bequthing parl of
their wealth Therefore bequests assume 1he form of a scrap value of terminal wealth of a
generation In AuerbachmiddotKotlikof (1983 1984 1987) each generation empathyz bullbull with Ihe
utility of future generations ovar bequethd wealth In addition Ballard (1ge3) includes detailed
demographic poiections wfill the intent of incorporating in the policy analysis the effects of the
bull Post World War tI baby-boom
In the real world household decisions also include the optimat allo~atiOn of saving among
shyalternative physical or financial assets Theories of household portfolio behavior are relatively
Iwen~establshed However they involve uncertainty as a crucial element Accordingly they have
middotnot been incorpora~ed in the context of the determinIstic dynamic models we sutyey here We will 1 Come backt-O this Issue below
32 Producer behavior
The efforts to build dynamic features into the economic behavior of producers are more recellt
and less widely adopted The first attempts are due to Bovnberg (1984 1985) and Summers
(1965) Dynamic behavior has been more fully incorporated In the recent models of Andersson
(1987) Auerbach-Kotlikoff (19B7) GouldermiddotSummers (1987) and Pereira (1986b 1987d)
Part of the reason why production side dinamlcs has been more slowty adopted is the weak
supply of accepted theories referring to the dynamic behavior of the firms In the models referred
I 17
to above gynamlc producion and investment b~havior are middotinduced by the existence of capital
adjustment costs and linked to Tobins q theory Adjuslment costs are designed to capture both the
incomplete mobility of capital aoross industries and installation costs ~e the costs of adjusting
capital towards its cpHmai level)
Adjllstment costs can be conceived as internal to the firm and measured in terms of foregone
outpul along the lines of Lucas (1967) Alternatively adjustment costs can be viewed as actual
external cests incurred together with Ihe purchase costs along the Unes of Gould (1969) With the
exception of Pereira (198Gb 1937d) all of the CGE authors follow the former approach
The firms in the economy maximize their market value as the present discountad value of the
future stream of dividends In Andersson (1981) and Pereira (198Gb 1987d) firms are seen as
maximizing the present discounted value of net cash flow Maximization is constrained by Ihe
adjustmenl cosl lechnology and ~n equation of motion describing Ihe evolution of the capital stock
It should be noted that undereogenous divldendrelention rules the two problems are equivalent II
should also be nOled Ihat a salisfectory economic rational for th bullbullistoo of dMdonds with the
present tax code is missing in the profession (Shoven (1986))
The models with static formulation of producers behavior are characterized by passive
investment behavior Investment merely accomodates to saving in the eccnomy With a dynamic
formulation induced b adjustment costs real investmen decisions are forward looking Investmenl
Is endogenously and optimally determined by the firms A fundamental difference between the
shonmiddotrun in which capital stock is given and longmiddotrun in which Ihe level of capital is allowed to be
optimally determined is emphasized
This extra richness of produc1ion dynamics is not without costs A careful look at the summary
tables will clearly show an inverse relation between the adoptlon of dynamic features in production
and the level of disagreggation of the production side of the economy In fact with production
bull
j
1 6
dynamics $e dimension of the problems is immensly increased Let IS be more specific
In a static framework with constant returns to scale production technology the output level ismiddot
indeterminate The optimal aliocation of inputs can be obtained by cost minimization with any
feasible output level generatillg zero profits Such zero profit conditions are used to solve for the
output prices in terms of the factor prices and hencemiddot to reduce the dlmensionality of the problem
from the -number of commodities and factors to the number of factors Thus even if the model deals
with 30 production sectors the computation of an economic equilibrium can take place using only
the dimensionality of the primary inputs in the economy
Now under certain regularity conditions on the production and adjustment costs technologies
leading to enough concavity of the optimality objective the intertemporal output path for the firm is
endogenously optimally and uniquely determined even with constant relums to scale technologies
(see Pereira (1985 1987a) Accordin~y with adjustment costs in the model optimal profits
will in general be non-zero This result has crucial implications for the computatiQn of equilibrium
in the model~ with production dynamics With adjustment costs no reduction of dimensionality is
possible The curse of dimensionality returns as a binding constraint
The introduction of adjustment costs adds another significant complication to the model With
capital being less that perfectly mobile in the economy different rates of return on capital will
exist in different sectors This is a difficult problem to tackle conceptually in the absence of
uncertainty Also in the real world producer maximization choices include also its choice of
financial ratios (debVequity) and payout rates (dividendretained earnings) These financial
subjects are difficult and much Clf this behavior is still taken as exogenous by the modellers We
will come back to these issues below
middot i
19
bull 33 government behavior
middot
Two issues dominate the modelling of government behavIor First govemment behavior has
Ibeen typically seen in Ihe CGE liIerature for lax poliCY evaluation as constrained by yearly balanced
I budsets (see for example 8allard-Fullrton-S~ovn-Whafley 098Sraquo The analysis of
govemment defidlS and public debl in a CelE context requires a dynamic selling Second the level of
government expenditures is either exegenouty given as in Auerbach-KoUikofi (1984 1987)
6ovenbrg (1984 1985) Feltenstein (1984 1986) and Jorgenson-Yun (1984) or
endogenously (but no optimally) determined by the balanced budget conditions as in
6allard-Fullerton-Shoven-Whalley (1985) Andersson (1987) Erllch-Ginshurgh-Heyden
~
(1987) Gouder (19aS) and Gouder-Summers (1987) In the second case the composition of
bull public expenditures is often optimally determined However the leve~ of government expendi1ures
can only 00 endogenously and optimally determined if the government is seen as an optimizing agent
and is allowed to run deficits
The first attempts to deal wiU the government defiCits in CGE tax models are due to
Auerbach-Kollilltoff (1984) and Feensain (1984) In Auerbach-Kotlikof (1984) government
expenditures afe exogenous They grow at the rate of growth of population However given 1he tax
structure and tax revenues yearly deficits and surpluses are atTowed subject to an in1ertempond
constraint that the present value of future tax revenues equals the present value of future
expenditures
In Fellenstein (1984 1986) government expenditures are also exogenously given He also
allows the government to run deficits to finance expenditures in excess to 1ax revenues However
surpluses are returned to CQnsumars in the form cf transfers Accordingly government ts not
subject to any constraint regarding the future repayment of public debt
In GouJder (19851 the 9Vernmenl maximizes a static social welfare ~nctio subject to a-
bull
20 I i
I
I I I
I bull
balanced budget conSirtlL This follows the optimal alocatkm of government expenditures alpng
the lines of Baliard-FHeiton-Shoven-Whalley (1985) Ttle model Is generallzea to allow for
exogenous changes in HIe time path of government expenditures Two financing alternatives are
consjdeft~d and contr8s1ed EldditionaJ tax revenues and bond issuance
Pereira (198Gb 1137d) auempts to address both the incorporation of deficits and the
determination of govement expenditures The path of government expenditures and the path of
delicilSlsurpluses (ano therefore the path lor debt) are endogenously and optimally determined The
government is seen as mltJximizing n intertemporal social welfare function given the 13x structure
Optimization is subject to a sequence of recursive equations of motion reflecting the evolution of the
ptlbllc debt allowing fDr government budget imbalances It should be stressathat this specification
01 the government cOMtcalnt is equivalent to the specificaUon in Auerbaeh-Kotlikoff (1983 1987)
when no liquidity constraints affect -gove~nment behavior
The extra richness in the treatment of government behavior allows tfte examination of
government debt poices in a truly dynamic setting Also financial crowding out effects induced by
government deficits can be analyzed (see Auerbach-Kottlikoff (1987) Feltenstein (1986) and
Perer (198Gb 1987e))
34 Foreign sector
Most of the open economy models follow the assumptions of balanced trade with import and
export net demands characterized by constant elasticities along the line of Ballardmiddot
-Fullerton-Shaven-Whalley (1965) Such is the case of Salfard (1963) Ballard-Goulder
(19S5) and Goulder-Sllmmers (1987)_ In Feltenstin (1985) the rest of the world is treated as
21
an addjlion~ consumer group
Bovcoberg (1986) develops a model In which two economies lIr considered each following
intertempoml perlect foresight paths These economies meet in the international forum Their
trade relaroMhips are characterized by yearly balanced trade accounts
None of the new generation of dynamic cae tax models has yet IncorPorated the lnternational
capital flows as done in aoulder-Shoven-Whalley (1983) for the earlier Ballarrshy
-Fullertonmiddot Shaven-Whalley (1985) model This is unfortunate as there is an important
Intertmporal aspect to international lending The first al1empts along Ihese lines are due to
Andersson (1987) and Erlich-Glnsburgh-Heyden (1987) Andersson (1987) in his model of the
Swedish economy adopts a closed economy appraoch in which rates of return in the domesJic
economy are largely determined by the international capital markets Fidinterest rate induce
Intlnatlonal capital flows which determine and finance the International lrade imbalance In
lurn In ErlichGinsburghHeydenmiddot (1987) foreign trade Is generaled according to an
intertemporal Irade welfare lunction with constant import and export elasticities In the shortmiddotrun
they allow inlernational trade inbalances which generate capilal ftows to the domestic households In
the long run however trade balance is assumed
35 The need for financial markets
A dynamic economic Wucture not only provides the ideal environment 10 model many features of
economic behavior it also permits one to Incorporate the financial side of the economy If the
government is allowed to run deficits the question 01 deficll financing automatically follows If
investment is optimally dewmined and returns to capital are different aClOSS sectors problems of
investment financing arise If there are seeral final1cfal assets In the economymiddot government
4-
I i
ibull
I I I I
I
j
22
bonds privpte bonds and equity (or simply physical capital installed inmiddot different sectors) the
problem of allocation of saving among assets with potentially different returns arises
The papers surveyed var greatly with respect to l~e extent of their attention to the financial
side of the economy At Clne extreme are the models in ~anardmiddotFuliertonmiddotShoven-Whalley (1985)
Andersson (1987) Ballltld (1983) Ballard-Goulder (1985) Bovenberg (1985 1986) and
Erlich-Gjnsburgh-Heyden C19B which are devoted exclusively to the real side of the economy
In turn Auerbach-KotliKoff (1984 1987) Feltenstein (1984 1986) and Goulde (1985)
allow for government debt In these models saving finances changes in government debt and physical
capital Private and publlc 2ssets are perceived by the households as perfect substitutes The
allocation of saving merely adjusts to the relative demands for funds
Feltenstein (1984 1986) is the only model surveyed which introduces money Government deficits are financed by issuing money and bonds according to an exogenously given rule Money is
demanded by consumers for transaction motives and an exogenously given fraction as a store of value
On the other hand government bonds and physical capital are the vehicles for the intertemporal
transfer of wealth
Summers (1985) and GOlldermiddotSummers (1987) introduce a whole menu of financial assets shy
firm specific equity capital Different assets earn different rates of retum However such rates
are equal up to constant and exogenous sector speCific risk premia Therefore the introduction of
constant exogenous risk premia as helpful as it may be in the context of calibration does not solve
the main issue of the non-optimality of the allocation of saving Also talking about risk premia in a
deterministic context is somewhat unsatisfactory
Pereira (1986b 1987d) also introduces a whole menu of assets private and public bonds and
firm specific equity However all the assets are expected to yield the same rate of return and
therefore perc~jved as perfect substitutes The different asset types allow consideration of
23
exogenOU$~8tUequjty triO dividendlrc~enlion rufes and therefore seve-ral sources of investment
financing bond equity and retained earnings
The nonmiddotcptimalHy of the allocation of saving and the absence or exogenelty of corporate
financial rules is _a reftectlon of the limitations of the determInistic approach Either in a context of
perfect anticipation oi the future prices Of in general within the t$atm of point price
expectations all the papers follow a deterministic approach Under such circumstances consumers
either expect diHerent rales of return (inclusive of risk premium) across assets in which case
they wilt buy only one asset (that with highest rate) Of they expect equal rat~s of relurn in which
case Ihey are indifferent about the asset composition of Iheir portfoliO There is no way of
Iradingmiddotoff rales of return and risks to obtain an optimal interiof solution to the problem of the
allocation of saving
The most advanced contribution in the modelling of saving allocation in a CGE setting is due to
Slemrod (1980 1983) in the context of a static one-period model In his model consumers act
according to a two stage separable decision process They Hrsl decide on how much to save Then
they decide on the allocation of saving according to an Indirect uttlity function dependent on the rates
of ratum and variances offered by the different assets in the eoonomy The sourCe of riskiness in the
ampConory comes from an uncertain marginal product of capital On the other hand aside from
portfolio decisions the res of the economy is insulated from uncertainty
The most complete contribution in terms of the treatment of the corporate financial rules is
FulienonmiddotGordon (1983) In a variant 01 Ihe BailardmiddotFulienon-ShovenmiddotWhelley (1985) model
Fullerton and Gordon have capital intensity and optimal financial decisions Jointly deter~ined
through a two siage process The cosl of financing capital is minimized by trading off the tax
advantages of debt against the epected real bankrupcy coSls inherenl with high debVequity ratios
Given the optimal debtequity ratio the level of inveSlment Is chosen such that at Ihe margin the
24 l
bull
rerurn pn ~uity equas the return on bonds plus an exogenous risk premium
36 Market assumptions equilibrium and expectations
Virtually all of the papers arc characterized by Walrasian market clearing assumptions All
markets are perlect1y competi1ive Atomistic competition among agents 1s assumed even thoughmiddot only
a flnile number of agents is considered Virtually no mark~t disequilibria or price stickiness afe
considered The only exception is ErlichmiddotGinsburghmiddotHeyden (1987) bull In thei model of the Belgium
economy I the _wage rate is fixed in the short-run Therefore in the shOrlwrun desequilibrium in
the labor market will generate endogenous unemployment However in lh$ long-run all prices
Including the wage rata are Ilexible and accordingly ali markets ciear
Given the dynamic nature of behavlor in the economy markelmiddotcfearing prices in each period
depend on expectations of fture prices and on tax variables In the economy There are essentially
two ways at interpreting the economic equilibrium in such a dynamic conte~bull If future prices are
perfectly anticipated (io expectations are self-fulfilling) a perfect foresight equilibrium
prevaHs Then future actions are merely the implementation of current decisions for future
periods However jf ptica expecla110ns are not perfect (Le agents make mistakes w1th re$pect 10
future prices) then a temporary or shortrun equilibrium prevails Markets ciear and clearing
prices depend on fu1ure price expectations Current plans about the future are typically not
precisely implemented They will be revised as more or better information becomes available to the
economic agents See Grandmant (1982) for a comprehensive survey 01 the temporary equilibrium
literature
Wilh the exception of Gould (1985) and Pereira (19S6b 19S7d) all the models surveyed
adopt th~ concept of a perfect foresight equilibrium In turn BaliardmiddotGouldr (1965) considers a
J
2S
flexible am9unt of foresight In terms of the number of years over which price movements are ~
foreseen Pereiras model (1986b 1987b)) is flexible in a somewhat different way in that it can
indude any range of foresight from myopia to perfect foresight The choice between the perfct foresight and lemporary equilibrium is ultimately to be made on
philosophic grounds It can be argued that less than perfect expectations Imply that agents are
irralional in some way (see Auerbach-Kctiikoff (1987 p 10) However the reverse argument can
be made One can question wiether agents are reaHy rationar and perfectly knowledgable about
future prices
Recent evidene of Balird (1987) Ballard-Goulder (19851 Goulder (19851 and Pereira
(1986b 1987d) suggests that the choice in modemng expectations is an important one They show
that the degree of foresight inl0 thc future (ranging from perfect foresight to-myopic expectations)
may have dramatic impacts on the pollcy conclusions of Ihe model Accordingly the best research
strategy may b to design models which are flexible enough 10 allow for different rules regarding the
formation of expectations With the exception of the articles just mentioned sensitivity analysis
bullhave previously not been performed along this dimension
In terms of implementation the two concepts of equilibrium - perfect foresight and temporary
equilibrium have different implications The dimensionality of the equ1ibrium soll1ion algorIthm
is involved Suppose we have a model with ten markets to be run for a period of 50 years Aside
from normalization a perfect foresight model implies computing prices In 500 dimensions while a
-t~mporary equilibrium model requires sorving SO equilibria each in ten dimensions Given that
computational speed often varies with the cube of the number of dimensions the lemporary
equilibrium formulation is potentially strikingly more feasible However Ballard (1987) and
Goulder-Ballard (1985) have devoloped techniques to greatly speed the computation of a perfect
foresight equilibrium
26
j
I
4 Jrnplemenlalion and Issues on policy evaluallon
41 Calibration and equilibrium comparlslon
CGE models at typically parameterized by the use of a calibration procedure Some parameters
are exogenously given However some crucial parameters are determined in such a way that the
model replicates the data for a given base year See MansurmiddotWhalleY 11984) for an eXlensive
discussion of 1~i$ issue
CalibratIon in a dynamic context is generaliy interpreted as requiring two properties First
replication of base year data is required Second the model is parameterized to simulate an
intertemporal balanced growth path when Ihe base policy is maintaineq This Is the approach
followed by Ballard (1983) BallardmiddotGoulder (1985) Goulder (1985) and SummersmiddotGoulde
(1986) It follows the practice of BallardmiddotFuliertonmiddotShovenmiddotWhalley (1985) with their
sequential cquilibrium dynamics
Other authors Auerbach-Kotlikof (1983 1984 1987) Boyenberg (1984 1985 1986)
and Pereira (198Gb 1987d) follow whal we cali a qualitalive calibration The slructural
palametars are exogenously chosen so that lh economy follows a reasonable path into Ihe fulure
ThaI has 10 do wllh the fact thaI given the recursive nature 01 the dynamic economy and aside tom
such structural parameters only iniUa stock values are needed to run these models Given initial
conditions on the stocks of say private wealth capltal and government debt agents will optimize
and Ihereby generate a lirst round of net demands and equilibrium conditions In lurn the
equilibrium prices will determine the evolution of the stock variables into the next perIod
There are several potential problems with calibration in the conlext of dynamic models The
assumption of a steadymiddotstate growth ~ath in the base case can be questioned First while
27 I Ibullbull
steadymiddotstaU is a possibity it certainly is not the only meaningfull solution to dynamic models
Even in the case of a perfect foresight equilibrium the model implies an equilibrium path which i may or may not involve colanced growth The model not the modeller should dictate the nature 01 I themiddotbase case path Second in the context of a temporarY equilibrium path a steady state solutlcn is Inol a likely model QutcOn1tl In fact unlessmiddot expectations are stattc short run behavior consistent
with a steady-state evolution will in general not be generated On the other hand if static I expectations are self~fufjili[1g we have in fact a perfect foresight model Third even the base year
replication requirement may cause problems In Ihe cOntexl of temporary equilibrium Any
calibration parameter would be conditional on expectation rules which is probably an undesirable
feature
The nalur of the two-requirement calibration strategy is very muchin the spirit of the bull
traditional design of the comparision of alternative equilibria comparis_ion between a steady-state
base case on one hand and alternative paths including a transition period and a finat sleadyyenstate on
the other hand Pereira (198Gb 1987d) compare different (not necessarily steadystate)
equilibrium paths The arguments against such a procedure are based on the idea that the impact of
policy changes can be observed most easily since all departures from the steadymiddotstate can be
attributed to the altemalive polioy On the other hand the base case so dEfined as a steadymiddotstate is
consist~nl with previous work in a less dynamtc setting and Iherefore allows a common standard
for comparing model results
42 Computation algorithms
We are still al a stage in which basically each author uses a different computational technique
Th development of dynamic models - in parlicular wilh adjustment costs andior perfect foresight
bull
28
has corresponded with the decline in the use of fixed point algorithms In -fact given the relative
arge dimensions inevitably involved such algorithms lend to be very inefficient at the best and
chen prohibitively stow See Stone (1985) and Preckel (1965) for a comparative assessment of
different computation techniques Among the models surveyed only Ball1rdFulle(ton~
-ShavenmiddotWhalley (1985) and Feltensteln 1985 use Merrills variant of the fixed point
algorithm 1echnique
AuerbachmiddotKoHikoff (1963 1984 1967) follow a three stage procedure They first compute a
base case steady-state then a revised case steady state and finally transition path for the economy
berveen these two steady-states tn all stages a Gauss-Seidel iterative procedure is used
Ballard (1982) Ballard-Goulder (1965) Goulder (HiSS) and Summers-Goulder (1986)
~
use a method developed by Ballard and Goutder which is similar in many awects to the FairmiddotTaylQr bull
(1983) algorl1hm Short-run equilibria are calculated (using Merrills algorithm) parametric on
nrice expeC~lions The model is then iterated to generate self-fulfilling intertemporaf expectations
and the corresponding perfect foresight equilibrium In a relatively similar approach Andersson
1987 uses a simulation program SIMNON developed In the University of Lund This program
can handle two-paint boundary problems In a fashion consistent with the multiple shooting
algcrlthm (see Lipton-Poterba-Sachs-Eummers (1982))
Boyenbergs computational approach (1985 1985) differs from the other models surveyed in
that he relies heavily on analytical techniques Computations are done by using a dynamic version of
Johansons linearization method Being essentially determined by the continuous time nature of the
I I
f
model thiS linearization model has the disadvantage of confining the analysls to infinltesinal changes
around the base case equilibrium (see Bovenberg (1985) p 53) Pereira (1986b 1987d) uses an optimization algorithm NPSOL - developped by
rGillMurray-Saunders-Wrlght (1986)) This algorithm is used 10 compute the sequence of
I
29
Shortmiddotrun temporary equilibrium which makes up the intertemporal equilibrium path The
equilibrium conditions are seen as nonlinear equality constraints in the minimization of an artlcial
objective function The prices are normalized to the unit simplex by an additional linear equality I I 1 bull constraint
II ErlichmiddotGinsburgh-Heyden (1987) follow a unique approach in that they use a variant of the
optimization technique introduced by Negishl (1960) The economic equilibrium can be generated i
as a solution of a mathemallcal program Ihe objective function of which is a weighted sum of the
utility functions of the various agents while tho constraints set consists of the market dearing 1
conditions Ginsburgh-Heyden (1985) have extended Negishis resulllO tho case of downward price
lrigidities I
Finally the paper by Jorgonson-Yun (1984) is also unique in tIlat it is the only I
econometrically estimated model Different blocks for the consumption and production Side of the
-
model are separately estimated to provide the necessary structural parameters i
I The diversity of computation techniques is yet another indicator of the exploratory nature of the
body of literature surveyed in this article
43 Equal yield comparisions
The link between a base case and counteriactual simulations Is usually provided by the concept of
equal yield Shaven-Whaley (1977) discuss the meaning 01 equal yield in a general equilibrium
context when government is not allowed to run defICits Equal yield is interpreted to moan constant
public utility Government base case utility is maintained in the counferfactual experiments With
balanced budgets the concepl of equal yield is unambiguous The new equUibrium prices and the
balanced budge condition will del ermine the minimal expendilure and taxes needed to maintain base
30
case publi~ utility Accordingly in general equal yield is inconsistent with equal nominal tax
revenue Some change in tax revenue Is necessary Different tax replacement schemes ate
considered to assure that enough tax revenue is collected This is the approach essentially followed
by most 01 the papers surveyed Only Feltenstein (1985) Goulder (198) and Jorgenson-Yun
(1984) chose not to follow an equal yield strategy
The question is of how to Interprete lila concept of equal yield when the government is alowed to
fUn deflcLts The optimal leve of expenditure for base ease public trliUty can now be taAinanced
bond financed or financed by a mix of bonds and taxation We have several versions of equal yield In
particular equal yield may now be consistent with equal tax revenUE Furthermore some meaSure
ltgtf financial crowding out effects of government deficns can be inferred from the comparislon of the
several equal yield alternatives These issues are extensively discussed ~Pereira (1987b)
44 Price expectations and the dynamic generalization of compensation
Indicators
The sum of equivalent and compensation variations over households is the most widely used
aggregated measure of efficiency gains or losses In the context of $teady~state eomparisions andor
when complete future markets exist andlor perfect foresighl is assumed there are no difficulties
associated with the use of the standard Hicksian indicators In fact correct future prices are known
in these cases
If expectations are not selfmiddot fulfilling andlor future markels are not available a dynamic
generalization is necessary 8allardmiddotGoulder (1985) provide some steps in that direction by
defining an indicator that accomodates periods far into the future when households do not have
perfect foresight but a steady~state prevails On he other hand this issue is more fundamental in
i
[
I
i I I I I [
I I
bullI
31
the contex~ of a Gner1 lemporary equilibrium ftamawor~ In such Circumstances Pereira
(1986a) dev~lops a cnamlc generalization which is obtained as the present discounted value of a i i
sequence of short-fUn optimal expenditure functions consistent with a base case expected future I stream of IlWities I
32
5 Empirical evidenee from selected policy Issues
Dynamic tax models have generated several important results which escaped Sialic modellers
The following is a seleCpoundd set of issuos whIch stress the marginal benefits of dynam~ modelHng of
economic behavior in innovltlttive areas of analysis The discussion of a consumption tax emphasizes
the benefits of dynamic tusehold behavior and intergenerational aspects The study of the
flliminatlon or the re~intToduction of the investsnemt tax credits is made meaningful by the dynamic
modelling of production betlavor Modelling governmenl in a dynamic conlexts allows the modeller
to study the irrrpact of government deflcits Finally a dynamic framework lets us appreCiate the I
relevance of consumer expectations in terms of the evaluation of policy alternatives
i
i
51 Consumption Tax
t
By Its very nature a consumption tax can not be adequately investigated with a static model bull
Fullerion-Shoven-Whaicy (1983) use Ihe model of the US economy s described in
Ballard-Fullerton-Shoven-Whalley (1935) to evaluate the movement from the currenl US tax Isys1em to a progressive consumption tax Since their modeJ Incorporates a Jaborleisure choice
where leisure is an untaxed commodlty their results reflect the fact that both the consumption tax
and Ihe present system are dslortionary
Concentrating on the rntertemporaf distortions they show that sheltering more saving from lhe
current US income tax could improve econornic efficiency even if marginal tax rate increases are
necessary in order to maintain government fevenue~ At first you have a revenue shortfall
However the economy moves to a higher sustatned growth path and uitimatefy lower tax rates
cangenerate the same revenue path Also wages increase in the long run with this policy The
bull
bullbull ~
33 shyt
i
I
I
present value of welfare gi[lS for a potiey of complete savin9 deduction wilh marginal tate
adjustments (consumption lax) is simulated to be around $500 billion to $600 billion 01 1973
dollars
They also investiga1e th 1ngth of ttme it takes th~ economy to adjust to these policy changes
Roughly they estimate the long run to b thirty years although this figure Is very sensitive to the
specification of the savings elasticity - a crucial parameler in this model
52 Investment T~x Credit
Goulder-$ummers (1 S87) address the impact of eliminating the lnvestmenl Tax Cradit (ITC) on
intersectoral capital formation and on economic growth Their model isYa1ticularly adequate to
addies$ this issu~ in that is postulates a forward looking investment m9we with adjustment costs
They show that the eliminating the ITC causes a reduction in the rale of investment Inveslment is
estimated to fall by about 7 in the short-run and by about 12 in Ihe new long run steady state
In lurn a previous policy anncuncement lowers Ihe overall al1ractiveness of Investment and leads 10
a downward shift of the Investrrent profile On the other hand the combined effect of a revenue
neutral simultaneous efimiraron of the ITC and reduction of the corporate tax rates is a long run
reduction ef the capital Sieck by 35 This pattern suggests that a revenue neutral increament in
both Ihe corporale tax rates and the inveSlment lax credits would be preferable 10 the formula
adopted in Ihe US Tax Reform Act of 1986
Pereira (1987d) addresses the impact of the ITe from Ihe oland point of Ihe curTenllax syslem
under the Tax Reform Act of 1986 Given Ihe concern over the potential depressiVe effects upon
savings and investment of the new tax system in conjunction with deficit reduC1ion mechanisms of
the GrammmiddotRudman type a pertinent question is should Ih lTC be e-introduced In the context
bull
34
-of his mod~1 of the US economy Pereira focuses on the trade-off between the distortion in the
intersectoral allocation of capital induced by the lTC the lower tax revenues it generates and
potential financial crowding-out effects on one hand and the positive effects of lowering the relative
price of new capital goods on the other hand Preliminary results confirm the qualitative results in
Goulder-Summers (1987) suggestting a welfare gain of about 2 of the present value of GNP in
the best scenario of absence of replacement taxes Furthermore preliminary results show an
important time pattern to welfare gains with the average benefits increasing the further you look
into the future This pattern is due to the presence of constraints to intersectoral mobility of capital
and inertia in the adjustment towards the optimal capital levels as reflected by the presence of
adjustment costs
I
I
bullAuerbach-Katlikaff (1gB) in the context of their intergenerational model of the US economy
consider the impact on savings capital formation and interest rates of deficit policies and balanced
budget increases in government consumption They conclude that deficit finance and government
consumption can significa-ntly crowd out capital formation and lower the welfare of future
generations However crowding out from deficit finance is a very slow process because it results
from increased government spending over potentially long horizons Also deficit policies may
substantially influence the longmiddotrun interest rates while leaving the short-run rates essentially
unaffected Finally and in opposition to the central role adjustment costs seem to play in both
53 Financial crowding outeff~cts of government deficits
Goulder-Summers and Pereiras models Auerbach-Kotlikoff suggest that the time path of interest
rates induced by a policy of deficit financing seems to be insensitive to the presence of adjustment
costs
bull
35
54 The role 01 consumer expectatlons
The expectations analysis has uncovered one of the benefits of dynamic modelling
Ballard-Goulder (1985) find that the appeal of adopting a consumption tax depends on the level of
foresight possessed by the consumers Furthermore additional foresight may be welfare worsening
This is a second best type of result This tends 10 occur under policies that lead to capital deepening
and declining rate of return to capital over time To the extent that consumers have more foresight
they will be beller equippod to anticipate the fall in the rental price of capital Consequently if a
saving incentive is enacted people save more with myopia than with perfect foresight Given the
exi~tence of taxes on capita and the discrepancy between the private an~--sccial returns to capital
the greater savings with myopia is beller socially Ballard-Goulder find that the welfare gains
from a consumption tax are reduced by about 10 when we move from myopia to farleet foresight
Therefore the attractivenness of adopting a consumption tax seems fairly [obust across these
foresight specifications
I
bull
36
6 Tbe power and weaknesses 01 dynamic modelling the example 01 corporale
tax Integrallon
The Issue of corporate tax integration allows us to sh~w the stre-ngths and weaknesses or the
dynamic approach Empirical evidence using CGE techniques indicates that depending on the precise
scheme of integration and the 1ax replacement methods integratiOn may have substantial effects In
Ihe work of FuliertonmiddotKingmiddotShoven-Whalley (1980 1981) total integration was found to yield an bull
annual static eHielency gain of $4 to $8 billion in 1973 dollars Simulated dynamic gains may be as
large as $695 billion or abcu 14 of the present value of tulUre consumption and leisure in the
US economy These gains result primarily from interindustry reallocalions of Investment and an
improved inter~emporal allocation of consumption
Mora recent work emphasizes hOw consumers asset portfolio decisions and firms finanCial
decisions affect the efficiency gains from integration Slemrod (1980) focusing on consumers
asset ponfoio deCisions finds static efficiency gains which are about twice -as Iarge as those
repOrted iiy Fulierton-King-ShovenWhalley (1961) FuUrlon-Gurdon (1983) focus on the
firms financial decisions They roporl efficiency gains of 6 of GNP from the elimination of the
tax distor1ions favoring debt However when they eliminate the corporate tax and repiace it with
increaSed persona income taxes addlHonal dls10t1ions are created in the optimal labot~leisure
decisions These distortions tend to dominate the analysis slJ(h that the overall effects of complete
integration are very modest Galpr-LuckmiddotTodr (1986) address simultaneously consumers
asset pOrlfollo decisions and the firms financial decisions They also find very modest efficiency
gains from integralion
The above results are important but may be severely biased There are severa aspects of
economic behavior and modelling crucial for the study of Income tax Integration whiCh have no been
i 37
captured innny of the above papers One aspect is the absence of government deficitsmiddot a balanced
budget is assumed It is true Ihal resource crowding OUI is caplured In these mOdels but financial
crowding out induced by government spending is nOI Second investment is not derived from
optimtztlcn behavior Investment behavior passively accomodates endogenous saving decisions As
a consequence the differential impacts of policies in the incentives to save and to invest are not
captured Aso full capital mobility across secors is assumed with instantaneous capital
adjustmenlS towards optimal levels This assumption rules qut different costs of capital across
seclors and therefore differentiated reactions to tax policies changes Finallyhe modelling of both
gov9rnmerlt deficits and of endogenous real and financial investment decisions necessitates the
~nsideraiion of a dynamilt framework and the introduction of financiaJ assets govemment bonds and
iprivate financial assets A dynamic framework also highlights the efficiency effects of Integration on r
ithe ptimal intertemporal decisions The above models are either sIalic (Slemrods and
GalpermiddotltJckemiddotTode(sj or a dynamic sequ~nce of otherwise Slatic model bullbull I Pereir a (1986b) develops a dynamic applied general equilibrium model bull to study the tbull
efficiency effects of integration on the growth and allocation of investmenl across sectors Special
Iattention is paid to the structural affeelS of resource and financial crowding oul induced by
gove mect spending and deficits and to the real and financial Investment reactions across industries
to both tax integration and flnancial crowding out His mode departs from previous work on income
lax Integraticn and for that matler from most 01 the CGIE literalure for tax policy evaluation in
several direcllons It encompasses an endogenous sequential equilibrium struelure founded on
dynamic behavior with flexible expectations Government deficits are optlmaUy determined
Investment decisions are tward looking and the resull of optimizing behavior Several financial
assetsmiddot public and privatemiddot are considered
Simulation resulls Indicale that the welfare gains from Integration under large deficits are at
32
best lJlodest when measured in terms of the GNP The elimination of the corporate tax and its I replacemenl by increased income lax rates yields long run benefits which are never larger than
2 of the presenl yalue ct telr consumption and leisure onder the previous tax regime and 1
under the current tax regime However the short run w~lare effects of full integration tend 10 be
very low and in in some scenarios even negative This is a flew intertemporar pattern of efficiency Ieffects which reflects an adjJsiment lag in the interindustry investment decisions due to the
existence of costs of adjustment On the of her hand partial integration achieved by excluding I dividends from the corporale lax base syslematically yields negalive efficlencyeffeclS In Ihe long Inm these can be as high as 3 of present value of the future value of consumption and leisure This
is a new second best effect suggesllng Ihat less than complete integration may have perverse I efficiency effects
The dynamic slructure of PerclrBs model snowing for an improved freatment of real investment
declsions and government deficts provides a potential setting for a more accurate measure of the
costs and benefits of inlegration The simulation resullS suggest lower benefits and an intertempora Ibull
pattern of growing benefits
Simulation results also clearly suggest robust negative effects from partial integration How
reliable is this result If dJvidends were deductible from the corporate lax base (as are interest
payments) the preferenlial Ireatmant of debl over equily would be ellminaled Corporallons should
be expecled 10 react by decreasing the optimal deWeqully rallo However corporale financial
gecisions are exogenously set In Pereiras model as in aft of the other models surveyed that go as far
as dealing with the issue
Also withdrawing dividends from Ihe corporate tax base is a way of eliminating Ihe double
taxation of dividends at belh Ihe corporate and personnal Income levels How will the optimal
aliocellon of household saving accomodate that change in Ihe relalive return to corporale equity
39 I
t However hOusehold portfolio decisions are exogenouly set in Pereiras model as in all of the other
models surveyeltj go as far as dealing with the issue
IIt is sate to say that the evaluation 01 the benefits 01 partial Integration as defined above are
sevrey underestima1eO Meanwhile the distortionSoenerated by the increase in the marginal
personal tax rates designed to ralse the revenue foregone by the dividend exclusion are fully
aCCltlunted for A good measure of the costs of integration together With a poor evaluation of the
benefils may very well be the reason why partial integration is Simulated 10 yield neg alive
eHikiency gains
On a different vein as most of Ihe model surveyed here Pereiras is a closed economy model It
is legitimate to wonder how the resullS would change If international capital flows were anowed
IThis almost certainly would affect financial crowding out since a good d~al of the capital inflows
could be used to finance public debt
For these various reasons the results should be treated as preliminary but they strongly I suggosl that the dynamic structure provides valuable new insights into the corporate tax integration SSU2
40
7 Concluding remarks I I I i
What has been acccmplished Ihis far The success of Ihe CGE research in laelding the challenge
of dynamic modelling can not be denied
I iGreat progress in the modelling of household behavior has been achieved Promising
developm~nts in the modeUing of production and government behavior have also been accomplished
Interesting innovations in the concept of equilibrium and computation techniques were discussed
The policy analysis was greatly enriched by considering transitional effects together with longer
run steadymiddotstate equilibrium paths generalized equal yield strategies accomodating different
financial crowding out impacts and generalized dynamic policy evaluation indicators
An important set of issues have been addresed by the models surveyed Traditional issues
focusing on the effects of capital taxation and consumption taxes have been pursued In terms of
capital taxation for example the intertemporal nature of the issue was enhanced by allowing an
optimal evolution of capital stock in the economy and forward looking optimal investment decisions
In Andersson (1987) GouldermiddotSummers (1987) and Pereira (1986b 1987d) this is coupled
with a improved treatment of several tax provisions like investment tax credits and depreciation
allowances In terms of the consumption taxes developments in the specification of Intertemporal
household behavior the introduction of overlapping generations and the modelling of
intergenerational links via bequest motives as well as a closer attention to demographic evolution
provided a much improved economic setting for the understanding of the several aspects of the
problem
Furthermore dynamic modelling has permitted the addressing of a set of new issues Policy
Issues ranging from dynamic tax policy analysis to the evaluation of exogenous changes in
government expenditures to the impact of different methods of financing government expenditure to
4 1
the importance of tinanciai croHrii1g out to 1he relevance of consumorS expectations in tne policy
results have been addressed
Despile all of the progress and in part due to such progress several avenues are wide open for
much neded additional research The following seem to be the most Interesting and promising
areasmiddot First the specincation of liquidity constraints to household behavior may help to obtain a
better understanding of policy chenges that affect directly the interest rates in the economy Second
major attention needs to be paid to the incorporafion of a roraign sector (with open capital markets)
Third some efforts are needed in terms of finding adequate computation techniques now that
dynamic modelling eally makes the curse of dimensionality worse than eVOrbull
The modelling of financial markets is the single most unsatisfactory speet of the dynamic
models surveyed here~ The problems associated with provlding a mo~ adequate treatment of
financial markets and in general endogenous and optimal financial deci$ions bull both at the corporate
and at the personal levelsmiddot have 10 be addressed That necessarily requires Introducing uncertainty
into the models To b adequate this must go well beyond the mere assumption of point expeetalions
with the whole array of modelling implementalion and evaluation problems thai generates
bull I i
42
REFERENCES
Abel A and O Blanchard 1983 An Intertemporal Model of Saving and Interest Econometrica
51 pp 675-92
~ Andersson K 1987 Tar-ation of Capital in Sweden - A General Equilibrium Model Ph D
Dissenation UnIversity of Lund Sweder
Auerbach A L KotlikoH 1983 National Savings Economic Welfare and the Structure of
Taxation in M Feldstein ed Behavioral Simulation Methods in Tax Policy- Analysis Chicago
University of Chicago Press
Auerbach A l Kotlikoff 1984 Social Security and the Economics of Demographic-
Transition in H Aaron and G Burtless eds Retirement and Economic Behavior Washington DC
Brookings Institution
Auerbach A L Kotlikoff 1987 Dynamic Fiscal Policy Cambridge Cambridge University
Press
Auerbach A L Kotlikof and J Skinner 1983 The Efficiency Gains from Dynamic Tax
Reform Internatiooaj Economic Reyiew 24 pp 81middot100
Ballard C 1983 Evaluation of the Consumption Tax with Dynamic General Equilibrium
Models Ph D Dissertation Stanford University
Ballard C 1987 Tax Policy and Consumer Foresight A General Equilibrium Simulation
Study Economic Inquiry 25 pp 267-284
Ballard C and L Goulder 1985 Consumption Taxes ForeSight and Welfare a Computatonal
General Equilibrium Analysis in J Piggott and J Whalley eds New Deyelooments in Apolied
General Equilibrium Analysis Cambridge Cambridge University Press
Ballard C D Fullerton J Shoven and J Whalley 1985 A General Equilibrium Model for
t 43
Tax Eoliey Eyalua~iQO NBER University or Chicago Press
Borges A (1987) bull A Survey of Compula1ional General Equilibnum Models CECD Economic
Slud~ 8
Bovenbig L 19B3 imperfect Capital Mobility in a Perfect Foresight Herberger Model
Mimea UC Br~eley
Bovenberg L 1984 Capitel Accumulation and Capital Immobility Qmiddottheory In a Dynamic General Equilibrium Framework PhD dissertation University of California at Berkeley t
Bovenberg L 19850 Dynamic General Equilibrium Tax Models with Adjustment Costs in A iManne ed bull 1985 EtQnomrc ECUjilbrium Model Formulation and Solution Mathematical
lProgramming Siudy 23 Arnsterdam NorthmiddotHoliand
Bovenberg L t 1986 Capital Income Taxation in Growing Open Eeondmies Journa of Pubtic
Eamp(omlcs 3t pp 347middot376 IBovenberg L ald W Keller 1981 Dynamics and Nonlinearlties in Applied General
Equilibrium Models Internal Report Mlmiddot81middot208 Department for Statistical Methods CBS bull
Vooiburg The Netherlands
Charnley C 1981 The Welfare Cost of C~pltal income Taxation In a Growing Economy
IJournal of Political ECQnQm~ 89 3 pp 468middot496 I
Charnley C 19810 Efficient Tax Reform in a Dynamic Model of General Equilibrium I Mimeo Yale University I Charnley C 1982 On the Optimal Taxation In Stylized Models 01 Inlartemporal General
Equilibrium Mimeo Yale University I Decaluw B and A Manens 1985 Pays en Oeveloppement at Modeles Calculables OEquilibre
General - Une Revue de al litterature empirique Monographie Sene 9 Volume 3 Centre de
Recherche et Developpemont Economlque Unrslt de Montreal
44
bull
I I
I
I
1
Erlich $ V Ginsburgh and L van der Heyden 1987 Wher~ do Real Wage Policies Lead
Belgium bull A General Equilibrium Analysis European Economic Reyiew (forthcoming)
air R and J Taylor 1983 Solution and Maximum Likelihood Estimation of Dynamic
Nonlinear Ratioant Expeclations Models Ecooometrica 51 pp 1169middot1186
Feltenstein A 1984 Money and Bonds In a Disaggregated Ec~nonYin Searl 1-1 J Shoven
Eds Aoolied ~eoeral EqJiliDCiulD tioalYsis Cambridge University Press Cambridge
Feltenstein A 19B6 A Dynamic General Equllibrlu Analysis of Financial Crowding Out
Theory with an Application to Australia Journal of Public Economics 31 No1
Fullerlon D R Gorden 1983 A Reexamination of Tax Distorllons in General Equilibrium
Models in M Feldstein Ed Behayora Simulation Methods in Tax pOlicy Analysis Chicago
University of Chipgo Press
Fvllerlon D J Shaven and J Whalley 1983 Replacing the US Income Tax with a
Progressive Consumption Tax JoulOal of Public ECQnomics 20 pp3middot23
Fullerton D Y Henderson J Shaven 1984 A COmparision of Methodologies in Empirical
General Equilibrium Models of Taxation in Searl H J Shoven Eds ~pQlied Geoaral Jguilibrium
lIoalyss Cambridge University Press Cambridge
Fullerton 0 A King J Shaven and J Whaney 1981 Corporate Tax Integration in the United
Statesmiddot a General Equiiibdufi1 Approach American Economic Review 71 pp~ 677~691
Galper H R Lucke E Tader 1985 Taxation portlor Choice and the Allocation of Capital A
General Equilibrium Approach Mimeo
Gill p W Murray M Saunders M Wright 1986 Users Guide for NPSOL (Version 40) A
Fortran Package for Nonlinear Programming T echnicsl Report SOL 86middot2 Deparlment of
Operations Research Stanford University
Glnsbur~h V and L van der Heyden 1985 General Equilibrium with Wage Rigidities An
45
AppJication to Belgium in A Manne ed t Economic EQumprit1m~ Made Formulation and $oJioo
Mathematical Programming Study 23 Amsterdam NorthmiddotHolland ~
Goulet J 1968 1 Adjustrent CO$1S in the Theory o(the Firm Reyiew of EconomiC Studies 35 bull
pp47middot55
Goulder L 1985 Tax-Financed versus BondmiddotFinancedGhanges in Governent Spending
bull Mimeo Harva(d UniversilYmiddot
Goulder l J Shaven and J Whalley 1983 Domestic Tax Policy and the Foreign Sector The
importance of altemative foreign policy formulations to results from a general equilibrium tax
analysis model In ~n methods in Ibe antico of income from caoia1 ed Martin
Feldstein Chicago University of Chicago riSS
Goulder l L Summers 1987 Tax Policy Asset Prices and 3rowtly ~
A General Equilibrium Analysis NBER Working Pper No 212B
Grandmont J 1982 Temporary General Equilibrium Theory Ch 20 in Handbook 01
Mathemalical Eccnornics Vol II NorthmiddotHoliand
Harberger A 1959 The Corporate Income Tax An empirical Appraisal In lax Revision
Comoendjum Voll House Cornmitee on Ways and Means Washington DC Govemment Printing
Office
Harberger A 1962 The Incidence of the Corporate Income Tax JOULn1 of Political
Economy 70 pp 215middot240
Harberger A 1964 Taxation Resource Aliocalrlon and Welfare fn lb Role of Dir bullbulll and
Indjresl Taxes io tbe Fede[al Reserve s~I~Ill Princeton Princeton Unlveresity Press
James J 1985 New Developments in the Applications of General Equilibrium models to
Economic History in Piggott aand Whalley Eds
Jorgenson D and K Yun 1984 Tax Policy and Capital Allocation Harvard Inslitute of
bull
I
I f
Economic Research WP 1107
Judd K 1965 ShonmiddotRun Analysis of Fiscal Policy in a Simple Perfect Foresight Model
Journal of poJiHcal 5CCJiW1l pp 298middot319 r
Upton D J Potbc J Sachsa nd L Summers 1982 Multiple Shooting in Rational
Expectations Models fuoometrlc 50 1329middot1333
Manne A ed 1985 fconQrnjc Eoullibrium Model FOrIDujaljQo and Solution Mathematical
Pogrammlng Study 23 Amsterdam NorthmiddotHolland
Manne A 1985 Or the Formulation and Solution 01 Economic Equilibrium Models In A
Manne ed ECQOQlli-Jr~ibmiddotum ~tQdel EQlJulailon and Solution Malhematical Programming
Study 23 Amsterdam NorthmiddotHolland
Mansur A and J Whalley 1984 Numerical Specification of Applied General Equilibrium
Models in Scarf H J Shoven Eds Aoplied General Eouillbrium AnaiysectIsect Cambridge Universily
Press Cambridge
Merrill 0 1972 Applications and Extensions of an Algorithm thaI Computes FixedmiddotPoints of
Cenain Upper Semimiddotcominuou Point to Set Mappings PhD Disserlation Unlvel$ity 01 Michigan
Negishi T 1960 Welfar Economics and Exislence I an Equnlbrlum for a Competitive
Economy MelroeCOOQ11ka 12 92middot97
Pereira A 1985 On the Existence and Uniqueness of Optimal Ou1put Path lor CRTS Firms in
Adjustment Costs Technologies Mimeo Stanford University
Pereira A 1985a Consumer Expectations and Ihe COSI 01 UllUty In an Intertemporat
Framework Mlmeo Stanford University
Pereira A 1985b Corporale Tax Integrallon and Government Deficits A Dynamic General
Equilibrium Anaiysls Mlmeo Stanford University
Pereir A 1987 On Ihe Computation of the Users Costs of Stocks Ecooomi (forthcoming)
47
Pereira A bull 1987b Equal Yield Alternatives and Government Oeficlts Mimeof Stanford
University
Pereira A 1987c A Dynamic Applied General ~uHlbrium Model for Tax Policy Evaluation
Comlletamp Docurnent2tion Mimeo University of California San Diego (in progress)
Pereira A 1987d Should the investment Tax Credit be Rei(ltroduced Mlmeo University
of California San Diego (ir progreSs)
Piggott J and J Whalley eds 1985 tJew QeyeQpments in Applied General EQuilibrium
bullenalysjs Cambridge Cambridge University Press
Preckel Pbull 1985 Alternative algorithms for computing economic equilibria In A Manne Ed
Radnor R 19amp3 Equllibrium under Uncertainty In K Arrow M Intrilligaor ods
I ibull
Robinson S 1 g86 Mullisectorai Models of Oeveloping CountrieS a Survey Oivision of
Agricullure and Resources UC Borkeley WP 401
Scarf H 1967 The approximation of fixed points of a continuous mapping ~MM Jouwal of
epplid Mathornallcs 15 pp 1328middot43
Scarf H with T Hansen 1973 00 the Computation of EconomiC Eguilibria New Haven Yale
University Press
Scarf H J Shoven Eds 1984 61l1llied Genera Elujlibrium 6nalyssect Cambridge University
~ Press Cambridge
Shoven J 1976 The Incidence and Efficiency Effects of Taxe on Income from Capital
Jurnal of ramie1 Economy 84 pp 1261-83
Shoven J 1983 Applied General Equilibrium Tax Modemng IMF Slaff PaoerS 30
Shoven J and LB Simon 1986 Share Repurchases and Acquisitions An Analysis of which
Firms Participate NBER Working Paper No 2243
bull
48
IShaven J and J Whalley 1972 ft General Equmbrium Calculation of lhe Effects of
Differential Taxation of Incoma from Capital in the US Jauroe of public Econgmics 1
pp281middot321 (
Shoven J bull and J Whalley 1973 General Equilibrium with Tax$s a computation procedure
and an existence pr00f ReviEhi of Economic Studies 40 pp 475middot490
Shoven J J Whalley 1977 Equal Yield Tax Alternatives A General Equilibrium
Computa~ui0nal Tfchnique J~HJrnal of PUblic EconomIcs 8 pp 211~224 bull
Shaven J J Whalley 1984 Applied GeneralmiddotEquilibrlum MOdels of Taxalion and
International Tr8de An Introduction and Survey Journal of Economic literature 23t pp
1007middot1051
Slemrod J 1980 A General Equilibrium Model of Capital Income Taxation Ph D bullDissertation Harvard Univnrsity
Slemrod J 1983 A General Equilibrium Model of Taxation with Endogenous Financial
Behavior in M Feldstein edbull Behavioral Simulation Methods In Tax Policy Analysis Chicago
UniVerSity of Chicago Press
Stone J 1985 Sequential Optimization and Complementarily Techniques for computing
Economic equilibrium in A Manne Ed
Summers L 1981 Taxation and Corporate Investment a q-Theory Approach BroQkinas
Summers L 1985 Taxation and the Size and ComposHion 01 the Capital Stock An Asset Price
Approach Mlmeo Harvard University
Whalley J 1975 A Geneal Equiloibrium Assessment of the 1973 United Kingdom tax
reform Economica 42 pp 139middot61
1JLTIllOS JIQll~)ltG PAPERS PUBLICADOS bull
I
I I
j
112 70 - LUCENA iJiCIIJo ~ IIEnvironrilent Honitoring and Orqanization Structure T (Setellibro 1987)
nI 71 - L~ENA Diogo A Note on the Representation of Information Stxueture (SeteJObro 1987)
nQ 72 - LUCENA D1CgO Envuorment 111)nitoring and Organization
Structure II (Setelilhro 1987)
ril 73 - AlmmE3 Antonio Pais e CAsPAR Yitor bull Tributacao Inoentivos e InvesUlllentos Analise Qualitativa noYellibro 19S7)
n2 74 - liAlltOSA rose Pedro Opt1lllal lIage R1gidity A SUggestedtlthodology to Teet tb Theory and an Application (Outubro 1987)
nI 75 - CABRAL L 11 B Three lIotes on SYJIIlletric Gemes witt itsYJIIlletric Equilibri (NoveJObro 1987)
n2 76 - LUCENlt D10gOTo Search or Not to Search (Fevereiro1967)
nQ 17 - SA Jorge asconcelos euroI A Theory of SynergyB (Fevereiro1987)
n2 76 - VILARES tanuei Jose Os Be IntenaMios IlIIportados ColIIO Factor de Produao (Julho 1987)
nQ 79 - SA J~rge VasCQncel03 e HoY to Copete and Co~unicate in tature Industrial Products_ (aio 1988)
n2 80 - ROIl Rafael Learning and Capacity Expansion in a New larket Under Uncertainty (Fevereiro 1988)
n2 81 PEREIRA Alfredo llarvilo Survey of Dynamic Computational General EquilibriUlll Hodel for Tax Policy Evaluation (Outubro 1987)
Qualquer in ormacao sabre o 1IIorlltlng Papers ja publlcados sera prestada pelo Secretariado de Apoio aos Docentes podendo os ~e3OSf
ser adquirido5 ne Secyo de Vlndas da Faculdade de EconollLia UNL na Travessa Esteviio Pinto Cltpolide - 1000 LISllOA
bull
13
respecl to tgte rbullbull1 aoer lax wage rae and the elasticity of saving with respecl 10 the real after lax- rate of retum to capital
In the model consumerS have myopic expectations regarding lulure prices and in particular
regarding the future rale of return 10 capilal Ballard (1983) and BallardmiddotGoulder (1985)
inco~ora1e both perfect foresight and limited foresight into this model Future consumption is
acquired by buying a fixed composition portfolio of real investments that offer an infinite annuity
of returns
The production side of Ihe model is compielely static The mod1 Incorporates a constantmiddot
olasicily 01 substitution betweeen prlmary inputs in production (capital and labor) and fixed I coefficients for intermediate inputs The model distinguIshes between Industrial outputs and
eonsumer goods for the simple reason Ihat tile data are classified differenUy Industrial sectol$ I I involve such categories as forestry and fisherle metal mino9 and publishing and printing while
consumers purchas-3 fUfHi1ure 2ulomoblles and books This fact is recognized in the model by Ibull
incorporatirg a secord stage Of production which converts industrial outputs intO consumer goods
This technology is uStally mod9-led as a flxed~coefficient conversion matrix
The BallardmiddotFINton-ShovenWhalley model assumeS that the private sector finance
marginai ioveslment with the same composition of debt and equity as currently exists in each sector
This is the same arsumrtion that Herberger originally used Investors all hold debt and equity in tile
same proportion Thampreiore this ownership can be aggregated into simply capital ownership For
tax purposes however the separate treatment of debt and equity is taken into account at both the
corporate and persofa Io=vel SimllarlYJ the dividend poIicies on corporale equitY are established
exogenously There are no government bonds in the model since there are no government deficits
The modol is solved for a sequence (as many as 100) of temporary equilibria wltll consumers
allocating income between present and future conumption at each point in time The palh for 1he
14
economy ~ a sel of connected equilibria The connectIon is pra~ided by capital accumulation
Capital accumulation is endogenous and determined by saving The model stans with a sociagt
accOunting mi1rix In the bzse case the economy is assumed to be in a steaqy state growth -path (along
which all relative prices ~fe constant) The madel solves for both the naw steady state growth path
and the transltion 10 it zf~er a policy intervention rhe authors have frequently addressed the
question-of howong rt tKes to effectively settle un10 a new steady S1ate growth path
The dynamics of the model are limited however in that future consumption is collapsed into a
composite commodity Aso the absence of government deficits and the lack ofproductlon dynamics
limits the realism of the BallardmiddotFuliertonmiddotShoven-Whatley model for the analysis of dynamic
policy issues (such as the adoption 01 a consumption tax or Ihe elimination of the investment tax
credit) The models we survey try to Improve upon Ihe Ballard-FlIiertonmiddotShaven-Whalley
dynamic modelling The models Jn~orporate more forward lookTng behavior They improve the
dynamics on the conampumrtkm side of the economy and some of them introduce true 1ynamlc behavior
on the part oI the producers We turn now to the issues inY9lv_ed in developing such dynamic models
15 I I ~3 ~ssues in the modelling Of dynamic behavior
We will 100k firt at the speclficalion of economic behavior of consume(s~ producers
goyemment and Ihe r~sl of Ihe world The modemog of corporale and household financial decisions
I is then addr5sed Finally Ihe eencept of equilibrium is discussed
I
i
I 1
31 Consumers behavior
Early efforts 10 build dynamic features into the economic behavior 01 consumers are due to
Ballard (1963) and Auerbach-Kotllkoff (1983 1984) Their work has been closely followed by
several subsequent authors In faet dynamic household behavior Is the single most pervasive aspect
amohg the models surveyed
All the models Inc-O1orale some loro of lila-cycle behavior Household behavior is dewmloed
by the maximization of an addWVely separable time invariant intertemporal utillty function The
utility lune1ion is defined oyer the domain of the consumption goods In the economy In most of the
models leisure is also an argument in utility so that labor supply is optimally determined
Utility maximi4ation is subject to a lifetime intertemporal budget constraint which equalizes
the presen1 value of consumers income and expenditure More reeentiy in Andersson (1987) bull
Bovenberg (1985) and Pereira (1986b 1987d) the constraint i$ defined as a sequence 01
recursive equations of motion on wealth This has the potential advantage 01 accomodating liquidity
constraints However It should be recognized that in the absence 01 liquidity constraints (ie when
consumers are free to completely borrow against future income) the two specificallons of the
househofd ccnstraint are essentiay equivatent Furthermore tn both versions saving is optimally
determined as a way of translaring wealth intertemporally
1 6
Some lodels now have a sophisticated dynamic specification of hoosehokl In Ballard (1983) -
BallardmiddotGoulder (1985) and Auerbach-Kotlikof (1983 1984 1987) consumers behavio is
embedded in an iterQEneraional setting Each of them has S5 age cohorts simulataneously alive
Som-e measure of jntergener1ional altruism is considered in the form of bequest motives In Ballard
(1983) and Ballard-Gouldr (1985) consumers derive utility directly from bequthing parl of
their wealth Therefore bequests assume 1he form of a scrap value of terminal wealth of a
generation In AuerbachmiddotKotlikof (1983 1984 1987) each generation empathyz bullbull with Ihe
utility of future generations ovar bequethd wealth In addition Ballard (1ge3) includes detailed
demographic poiections wfill the intent of incorporating in the policy analysis the effects of the
bull Post World War tI baby-boom
In the real world household decisions also include the optimat allo~atiOn of saving among
shyalternative physical or financial assets Theories of household portfolio behavior are relatively
Iwen~establshed However they involve uncertainty as a crucial element Accordingly they have
middotnot been incorpora~ed in the context of the determinIstic dynamic models we sutyey here We will 1 Come backt-O this Issue below
32 Producer behavior
The efforts to build dynamic features into the economic behavior of producers are more recellt
and less widely adopted The first attempts are due to Bovnberg (1984 1985) and Summers
(1965) Dynamic behavior has been more fully incorporated In the recent models of Andersson
(1987) Auerbach-Kotlikoff (19B7) GouldermiddotSummers (1987) and Pereira (1986b 1987d)
Part of the reason why production side dinamlcs has been more slowty adopted is the weak
supply of accepted theories referring to the dynamic behavior of the firms In the models referred
I 17
to above gynamlc producion and investment b~havior are middotinduced by the existence of capital
adjustment costs and linked to Tobins q theory Adjuslment costs are designed to capture both the
incomplete mobility of capital aoross industries and installation costs ~e the costs of adjusting
capital towards its cpHmai level)
Adjllstment costs can be conceived as internal to the firm and measured in terms of foregone
outpul along the lines of Lucas (1967) Alternatively adjustment costs can be viewed as actual
external cests incurred together with Ihe purchase costs along the Unes of Gould (1969) With the
exception of Pereira (198Gb 1937d) all of the CGE authors follow the former approach
The firms in the economy maximize their market value as the present discountad value of the
future stream of dividends In Andersson (1981) and Pereira (198Gb 1987d) firms are seen as
maximizing the present discounted value of net cash flow Maximization is constrained by Ihe
adjustmenl cosl lechnology and ~n equation of motion describing Ihe evolution of the capital stock
It should be noted that undereogenous divldendrelention rules the two problems are equivalent II
should also be nOled Ihat a salisfectory economic rational for th bullbullistoo of dMdonds with the
present tax code is missing in the profession (Shoven (1986))
The models with static formulation of producers behavior are characterized by passive
investment behavior Investment merely accomodates to saving in the eccnomy With a dynamic
formulation induced b adjustment costs real investmen decisions are forward looking Investmenl
Is endogenously and optimally determined by the firms A fundamental difference between the
shonmiddotrun in which capital stock is given and longmiddotrun in which Ihe level of capital is allowed to be
optimally determined is emphasized
This extra richness of produc1ion dynamics is not without costs A careful look at the summary
tables will clearly show an inverse relation between the adoptlon of dynamic features in production
and the level of disagreggation of the production side of the economy In fact with production
bull
j
1 6
dynamics $e dimension of the problems is immensly increased Let IS be more specific
In a static framework with constant returns to scale production technology the output level ismiddot
indeterminate The optimal aliocation of inputs can be obtained by cost minimization with any
feasible output level generatillg zero profits Such zero profit conditions are used to solve for the
output prices in terms of the factor prices and hencemiddot to reduce the dlmensionality of the problem
from the -number of commodities and factors to the number of factors Thus even if the model deals
with 30 production sectors the computation of an economic equilibrium can take place using only
the dimensionality of the primary inputs in the economy
Now under certain regularity conditions on the production and adjustment costs technologies
leading to enough concavity of the optimality objective the intertemporal output path for the firm is
endogenously optimally and uniquely determined even with constant relums to scale technologies
(see Pereira (1985 1987a) Accordin~y with adjustment costs in the model optimal profits
will in general be non-zero This result has crucial implications for the computatiQn of equilibrium
in the model~ with production dynamics With adjustment costs no reduction of dimensionality is
possible The curse of dimensionality returns as a binding constraint
The introduction of adjustment costs adds another significant complication to the model With
capital being less that perfectly mobile in the economy different rates of return on capital will
exist in different sectors This is a difficult problem to tackle conceptually in the absence of
uncertainty Also in the real world producer maximization choices include also its choice of
financial ratios (debVequity) and payout rates (dividendretained earnings) These financial
subjects are difficult and much Clf this behavior is still taken as exogenous by the modellers We
will come back to these issues below
middot i
19
bull 33 government behavior
middot
Two issues dominate the modelling of government behavIor First govemment behavior has
Ibeen typically seen in Ihe CGE liIerature for lax poliCY evaluation as constrained by yearly balanced
I budsets (see for example 8allard-Fullrton-S~ovn-Whafley 098Sraquo The analysis of
govemment defidlS and public debl in a CelE context requires a dynamic selling Second the level of
government expenditures is either exegenouty given as in Auerbach-KoUikofi (1984 1987)
6ovenbrg (1984 1985) Feltenstein (1984 1986) and Jorgenson-Yun (1984) or
endogenously (but no optimally) determined by the balanced budget conditions as in
6allard-Fullerton-Shoven-Whalley (1985) Andersson (1987) Erllch-Ginshurgh-Heyden
~
(1987) Gouder (19aS) and Gouder-Summers (1987) In the second case the composition of
bull public expenditures is often optimally determined However the leve~ of government expendi1ures
can only 00 endogenously and optimally determined if the government is seen as an optimizing agent
and is allowed to run deficits
The first attempts to deal wiU the government defiCits in CGE tax models are due to
Auerbach-Kollilltoff (1984) and Feensain (1984) In Auerbach-Kotlikof (1984) government
expenditures afe exogenous They grow at the rate of growth of population However given 1he tax
structure and tax revenues yearly deficits and surpluses are atTowed subject to an in1ertempond
constraint that the present value of future tax revenues equals the present value of future
expenditures
In Fellenstein (1984 1986) government expenditures are also exogenously given He also
allows the government to run deficits to finance expenditures in excess to 1ax revenues However
surpluses are returned to CQnsumars in the form cf transfers Accordingly government ts not
subject to any constraint regarding the future repayment of public debt
In GouJder (19851 the 9Vernmenl maximizes a static social welfare ~nctio subject to a-
bull
20 I i
I
I I I
I bull
balanced budget conSirtlL This follows the optimal alocatkm of government expenditures alpng
the lines of Baliard-FHeiton-Shoven-Whalley (1985) Ttle model Is generallzea to allow for
exogenous changes in HIe time path of government expenditures Two financing alternatives are
consjdeft~d and contr8s1ed EldditionaJ tax revenues and bond issuance
Pereira (198Gb 1137d) auempts to address both the incorporation of deficits and the
determination of govement expenditures The path of government expenditures and the path of
delicilSlsurpluses (ano therefore the path lor debt) are endogenously and optimally determined The
government is seen as mltJximizing n intertemporal social welfare function given the 13x structure
Optimization is subject to a sequence of recursive equations of motion reflecting the evolution of the
ptlbllc debt allowing fDr government budget imbalances It should be stressathat this specification
01 the government cOMtcalnt is equivalent to the specificaUon in Auerbaeh-Kotlikoff (1983 1987)
when no liquidity constraints affect -gove~nment behavior
The extra richness in the treatment of government behavior allows tfte examination of
government debt poices in a truly dynamic setting Also financial crowding out effects induced by
government deficits can be analyzed (see Auerbach-Kottlikoff (1987) Feltenstein (1986) and
Perer (198Gb 1987e))
34 Foreign sector
Most of the open economy models follow the assumptions of balanced trade with import and
export net demands characterized by constant elasticities along the line of Ballardmiddot
-Fullerton-Shaven-Whalley (1965) Such is the case of Salfard (1963) Ballard-Goulder
(19S5) and Goulder-Sllmmers (1987)_ In Feltenstin (1985) the rest of the world is treated as
21
an addjlion~ consumer group
Bovcoberg (1986) develops a model In which two economies lIr considered each following
intertempoml perlect foresight paths These economies meet in the international forum Their
trade relaroMhips are characterized by yearly balanced trade accounts
None of the new generation of dynamic cae tax models has yet IncorPorated the lnternational
capital flows as done in aoulder-Shoven-Whalley (1983) for the earlier Ballarrshy
-Fullertonmiddot Shaven-Whalley (1985) model This is unfortunate as there is an important
Intertmporal aspect to international lending The first al1empts along Ihese lines are due to
Andersson (1987) and Erlich-Glnsburgh-Heyden (1987) Andersson (1987) in his model of the
Swedish economy adopts a closed economy appraoch in which rates of return in the domesJic
economy are largely determined by the international capital markets Fidinterest rate induce
Intlnatlonal capital flows which determine and finance the International lrade imbalance In
lurn In ErlichGinsburghHeydenmiddot (1987) foreign trade Is generaled according to an
intertemporal Irade welfare lunction with constant import and export elasticities In the shortmiddotrun
they allow inlernational trade inbalances which generate capilal ftows to the domestic households In
the long run however trade balance is assumed
35 The need for financial markets
A dynamic economic Wucture not only provides the ideal environment 10 model many features of
economic behavior it also permits one to Incorporate the financial side of the economy If the
government is allowed to run deficits the question 01 deficll financing automatically follows If
investment is optimally dewmined and returns to capital are different aClOSS sectors problems of
investment financing arise If there are seeral final1cfal assets In the economymiddot government
4-
I i
ibull
I I I I
I
j
22
bonds privpte bonds and equity (or simply physical capital installed inmiddot different sectors) the
problem of allocation of saving among assets with potentially different returns arises
The papers surveyed var greatly with respect to l~e extent of their attention to the financial
side of the economy At Clne extreme are the models in ~anardmiddotFuliertonmiddotShoven-Whalley (1985)
Andersson (1987) Ballltld (1983) Ballard-Goulder (1985) Bovenberg (1985 1986) and
Erlich-Gjnsburgh-Heyden C19B which are devoted exclusively to the real side of the economy
In turn Auerbach-KotliKoff (1984 1987) Feltenstein (1984 1986) and Goulde (1985)
allow for government debt In these models saving finances changes in government debt and physical
capital Private and publlc 2ssets are perceived by the households as perfect substitutes The
allocation of saving merely adjusts to the relative demands for funds
Feltenstein (1984 1986) is the only model surveyed which introduces money Government deficits are financed by issuing money and bonds according to an exogenously given rule Money is
demanded by consumers for transaction motives and an exogenously given fraction as a store of value
On the other hand government bonds and physical capital are the vehicles for the intertemporal
transfer of wealth
Summers (1985) and GOlldermiddotSummers (1987) introduce a whole menu of financial assets shy
firm specific equity capital Different assets earn different rates of retum However such rates
are equal up to constant and exogenous sector speCific risk premia Therefore the introduction of
constant exogenous risk premia as helpful as it may be in the context of calibration does not solve
the main issue of the non-optimality of the allocation of saving Also talking about risk premia in a
deterministic context is somewhat unsatisfactory
Pereira (1986b 1987d) also introduces a whole menu of assets private and public bonds and
firm specific equity However all the assets are expected to yield the same rate of return and
therefore perc~jved as perfect substitutes The different asset types allow consideration of
23
exogenOU$~8tUequjty triO dividendlrc~enlion rufes and therefore seve-ral sources of investment
financing bond equity and retained earnings
The nonmiddotcptimalHy of the allocation of saving and the absence or exogenelty of corporate
financial rules is _a reftectlon of the limitations of the determInistic approach Either in a context of
perfect anticipation oi the future prices Of in general within the t$atm of point price
expectations all the papers follow a deterministic approach Under such circumstances consumers
either expect diHerent rales of return (inclusive of risk premium) across assets in which case
they wilt buy only one asset (that with highest rate) Of they expect equal rat~s of relurn in which
case Ihey are indifferent about the asset composition of Iheir portfoliO There is no way of
Iradingmiddotoff rales of return and risks to obtain an optimal interiof solution to the problem of the
allocation of saving
The most advanced contribution in the modelling of saving allocation in a CGE setting is due to
Slemrod (1980 1983) in the context of a static one-period model In his model consumers act
according to a two stage separable decision process They Hrsl decide on how much to save Then
they decide on the allocation of saving according to an Indirect uttlity function dependent on the rates
of ratum and variances offered by the different assets in the eoonomy The sourCe of riskiness in the
ampConory comes from an uncertain marginal product of capital On the other hand aside from
portfolio decisions the res of the economy is insulated from uncertainty
The most complete contribution in terms of the treatment of the corporate financial rules is
FulienonmiddotGordon (1983) In a variant 01 Ihe BailardmiddotFulienon-ShovenmiddotWhelley (1985) model
Fullerton and Gordon have capital intensity and optimal financial decisions Jointly deter~ined
through a two siage process The cosl of financing capital is minimized by trading off the tax
advantages of debt against the epected real bankrupcy coSls inherenl with high debVequity ratios
Given the optimal debtequity ratio the level of inveSlment Is chosen such that at Ihe margin the
24 l
bull
rerurn pn ~uity equas the return on bonds plus an exogenous risk premium
36 Market assumptions equilibrium and expectations
Virtually all of the papers arc characterized by Walrasian market clearing assumptions All
markets are perlect1y competi1ive Atomistic competition among agents 1s assumed even thoughmiddot only
a flnile number of agents is considered Virtually no mark~t disequilibria or price stickiness afe
considered The only exception is ErlichmiddotGinsburghmiddotHeyden (1987) bull In thei model of the Belgium
economy I the _wage rate is fixed in the short-run Therefore in the shOrlwrun desequilibrium in
the labor market will generate endogenous unemployment However in lh$ long-run all prices
Including the wage rata are Ilexible and accordingly ali markets ciear
Given the dynamic nature of behavlor in the economy markelmiddotcfearing prices in each period
depend on expectations of fture prices and on tax variables In the economy There are essentially
two ways at interpreting the economic equilibrium in such a dynamic conte~bull If future prices are
perfectly anticipated (io expectations are self-fulfilling) a perfect foresight equilibrium
prevaHs Then future actions are merely the implementation of current decisions for future
periods However jf ptica expecla110ns are not perfect (Le agents make mistakes w1th re$pect 10
future prices) then a temporary or shortrun equilibrium prevails Markets ciear and clearing
prices depend on fu1ure price expectations Current plans about the future are typically not
precisely implemented They will be revised as more or better information becomes available to the
economic agents See Grandmant (1982) for a comprehensive survey 01 the temporary equilibrium
literature
Wilh the exception of Gould (1985) and Pereira (19S6b 19S7d) all the models surveyed
adopt th~ concept of a perfect foresight equilibrium In turn BaliardmiddotGouldr (1965) considers a
J
2S
flexible am9unt of foresight In terms of the number of years over which price movements are ~
foreseen Pereiras model (1986b 1987b)) is flexible in a somewhat different way in that it can
indude any range of foresight from myopia to perfect foresight The choice between the perfct foresight and lemporary equilibrium is ultimately to be made on
philosophic grounds It can be argued that less than perfect expectations Imply that agents are
irralional in some way (see Auerbach-Kctiikoff (1987 p 10) However the reverse argument can
be made One can question wiether agents are reaHy rationar and perfectly knowledgable about
future prices
Recent evidene of Balird (1987) Ballard-Goulder (19851 Goulder (19851 and Pereira
(1986b 1987d) suggests that the choice in modemng expectations is an important one They show
that the degree of foresight inl0 thc future (ranging from perfect foresight to-myopic expectations)
may have dramatic impacts on the pollcy conclusions of Ihe model Accordingly the best research
strategy may b to design models which are flexible enough 10 allow for different rules regarding the
formation of expectations With the exception of the articles just mentioned sensitivity analysis
bullhave previously not been performed along this dimension
In terms of implementation the two concepts of equilibrium - perfect foresight and temporary
equilibrium have different implications The dimensionality of the equ1ibrium soll1ion algorIthm
is involved Suppose we have a model with ten markets to be run for a period of 50 years Aside
from normalization a perfect foresight model implies computing prices In 500 dimensions while a
-t~mporary equilibrium model requires sorving SO equilibria each in ten dimensions Given that
computational speed often varies with the cube of the number of dimensions the lemporary
equilibrium formulation is potentially strikingly more feasible However Ballard (1987) and
Goulder-Ballard (1985) have devoloped techniques to greatly speed the computation of a perfect
foresight equilibrium
26
j
I
4 Jrnplemenlalion and Issues on policy evaluallon
41 Calibration and equilibrium comparlslon
CGE models at typically parameterized by the use of a calibration procedure Some parameters
are exogenously given However some crucial parameters are determined in such a way that the
model replicates the data for a given base year See MansurmiddotWhalleY 11984) for an eXlensive
discussion of 1~i$ issue
CalibratIon in a dynamic context is generaliy interpreted as requiring two properties First
replication of base year data is required Second the model is parameterized to simulate an
intertemporal balanced growth path when Ihe base policy is maintaineq This Is the approach
followed by Ballard (1983) BallardmiddotGoulder (1985) Goulder (1985) and SummersmiddotGoulde
(1986) It follows the practice of BallardmiddotFuliertonmiddotShovenmiddotWhalley (1985) with their
sequential cquilibrium dynamics
Other authors Auerbach-Kotlikof (1983 1984 1987) Boyenberg (1984 1985 1986)
and Pereira (198Gb 1987d) follow whal we cali a qualitalive calibration The slructural
palametars are exogenously chosen so that lh economy follows a reasonable path into Ihe fulure
ThaI has 10 do wllh the fact thaI given the recursive nature 01 the dynamic economy and aside tom
such structural parameters only iniUa stock values are needed to run these models Given initial
conditions on the stocks of say private wealth capltal and government debt agents will optimize
and Ihereby generate a lirst round of net demands and equilibrium conditions In lurn the
equilibrium prices will determine the evolution of the stock variables into the next perIod
There are several potential problems with calibration in the conlext of dynamic models The
assumption of a steadymiddotstate growth ~ath in the base case can be questioned First while
27 I Ibullbull
steadymiddotstaU is a possibity it certainly is not the only meaningfull solution to dynamic models
Even in the case of a perfect foresight equilibrium the model implies an equilibrium path which i may or may not involve colanced growth The model not the modeller should dictate the nature 01 I themiddotbase case path Second in the context of a temporarY equilibrium path a steady state solutlcn is Inol a likely model QutcOn1tl In fact unlessmiddot expectations are stattc short run behavior consistent
with a steady-state evolution will in general not be generated On the other hand if static I expectations are self~fufjili[1g we have in fact a perfect foresight model Third even the base year
replication requirement may cause problems In Ihe cOntexl of temporary equilibrium Any
calibration parameter would be conditional on expectation rules which is probably an undesirable
feature
The nalur of the two-requirement calibration strategy is very muchin the spirit of the bull
traditional design of the comparision of alternative equilibria comparis_ion between a steady-state
base case on one hand and alternative paths including a transition period and a finat sleadyyenstate on
the other hand Pereira (198Gb 1987d) compare different (not necessarily steadystate)
equilibrium paths The arguments against such a procedure are based on the idea that the impact of
policy changes can be observed most easily since all departures from the steadymiddotstate can be
attributed to the altemalive polioy On the other hand the base case so dEfined as a steadymiddotstate is
consist~nl with previous work in a less dynamtc setting and Iherefore allows a common standard
for comparing model results
42 Computation algorithms
We are still al a stage in which basically each author uses a different computational technique
Th development of dynamic models - in parlicular wilh adjustment costs andior perfect foresight
bull
28
has corresponded with the decline in the use of fixed point algorithms In -fact given the relative
arge dimensions inevitably involved such algorithms lend to be very inefficient at the best and
chen prohibitively stow See Stone (1985) and Preckel (1965) for a comparative assessment of
different computation techniques Among the models surveyed only Ball1rdFulle(ton~
-ShavenmiddotWhalley (1985) and Feltensteln 1985 use Merrills variant of the fixed point
algorithm 1echnique
AuerbachmiddotKoHikoff (1963 1984 1967) follow a three stage procedure They first compute a
base case steady-state then a revised case steady state and finally transition path for the economy
berveen these two steady-states tn all stages a Gauss-Seidel iterative procedure is used
Ballard (1982) Ballard-Goulder (1965) Goulder (HiSS) and Summers-Goulder (1986)
~
use a method developed by Ballard and Goutder which is similar in many awects to the FairmiddotTaylQr bull
(1983) algorl1hm Short-run equilibria are calculated (using Merrills algorithm) parametric on
nrice expeC~lions The model is then iterated to generate self-fulfilling intertemporaf expectations
and the corresponding perfect foresight equilibrium In a relatively similar approach Andersson
1987 uses a simulation program SIMNON developed In the University of Lund This program
can handle two-paint boundary problems In a fashion consistent with the multiple shooting
algcrlthm (see Lipton-Poterba-Sachs-Eummers (1982))
Boyenbergs computational approach (1985 1985) differs from the other models surveyed in
that he relies heavily on analytical techniques Computations are done by using a dynamic version of
Johansons linearization method Being essentially determined by the continuous time nature of the
I I
f
model thiS linearization model has the disadvantage of confining the analysls to infinltesinal changes
around the base case equilibrium (see Bovenberg (1985) p 53) Pereira (1986b 1987d) uses an optimization algorithm NPSOL - developped by
rGillMurray-Saunders-Wrlght (1986)) This algorithm is used 10 compute the sequence of
I
29
Shortmiddotrun temporary equilibrium which makes up the intertemporal equilibrium path The
equilibrium conditions are seen as nonlinear equality constraints in the minimization of an artlcial
objective function The prices are normalized to the unit simplex by an additional linear equality I I 1 bull constraint
II ErlichmiddotGinsburgh-Heyden (1987) follow a unique approach in that they use a variant of the
optimization technique introduced by Negishl (1960) The economic equilibrium can be generated i
as a solution of a mathemallcal program Ihe objective function of which is a weighted sum of the
utility functions of the various agents while tho constraints set consists of the market dearing 1
conditions Ginsburgh-Heyden (1985) have extended Negishis resulllO tho case of downward price
lrigidities I
Finally the paper by Jorgonson-Yun (1984) is also unique in tIlat it is the only I
econometrically estimated model Different blocks for the consumption and production Side of the
-
model are separately estimated to provide the necessary structural parameters i
I The diversity of computation techniques is yet another indicator of the exploratory nature of the
body of literature surveyed in this article
43 Equal yield comparisions
The link between a base case and counteriactual simulations Is usually provided by the concept of
equal yield Shaven-Whaley (1977) discuss the meaning 01 equal yield in a general equilibrium
context when government is not allowed to run defICits Equal yield is interpreted to moan constant
public utility Government base case utility is maintained in the counferfactual experiments With
balanced budgets the concepl of equal yield is unambiguous The new equUibrium prices and the
balanced budge condition will del ermine the minimal expendilure and taxes needed to maintain base
30
case publi~ utility Accordingly in general equal yield is inconsistent with equal nominal tax
revenue Some change in tax revenue Is necessary Different tax replacement schemes ate
considered to assure that enough tax revenue is collected This is the approach essentially followed
by most 01 the papers surveyed Only Feltenstein (1985) Goulder (198) and Jorgenson-Yun
(1984) chose not to follow an equal yield strategy
The question is of how to Interprete lila concept of equal yield when the government is alowed to
fUn deflcLts The optimal leve of expenditure for base ease public trliUty can now be taAinanced
bond financed or financed by a mix of bonds and taxation We have several versions of equal yield In
particular equal yield may now be consistent with equal tax revenUE Furthermore some meaSure
ltgtf financial crowding out effects of government deficns can be inferred from the comparislon of the
several equal yield alternatives These issues are extensively discussed ~Pereira (1987b)
44 Price expectations and the dynamic generalization of compensation
Indicators
The sum of equivalent and compensation variations over households is the most widely used
aggregated measure of efficiency gains or losses In the context of $teady~state eomparisions andor
when complete future markets exist andlor perfect foresighl is assumed there are no difficulties
associated with the use of the standard Hicksian indicators In fact correct future prices are known
in these cases
If expectations are not selfmiddot fulfilling andlor future markels are not available a dynamic
generalization is necessary 8allardmiddotGoulder (1985) provide some steps in that direction by
defining an indicator that accomodates periods far into the future when households do not have
perfect foresight but a steady~state prevails On he other hand this issue is more fundamental in
i
[
I
i I I I I [
I I
bullI
31
the contex~ of a Gner1 lemporary equilibrium ftamawor~ In such Circumstances Pereira
(1986a) dev~lops a cnamlc generalization which is obtained as the present discounted value of a i i
sequence of short-fUn optimal expenditure functions consistent with a base case expected future I stream of IlWities I
32
5 Empirical evidenee from selected policy Issues
Dynamic tax models have generated several important results which escaped Sialic modellers
The following is a seleCpoundd set of issuos whIch stress the marginal benefits of dynam~ modelHng of
economic behavior in innovltlttive areas of analysis The discussion of a consumption tax emphasizes
the benefits of dynamic tusehold behavior and intergenerational aspects The study of the
flliminatlon or the re~intToduction of the investsnemt tax credits is made meaningful by the dynamic
modelling of production betlavor Modelling governmenl in a dynamic conlexts allows the modeller
to study the irrrpact of government deflcits Finally a dynamic framework lets us appreCiate the I
relevance of consumer expectations in terms of the evaluation of policy alternatives
i
i
51 Consumption Tax
t
By Its very nature a consumption tax can not be adequately investigated with a static model bull
Fullerion-Shoven-Whaicy (1983) use Ihe model of the US economy s described in
Ballard-Fullerton-Shoven-Whalley (1935) to evaluate the movement from the currenl US tax Isys1em to a progressive consumption tax Since their modeJ Incorporates a Jaborleisure choice
where leisure is an untaxed commodlty their results reflect the fact that both the consumption tax
and Ihe present system are dslortionary
Concentrating on the rntertemporaf distortions they show that sheltering more saving from lhe
current US income tax could improve econornic efficiency even if marginal tax rate increases are
necessary in order to maintain government fevenue~ At first you have a revenue shortfall
However the economy moves to a higher sustatned growth path and uitimatefy lower tax rates
cangenerate the same revenue path Also wages increase in the long run with this policy The
bull
bullbull ~
33 shyt
i
I
I
present value of welfare gi[lS for a potiey of complete savin9 deduction wilh marginal tate
adjustments (consumption lax) is simulated to be around $500 billion to $600 billion 01 1973
dollars
They also investiga1e th 1ngth of ttme it takes th~ economy to adjust to these policy changes
Roughly they estimate the long run to b thirty years although this figure Is very sensitive to the
specification of the savings elasticity - a crucial parameler in this model
52 Investment T~x Credit
Goulder-$ummers (1 S87) address the impact of eliminating the lnvestmenl Tax Cradit (ITC) on
intersectoral capital formation and on economic growth Their model isYa1ticularly adequate to
addies$ this issu~ in that is postulates a forward looking investment m9we with adjustment costs
They show that the eliminating the ITC causes a reduction in the rale of investment Inveslment is
estimated to fall by about 7 in the short-run and by about 12 in Ihe new long run steady state
In lurn a previous policy anncuncement lowers Ihe overall al1ractiveness of Investment and leads 10
a downward shift of the Investrrent profile On the other hand the combined effect of a revenue
neutral simultaneous efimiraron of the ITC and reduction of the corporate tax rates is a long run
reduction ef the capital Sieck by 35 This pattern suggests that a revenue neutral increament in
both Ihe corporale tax rates and the inveSlment lax credits would be preferable 10 the formula
adopted in Ihe US Tax Reform Act of 1986
Pereira (1987d) addresses the impact of the ITe from Ihe oland point of Ihe curTenllax syslem
under the Tax Reform Act of 1986 Given Ihe concern over the potential depressiVe effects upon
savings and investment of the new tax system in conjunction with deficit reduC1ion mechanisms of
the GrammmiddotRudman type a pertinent question is should Ih lTC be e-introduced In the context
bull
34
-of his mod~1 of the US economy Pereira focuses on the trade-off between the distortion in the
intersectoral allocation of capital induced by the lTC the lower tax revenues it generates and
potential financial crowding-out effects on one hand and the positive effects of lowering the relative
price of new capital goods on the other hand Preliminary results confirm the qualitative results in
Goulder-Summers (1987) suggestting a welfare gain of about 2 of the present value of GNP in
the best scenario of absence of replacement taxes Furthermore preliminary results show an
important time pattern to welfare gains with the average benefits increasing the further you look
into the future This pattern is due to the presence of constraints to intersectoral mobility of capital
and inertia in the adjustment towards the optimal capital levels as reflected by the presence of
adjustment costs
I
I
bullAuerbach-Katlikaff (1gB) in the context of their intergenerational model of the US economy
consider the impact on savings capital formation and interest rates of deficit policies and balanced
budget increases in government consumption They conclude that deficit finance and government
consumption can significa-ntly crowd out capital formation and lower the welfare of future
generations However crowding out from deficit finance is a very slow process because it results
from increased government spending over potentially long horizons Also deficit policies may
substantially influence the longmiddotrun interest rates while leaving the short-run rates essentially
unaffected Finally and in opposition to the central role adjustment costs seem to play in both
53 Financial crowding outeff~cts of government deficits
Goulder-Summers and Pereiras models Auerbach-Kotlikoff suggest that the time path of interest
rates induced by a policy of deficit financing seems to be insensitive to the presence of adjustment
costs
bull
35
54 The role 01 consumer expectatlons
The expectations analysis has uncovered one of the benefits of dynamic modelling
Ballard-Goulder (1985) find that the appeal of adopting a consumption tax depends on the level of
foresight possessed by the consumers Furthermore additional foresight may be welfare worsening
This is a second best type of result This tends 10 occur under policies that lead to capital deepening
and declining rate of return to capital over time To the extent that consumers have more foresight
they will be beller equippod to anticipate the fall in the rental price of capital Consequently if a
saving incentive is enacted people save more with myopia than with perfect foresight Given the
exi~tence of taxes on capita and the discrepancy between the private an~--sccial returns to capital
the greater savings with myopia is beller socially Ballard-Goulder find that the welfare gains
from a consumption tax are reduced by about 10 when we move from myopia to farleet foresight
Therefore the attractivenness of adopting a consumption tax seems fairly [obust across these
foresight specifications
I
bull
36
6 Tbe power and weaknesses 01 dynamic modelling the example 01 corporale
tax Integrallon
The Issue of corporate tax integration allows us to sh~w the stre-ngths and weaknesses or the
dynamic approach Empirical evidence using CGE techniques indicates that depending on the precise
scheme of integration and the 1ax replacement methods integratiOn may have substantial effects In
Ihe work of FuliertonmiddotKingmiddotShoven-Whalley (1980 1981) total integration was found to yield an bull
annual static eHielency gain of $4 to $8 billion in 1973 dollars Simulated dynamic gains may be as
large as $695 billion or abcu 14 of the present value of tulUre consumption and leisure in the
US economy These gains result primarily from interindustry reallocalions of Investment and an
improved inter~emporal allocation of consumption
Mora recent work emphasizes hOw consumers asset portfolio decisions and firms finanCial
decisions affect the efficiency gains from integration Slemrod (1980) focusing on consumers
asset ponfoio deCisions finds static efficiency gains which are about twice -as Iarge as those
repOrted iiy Fulierton-King-ShovenWhalley (1961) FuUrlon-Gurdon (1983) focus on the
firms financial decisions They roporl efficiency gains of 6 of GNP from the elimination of the
tax distor1ions favoring debt However when they eliminate the corporate tax and repiace it with
increaSed persona income taxes addlHonal dls10t1ions are created in the optimal labot~leisure
decisions These distortions tend to dominate the analysis slJ(h that the overall effects of complete
integration are very modest Galpr-LuckmiddotTodr (1986) address simultaneously consumers
asset pOrlfollo decisions and the firms financial decisions They also find very modest efficiency
gains from integralion
The above results are important but may be severely biased There are severa aspects of
economic behavior and modelling crucial for the study of Income tax Integration whiCh have no been
i 37
captured innny of the above papers One aspect is the absence of government deficitsmiddot a balanced
budget is assumed It is true Ihal resource crowding OUI is caplured In these mOdels but financial
crowding out induced by government spending is nOI Second investment is not derived from
optimtztlcn behavior Investment behavior passively accomodates endogenous saving decisions As
a consequence the differential impacts of policies in the incentives to save and to invest are not
captured Aso full capital mobility across secors is assumed with instantaneous capital
adjustmenlS towards optimal levels This assumption rules qut different costs of capital across
seclors and therefore differentiated reactions to tax policies changes Finallyhe modelling of both
gov9rnmerlt deficits and of endogenous real and financial investment decisions necessitates the
~nsideraiion of a dynamilt framework and the introduction of financiaJ assets govemment bonds and
iprivate financial assets A dynamic framework also highlights the efficiency effects of Integration on r
ithe ptimal intertemporal decisions The above models are either sIalic (Slemrods and
GalpermiddotltJckemiddotTode(sj or a dynamic sequ~nce of otherwise Slatic model bullbull I Pereir a (1986b) develops a dynamic applied general equilibrium model bull to study the tbull
efficiency effects of integration on the growth and allocation of investmenl across sectors Special
Iattention is paid to the structural affeelS of resource and financial crowding oul induced by
gove mect spending and deficits and to the real and financial Investment reactions across industries
to both tax integration and flnancial crowding out His mode departs from previous work on income
lax Integraticn and for that matler from most 01 the CGIE literalure for tax policy evaluation in
several direcllons It encompasses an endogenous sequential equilibrium struelure founded on
dynamic behavior with flexible expectations Government deficits are optlmaUy determined
Investment decisions are tward looking and the resull of optimizing behavior Several financial
assetsmiddot public and privatemiddot are considered
Simulation resulls Indicale that the welfare gains from Integration under large deficits are at
32
best lJlodest when measured in terms of the GNP The elimination of the corporate tax and its I replacemenl by increased income lax rates yields long run benefits which are never larger than
2 of the presenl yalue ct telr consumption and leisure onder the previous tax regime and 1
under the current tax regime However the short run w~lare effects of full integration tend 10 be
very low and in in some scenarios even negative This is a flew intertemporar pattern of efficiency Ieffects which reflects an adjJsiment lag in the interindustry investment decisions due to the
existence of costs of adjustment On the of her hand partial integration achieved by excluding I dividends from the corporale lax base syslematically yields negalive efficlencyeffeclS In Ihe long Inm these can be as high as 3 of present value of the future value of consumption and leisure This
is a new second best effect suggesllng Ihat less than complete integration may have perverse I efficiency effects
The dynamic slructure of PerclrBs model snowing for an improved freatment of real investment
declsions and government deficts provides a potential setting for a more accurate measure of the
costs and benefits of inlegration The simulation resullS suggest lower benefits and an intertempora Ibull
pattern of growing benefits
Simulation results also clearly suggest robust negative effects from partial integration How
reliable is this result If dJvidends were deductible from the corporate lax base (as are interest
payments) the preferenlial Ireatmant of debl over equily would be ellminaled Corporallons should
be expecled 10 react by decreasing the optimal deWeqully rallo However corporale financial
gecisions are exogenously set In Pereiras model as in aft of the other models surveyed that go as far
as dealing with the issue
Also withdrawing dividends from Ihe corporate tax base is a way of eliminating Ihe double
taxation of dividends at belh Ihe corporate and personnal Income levels How will the optimal
aliocellon of household saving accomodate that change in Ihe relalive return to corporale equity
39 I
t However hOusehold portfolio decisions are exogenouly set in Pereiras model as in all of the other
models surveyeltj go as far as dealing with the issue
IIt is sate to say that the evaluation 01 the benefits 01 partial Integration as defined above are
sevrey underestima1eO Meanwhile the distortionSoenerated by the increase in the marginal
personal tax rates designed to ralse the revenue foregone by the dividend exclusion are fully
aCCltlunted for A good measure of the costs of integration together With a poor evaluation of the
benefils may very well be the reason why partial integration is Simulated 10 yield neg alive
eHikiency gains
On a different vein as most of Ihe model surveyed here Pereiras is a closed economy model It
is legitimate to wonder how the resullS would change If international capital flows were anowed
IThis almost certainly would affect financial crowding out since a good d~al of the capital inflows
could be used to finance public debt
For these various reasons the results should be treated as preliminary but they strongly I suggosl that the dynamic structure provides valuable new insights into the corporate tax integration SSU2
40
7 Concluding remarks I I I i
What has been acccmplished Ihis far The success of Ihe CGE research in laelding the challenge
of dynamic modelling can not be denied
I iGreat progress in the modelling of household behavior has been achieved Promising
developm~nts in the modeUing of production and government behavior have also been accomplished
Interesting innovations in the concept of equilibrium and computation techniques were discussed
The policy analysis was greatly enriched by considering transitional effects together with longer
run steadymiddotstate equilibrium paths generalized equal yield strategies accomodating different
financial crowding out impacts and generalized dynamic policy evaluation indicators
An important set of issues have been addresed by the models surveyed Traditional issues
focusing on the effects of capital taxation and consumption taxes have been pursued In terms of
capital taxation for example the intertemporal nature of the issue was enhanced by allowing an
optimal evolution of capital stock in the economy and forward looking optimal investment decisions
In Andersson (1987) GouldermiddotSummers (1987) and Pereira (1986b 1987d) this is coupled
with a improved treatment of several tax provisions like investment tax credits and depreciation
allowances In terms of the consumption taxes developments in the specification of Intertemporal
household behavior the introduction of overlapping generations and the modelling of
intergenerational links via bequest motives as well as a closer attention to demographic evolution
provided a much improved economic setting for the understanding of the several aspects of the
problem
Furthermore dynamic modelling has permitted the addressing of a set of new issues Policy
Issues ranging from dynamic tax policy analysis to the evaluation of exogenous changes in
government expenditures to the impact of different methods of financing government expenditure to
4 1
the importance of tinanciai croHrii1g out to 1he relevance of consumorS expectations in tne policy
results have been addressed
Despile all of the progress and in part due to such progress several avenues are wide open for
much neded additional research The following seem to be the most Interesting and promising
areasmiddot First the specincation of liquidity constraints to household behavior may help to obtain a
better understanding of policy chenges that affect directly the interest rates in the economy Second
major attention needs to be paid to the incorporafion of a roraign sector (with open capital markets)
Third some efforts are needed in terms of finding adequate computation techniques now that
dynamic modelling eally makes the curse of dimensionality worse than eVOrbull
The modelling of financial markets is the single most unsatisfactory speet of the dynamic
models surveyed here~ The problems associated with provlding a mo~ adequate treatment of
financial markets and in general endogenous and optimal financial deci$ions bull both at the corporate
and at the personal levelsmiddot have 10 be addressed That necessarily requires Introducing uncertainty
into the models To b adequate this must go well beyond the mere assumption of point expeetalions
with the whole array of modelling implementalion and evaluation problems thai generates
bull I i
42
REFERENCES
Abel A and O Blanchard 1983 An Intertemporal Model of Saving and Interest Econometrica
51 pp 675-92
~ Andersson K 1987 Tar-ation of Capital in Sweden - A General Equilibrium Model Ph D
Dissenation UnIversity of Lund Sweder
Auerbach A L KotlikoH 1983 National Savings Economic Welfare and the Structure of
Taxation in M Feldstein ed Behavioral Simulation Methods in Tax Policy- Analysis Chicago
University of Chicago Press
Auerbach A l Kotlikoff 1984 Social Security and the Economics of Demographic-
Transition in H Aaron and G Burtless eds Retirement and Economic Behavior Washington DC
Brookings Institution
Auerbach A L Kotlikoff 1987 Dynamic Fiscal Policy Cambridge Cambridge University
Press
Auerbach A L Kotlikof and J Skinner 1983 The Efficiency Gains from Dynamic Tax
Reform Internatiooaj Economic Reyiew 24 pp 81middot100
Ballard C 1983 Evaluation of the Consumption Tax with Dynamic General Equilibrium
Models Ph D Dissertation Stanford University
Ballard C 1987 Tax Policy and Consumer Foresight A General Equilibrium Simulation
Study Economic Inquiry 25 pp 267-284
Ballard C and L Goulder 1985 Consumption Taxes ForeSight and Welfare a Computatonal
General Equilibrium Analysis in J Piggott and J Whalley eds New Deyelooments in Apolied
General Equilibrium Analysis Cambridge Cambridge University Press
Ballard C D Fullerton J Shoven and J Whalley 1985 A General Equilibrium Model for
t 43
Tax Eoliey Eyalua~iQO NBER University or Chicago Press
Borges A (1987) bull A Survey of Compula1ional General Equilibnum Models CECD Economic
Slud~ 8
Bovenbig L 19B3 imperfect Capital Mobility in a Perfect Foresight Herberger Model
Mimea UC Br~eley
Bovenberg L 1984 Capitel Accumulation and Capital Immobility Qmiddottheory In a Dynamic General Equilibrium Framework PhD dissertation University of California at Berkeley t
Bovenberg L 19850 Dynamic General Equilibrium Tax Models with Adjustment Costs in A iManne ed bull 1985 EtQnomrc ECUjilbrium Model Formulation and Solution Mathematical
lProgramming Siudy 23 Arnsterdam NorthmiddotHoliand
Bovenberg L t 1986 Capital Income Taxation in Growing Open Eeondmies Journa of Pubtic
Eamp(omlcs 3t pp 347middot376 IBovenberg L ald W Keller 1981 Dynamics and Nonlinearlties in Applied General
Equilibrium Models Internal Report Mlmiddot81middot208 Department for Statistical Methods CBS bull
Vooiburg The Netherlands
Charnley C 1981 The Welfare Cost of C~pltal income Taxation In a Growing Economy
IJournal of Political ECQnQm~ 89 3 pp 468middot496 I
Charnley C 19810 Efficient Tax Reform in a Dynamic Model of General Equilibrium I Mimeo Yale University I Charnley C 1982 On the Optimal Taxation In Stylized Models 01 Inlartemporal General
Equilibrium Mimeo Yale University I Decaluw B and A Manens 1985 Pays en Oeveloppement at Modeles Calculables OEquilibre
General - Une Revue de al litterature empirique Monographie Sene 9 Volume 3 Centre de
Recherche et Developpemont Economlque Unrslt de Montreal
44
bull
I I
I
I
1
Erlich $ V Ginsburgh and L van der Heyden 1987 Wher~ do Real Wage Policies Lead
Belgium bull A General Equilibrium Analysis European Economic Reyiew (forthcoming)
air R and J Taylor 1983 Solution and Maximum Likelihood Estimation of Dynamic
Nonlinear Ratioant Expeclations Models Ecooometrica 51 pp 1169middot1186
Feltenstein A 1984 Money and Bonds In a Disaggregated Ec~nonYin Searl 1-1 J Shoven
Eds Aoolied ~eoeral EqJiliDCiulD tioalYsis Cambridge University Press Cambridge
Feltenstein A 19B6 A Dynamic General Equllibrlu Analysis of Financial Crowding Out
Theory with an Application to Australia Journal of Public Economics 31 No1
Fullerlon D R Gorden 1983 A Reexamination of Tax Distorllons in General Equilibrium
Models in M Feldstein Ed Behayora Simulation Methods in Tax pOlicy Analysis Chicago
University of Chipgo Press
Fvllerlon D J Shaven and J Whalley 1983 Replacing the US Income Tax with a
Progressive Consumption Tax JoulOal of Public ECQnomics 20 pp3middot23
Fullerton D Y Henderson J Shaven 1984 A COmparision of Methodologies in Empirical
General Equilibrium Models of Taxation in Searl H J Shoven Eds ~pQlied Geoaral Jguilibrium
lIoalyss Cambridge University Press Cambridge
Fullerton 0 A King J Shaven and J Whaney 1981 Corporate Tax Integration in the United
Statesmiddot a General Equiiibdufi1 Approach American Economic Review 71 pp~ 677~691
Galper H R Lucke E Tader 1985 Taxation portlor Choice and the Allocation of Capital A
General Equilibrium Approach Mimeo
Gill p W Murray M Saunders M Wright 1986 Users Guide for NPSOL (Version 40) A
Fortran Package for Nonlinear Programming T echnicsl Report SOL 86middot2 Deparlment of
Operations Research Stanford University
Glnsbur~h V and L van der Heyden 1985 General Equilibrium with Wage Rigidities An
45
AppJication to Belgium in A Manne ed t Economic EQumprit1m~ Made Formulation and $oJioo
Mathematical Programming Study 23 Amsterdam NorthmiddotHolland ~
Goulet J 1968 1 Adjustrent CO$1S in the Theory o(the Firm Reyiew of EconomiC Studies 35 bull
pp47middot55
Goulder L 1985 Tax-Financed versus BondmiddotFinancedGhanges in Governent Spending
bull Mimeo Harva(d UniversilYmiddot
Goulder l J Shaven and J Whalley 1983 Domestic Tax Policy and the Foreign Sector The
importance of altemative foreign policy formulations to results from a general equilibrium tax
analysis model In ~n methods in Ibe antico of income from caoia1 ed Martin
Feldstein Chicago University of Chicago riSS
Goulder l L Summers 1987 Tax Policy Asset Prices and 3rowtly ~
A General Equilibrium Analysis NBER Working Pper No 212B
Grandmont J 1982 Temporary General Equilibrium Theory Ch 20 in Handbook 01
Mathemalical Eccnornics Vol II NorthmiddotHoliand
Harberger A 1959 The Corporate Income Tax An empirical Appraisal In lax Revision
Comoendjum Voll House Cornmitee on Ways and Means Washington DC Govemment Printing
Office
Harberger A 1962 The Incidence of the Corporate Income Tax JOULn1 of Political
Economy 70 pp 215middot240
Harberger A 1964 Taxation Resource Aliocalrlon and Welfare fn lb Role of Dir bullbulll and
Indjresl Taxes io tbe Fede[al Reserve s~I~Ill Princeton Princeton Unlveresity Press
James J 1985 New Developments in the Applications of General Equilibrium models to
Economic History in Piggott aand Whalley Eds
Jorgenson D and K Yun 1984 Tax Policy and Capital Allocation Harvard Inslitute of
bull
I
I f
Economic Research WP 1107
Judd K 1965 ShonmiddotRun Analysis of Fiscal Policy in a Simple Perfect Foresight Model
Journal of poJiHcal 5CCJiW1l pp 298middot319 r
Upton D J Potbc J Sachsa nd L Summers 1982 Multiple Shooting in Rational
Expectations Models fuoometrlc 50 1329middot1333
Manne A ed 1985 fconQrnjc Eoullibrium Model FOrIDujaljQo and Solution Mathematical
Pogrammlng Study 23 Amsterdam NorthmiddotHolland
Manne A 1985 Or the Formulation and Solution 01 Economic Equilibrium Models In A
Manne ed ECQOQlli-Jr~ibmiddotum ~tQdel EQlJulailon and Solution Malhematical Programming
Study 23 Amsterdam NorthmiddotHolland
Mansur A and J Whalley 1984 Numerical Specification of Applied General Equilibrium
Models in Scarf H J Shoven Eds Aoplied General Eouillbrium AnaiysectIsect Cambridge Universily
Press Cambridge
Merrill 0 1972 Applications and Extensions of an Algorithm thaI Computes FixedmiddotPoints of
Cenain Upper Semimiddotcominuou Point to Set Mappings PhD Disserlation Unlvel$ity 01 Michigan
Negishi T 1960 Welfar Economics and Exislence I an Equnlbrlum for a Competitive
Economy MelroeCOOQ11ka 12 92middot97
Pereira A 1985 On the Existence and Uniqueness of Optimal Ou1put Path lor CRTS Firms in
Adjustment Costs Technologies Mimeo Stanford University
Pereira A 1985a Consumer Expectations and Ihe COSI 01 UllUty In an Intertemporat
Framework Mlmeo Stanford University
Pereira A 1985b Corporale Tax Integrallon and Government Deficits A Dynamic General
Equilibrium Anaiysls Mlmeo Stanford University
Pereir A 1987 On Ihe Computation of the Users Costs of Stocks Ecooomi (forthcoming)
47
Pereira A bull 1987b Equal Yield Alternatives and Government Oeficlts Mimeof Stanford
University
Pereira A 1987c A Dynamic Applied General ~uHlbrium Model for Tax Policy Evaluation
Comlletamp Docurnent2tion Mimeo University of California San Diego (in progress)
Pereira A 1987d Should the investment Tax Credit be Rei(ltroduced Mlmeo University
of California San Diego (ir progreSs)
Piggott J and J Whalley eds 1985 tJew QeyeQpments in Applied General EQuilibrium
bullenalysjs Cambridge Cambridge University Press
Preckel Pbull 1985 Alternative algorithms for computing economic equilibria In A Manne Ed
Radnor R 19amp3 Equllibrium under Uncertainty In K Arrow M Intrilligaor ods
I ibull
Robinson S 1 g86 Mullisectorai Models of Oeveloping CountrieS a Survey Oivision of
Agricullure and Resources UC Borkeley WP 401
Scarf H 1967 The approximation of fixed points of a continuous mapping ~MM Jouwal of
epplid Mathornallcs 15 pp 1328middot43
Scarf H with T Hansen 1973 00 the Computation of EconomiC Eguilibria New Haven Yale
University Press
Scarf H J Shoven Eds 1984 61l1llied Genera Elujlibrium 6nalyssect Cambridge University
~ Press Cambridge
Shoven J 1976 The Incidence and Efficiency Effects of Taxe on Income from Capital
Jurnal of ramie1 Economy 84 pp 1261-83
Shoven J 1983 Applied General Equilibrium Tax Modemng IMF Slaff PaoerS 30
Shoven J and LB Simon 1986 Share Repurchases and Acquisitions An Analysis of which
Firms Participate NBER Working Paper No 2243
bull
48
IShaven J and J Whalley 1972 ft General Equmbrium Calculation of lhe Effects of
Differential Taxation of Incoma from Capital in the US Jauroe of public Econgmics 1
pp281middot321 (
Shoven J bull and J Whalley 1973 General Equilibrium with Tax$s a computation procedure
and an existence pr00f ReviEhi of Economic Studies 40 pp 475middot490
Shoven J J Whalley 1977 Equal Yield Tax Alternatives A General Equilibrium
Computa~ui0nal Tfchnique J~HJrnal of PUblic EconomIcs 8 pp 211~224 bull
Shaven J J Whalley 1984 Applied GeneralmiddotEquilibrlum MOdels of Taxalion and
International Tr8de An Introduction and Survey Journal of Economic literature 23t pp
1007middot1051
Slemrod J 1980 A General Equilibrium Model of Capital Income Taxation Ph D bullDissertation Harvard Univnrsity
Slemrod J 1983 A General Equilibrium Model of Taxation with Endogenous Financial
Behavior in M Feldstein edbull Behavioral Simulation Methods In Tax Policy Analysis Chicago
UniVerSity of Chicago Press
Stone J 1985 Sequential Optimization and Complementarily Techniques for computing
Economic equilibrium in A Manne Ed
Summers L 1981 Taxation and Corporate Investment a q-Theory Approach BroQkinas
Summers L 1985 Taxation and the Size and ComposHion 01 the Capital Stock An Asset Price
Approach Mlmeo Harvard University
Whalley J 1975 A Geneal Equiloibrium Assessment of the 1973 United Kingdom tax
reform Economica 42 pp 139middot61
1JLTIllOS JIQll~)ltG PAPERS PUBLICADOS bull
I
I I
j
112 70 - LUCENA iJiCIIJo ~ IIEnvironrilent Honitoring and Orqanization Structure T (Setellibro 1987)
nI 71 - L~ENA Diogo A Note on the Representation of Information Stxueture (SeteJObro 1987)
nQ 72 - LUCENA D1CgO Envuorment 111)nitoring and Organization
Structure II (Setelilhro 1987)
ril 73 - AlmmE3 Antonio Pais e CAsPAR Yitor bull Tributacao Inoentivos e InvesUlllentos Analise Qualitativa noYellibro 19S7)
n2 74 - liAlltOSA rose Pedro Opt1lllal lIage R1gidity A SUggestedtlthodology to Teet tb Theory and an Application (Outubro 1987)
nI 75 - CABRAL L 11 B Three lIotes on SYJIIlletric Gemes witt itsYJIIlletric Equilibri (NoveJObro 1987)
n2 76 - LUCENlt D10gOTo Search or Not to Search (Fevereiro1967)
nQ 17 - SA Jorge asconcelos euroI A Theory of SynergyB (Fevereiro1987)
n2 76 - VILARES tanuei Jose Os Be IntenaMios IlIIportados ColIIO Factor de Produao (Julho 1987)
nQ 79 - SA J~rge VasCQncel03 e HoY to Copete and Co~unicate in tature Industrial Products_ (aio 1988)
n2 80 - ROIl Rafael Learning and Capacity Expansion in a New larket Under Uncertainty (Fevereiro 1988)
n2 81 PEREIRA Alfredo llarvilo Survey of Dynamic Computational General EquilibriUlll Hodel for Tax Policy Evaluation (Outubro 1987)
Qualquer in ormacao sabre o 1IIorlltlng Papers ja publlcados sera prestada pelo Secretariado de Apoio aos Docentes podendo os ~e3OSf
ser adquirido5 ne Secyo de Vlndas da Faculdade de EconollLia UNL na Travessa Esteviio Pinto Cltpolide - 1000 LISllOA
14
economy ~ a sel of connected equilibria The connectIon is pra~ided by capital accumulation
Capital accumulation is endogenous and determined by saving The model stans with a sociagt
accOunting mi1rix In the bzse case the economy is assumed to be in a steaqy state growth -path (along
which all relative prices ~fe constant) The madel solves for both the naw steady state growth path
and the transltion 10 it zf~er a policy intervention rhe authors have frequently addressed the
question-of howong rt tKes to effectively settle un10 a new steady S1ate growth path
The dynamics of the model are limited however in that future consumption is collapsed into a
composite commodity Aso the absence of government deficits and the lack ofproductlon dynamics
limits the realism of the BallardmiddotFuliertonmiddotShoven-Whatley model for the analysis of dynamic
policy issues (such as the adoption 01 a consumption tax or Ihe elimination of the investment tax
credit) The models we survey try to Improve upon Ihe Ballard-FlIiertonmiddotShaven-Whalley
dynamic modelling The models Jn~orporate more forward lookTng behavior They improve the
dynamics on the conampumrtkm side of the economy and some of them introduce true 1ynamlc behavior
on the part oI the producers We turn now to the issues inY9lv_ed in developing such dynamic models
15 I I ~3 ~ssues in the modelling Of dynamic behavior
We will 100k firt at the speclficalion of economic behavior of consume(s~ producers
goyemment and Ihe r~sl of Ihe world The modemog of corporale and household financial decisions
I is then addr5sed Finally Ihe eencept of equilibrium is discussed
I
i
I 1
31 Consumers behavior
Early efforts 10 build dynamic features into the economic behavior 01 consumers are due to
Ballard (1963) and Auerbach-Kotllkoff (1983 1984) Their work has been closely followed by
several subsequent authors In faet dynamic household behavior Is the single most pervasive aspect
amohg the models surveyed
All the models Inc-O1orale some loro of lila-cycle behavior Household behavior is dewmloed
by the maximization of an addWVely separable time invariant intertemporal utillty function The
utility lune1ion is defined oyer the domain of the consumption goods In the economy In most of the
models leisure is also an argument in utility so that labor supply is optimally determined
Utility maximi4ation is subject to a lifetime intertemporal budget constraint which equalizes
the presen1 value of consumers income and expenditure More reeentiy in Andersson (1987) bull
Bovenberg (1985) and Pereira (1986b 1987d) the constraint i$ defined as a sequence 01
recursive equations of motion on wealth This has the potential advantage 01 accomodating liquidity
constraints However It should be recognized that in the absence 01 liquidity constraints (ie when
consumers are free to completely borrow against future income) the two specificallons of the
househofd ccnstraint are essentiay equivatent Furthermore tn both versions saving is optimally
determined as a way of translaring wealth intertemporally
1 6
Some lodels now have a sophisticated dynamic specification of hoosehokl In Ballard (1983) -
BallardmiddotGoulder (1985) and Auerbach-Kotlikof (1983 1984 1987) consumers behavio is
embedded in an iterQEneraional setting Each of them has S5 age cohorts simulataneously alive
Som-e measure of jntergener1ional altruism is considered in the form of bequest motives In Ballard
(1983) and Ballard-Gouldr (1985) consumers derive utility directly from bequthing parl of
their wealth Therefore bequests assume 1he form of a scrap value of terminal wealth of a
generation In AuerbachmiddotKotlikof (1983 1984 1987) each generation empathyz bullbull with Ihe
utility of future generations ovar bequethd wealth In addition Ballard (1ge3) includes detailed
demographic poiections wfill the intent of incorporating in the policy analysis the effects of the
bull Post World War tI baby-boom
In the real world household decisions also include the optimat allo~atiOn of saving among
shyalternative physical or financial assets Theories of household portfolio behavior are relatively
Iwen~establshed However they involve uncertainty as a crucial element Accordingly they have
middotnot been incorpora~ed in the context of the determinIstic dynamic models we sutyey here We will 1 Come backt-O this Issue below
32 Producer behavior
The efforts to build dynamic features into the economic behavior of producers are more recellt
and less widely adopted The first attempts are due to Bovnberg (1984 1985) and Summers
(1965) Dynamic behavior has been more fully incorporated In the recent models of Andersson
(1987) Auerbach-Kotlikoff (19B7) GouldermiddotSummers (1987) and Pereira (1986b 1987d)
Part of the reason why production side dinamlcs has been more slowty adopted is the weak
supply of accepted theories referring to the dynamic behavior of the firms In the models referred
I 17
to above gynamlc producion and investment b~havior are middotinduced by the existence of capital
adjustment costs and linked to Tobins q theory Adjuslment costs are designed to capture both the
incomplete mobility of capital aoross industries and installation costs ~e the costs of adjusting
capital towards its cpHmai level)
Adjllstment costs can be conceived as internal to the firm and measured in terms of foregone
outpul along the lines of Lucas (1967) Alternatively adjustment costs can be viewed as actual
external cests incurred together with Ihe purchase costs along the Unes of Gould (1969) With the
exception of Pereira (198Gb 1937d) all of the CGE authors follow the former approach
The firms in the economy maximize their market value as the present discountad value of the
future stream of dividends In Andersson (1981) and Pereira (198Gb 1987d) firms are seen as
maximizing the present discounted value of net cash flow Maximization is constrained by Ihe
adjustmenl cosl lechnology and ~n equation of motion describing Ihe evolution of the capital stock
It should be noted that undereogenous divldendrelention rules the two problems are equivalent II
should also be nOled Ihat a salisfectory economic rational for th bullbullistoo of dMdonds with the
present tax code is missing in the profession (Shoven (1986))
The models with static formulation of producers behavior are characterized by passive
investment behavior Investment merely accomodates to saving in the eccnomy With a dynamic
formulation induced b adjustment costs real investmen decisions are forward looking Investmenl
Is endogenously and optimally determined by the firms A fundamental difference between the
shonmiddotrun in which capital stock is given and longmiddotrun in which Ihe level of capital is allowed to be
optimally determined is emphasized
This extra richness of produc1ion dynamics is not without costs A careful look at the summary
tables will clearly show an inverse relation between the adoptlon of dynamic features in production
and the level of disagreggation of the production side of the economy In fact with production
bull
j
1 6
dynamics $e dimension of the problems is immensly increased Let IS be more specific
In a static framework with constant returns to scale production technology the output level ismiddot
indeterminate The optimal aliocation of inputs can be obtained by cost minimization with any
feasible output level generatillg zero profits Such zero profit conditions are used to solve for the
output prices in terms of the factor prices and hencemiddot to reduce the dlmensionality of the problem
from the -number of commodities and factors to the number of factors Thus even if the model deals
with 30 production sectors the computation of an economic equilibrium can take place using only
the dimensionality of the primary inputs in the economy
Now under certain regularity conditions on the production and adjustment costs technologies
leading to enough concavity of the optimality objective the intertemporal output path for the firm is
endogenously optimally and uniquely determined even with constant relums to scale technologies
(see Pereira (1985 1987a) Accordin~y with adjustment costs in the model optimal profits
will in general be non-zero This result has crucial implications for the computatiQn of equilibrium
in the model~ with production dynamics With adjustment costs no reduction of dimensionality is
possible The curse of dimensionality returns as a binding constraint
The introduction of adjustment costs adds another significant complication to the model With
capital being less that perfectly mobile in the economy different rates of return on capital will
exist in different sectors This is a difficult problem to tackle conceptually in the absence of
uncertainty Also in the real world producer maximization choices include also its choice of
financial ratios (debVequity) and payout rates (dividendretained earnings) These financial
subjects are difficult and much Clf this behavior is still taken as exogenous by the modellers We
will come back to these issues below
middot i
19
bull 33 government behavior
middot
Two issues dominate the modelling of government behavIor First govemment behavior has
Ibeen typically seen in Ihe CGE liIerature for lax poliCY evaluation as constrained by yearly balanced
I budsets (see for example 8allard-Fullrton-S~ovn-Whafley 098Sraquo The analysis of
govemment defidlS and public debl in a CelE context requires a dynamic selling Second the level of
government expenditures is either exegenouty given as in Auerbach-KoUikofi (1984 1987)
6ovenbrg (1984 1985) Feltenstein (1984 1986) and Jorgenson-Yun (1984) or
endogenously (but no optimally) determined by the balanced budget conditions as in
6allard-Fullerton-Shoven-Whalley (1985) Andersson (1987) Erllch-Ginshurgh-Heyden
~
(1987) Gouder (19aS) and Gouder-Summers (1987) In the second case the composition of
bull public expenditures is often optimally determined However the leve~ of government expendi1ures
can only 00 endogenously and optimally determined if the government is seen as an optimizing agent
and is allowed to run deficits
The first attempts to deal wiU the government defiCits in CGE tax models are due to
Auerbach-Kollilltoff (1984) and Feensain (1984) In Auerbach-Kotlikof (1984) government
expenditures afe exogenous They grow at the rate of growth of population However given 1he tax
structure and tax revenues yearly deficits and surpluses are atTowed subject to an in1ertempond
constraint that the present value of future tax revenues equals the present value of future
expenditures
In Fellenstein (1984 1986) government expenditures are also exogenously given He also
allows the government to run deficits to finance expenditures in excess to 1ax revenues However
surpluses are returned to CQnsumars in the form cf transfers Accordingly government ts not
subject to any constraint regarding the future repayment of public debt
In GouJder (19851 the 9Vernmenl maximizes a static social welfare ~nctio subject to a-
bull
20 I i
I
I I I
I bull
balanced budget conSirtlL This follows the optimal alocatkm of government expenditures alpng
the lines of Baliard-FHeiton-Shoven-Whalley (1985) Ttle model Is generallzea to allow for
exogenous changes in HIe time path of government expenditures Two financing alternatives are
consjdeft~d and contr8s1ed EldditionaJ tax revenues and bond issuance
Pereira (198Gb 1137d) auempts to address both the incorporation of deficits and the
determination of govement expenditures The path of government expenditures and the path of
delicilSlsurpluses (ano therefore the path lor debt) are endogenously and optimally determined The
government is seen as mltJximizing n intertemporal social welfare function given the 13x structure
Optimization is subject to a sequence of recursive equations of motion reflecting the evolution of the
ptlbllc debt allowing fDr government budget imbalances It should be stressathat this specification
01 the government cOMtcalnt is equivalent to the specificaUon in Auerbaeh-Kotlikoff (1983 1987)
when no liquidity constraints affect -gove~nment behavior
The extra richness in the treatment of government behavior allows tfte examination of
government debt poices in a truly dynamic setting Also financial crowding out effects induced by
government deficits can be analyzed (see Auerbach-Kottlikoff (1987) Feltenstein (1986) and
Perer (198Gb 1987e))
34 Foreign sector
Most of the open economy models follow the assumptions of balanced trade with import and
export net demands characterized by constant elasticities along the line of Ballardmiddot
-Fullerton-Shaven-Whalley (1965) Such is the case of Salfard (1963) Ballard-Goulder
(19S5) and Goulder-Sllmmers (1987)_ In Feltenstin (1985) the rest of the world is treated as
21
an addjlion~ consumer group
Bovcoberg (1986) develops a model In which two economies lIr considered each following
intertempoml perlect foresight paths These economies meet in the international forum Their
trade relaroMhips are characterized by yearly balanced trade accounts
None of the new generation of dynamic cae tax models has yet IncorPorated the lnternational
capital flows as done in aoulder-Shoven-Whalley (1983) for the earlier Ballarrshy
-Fullertonmiddot Shaven-Whalley (1985) model This is unfortunate as there is an important
Intertmporal aspect to international lending The first al1empts along Ihese lines are due to
Andersson (1987) and Erlich-Glnsburgh-Heyden (1987) Andersson (1987) in his model of the
Swedish economy adopts a closed economy appraoch in which rates of return in the domesJic
economy are largely determined by the international capital markets Fidinterest rate induce
Intlnatlonal capital flows which determine and finance the International lrade imbalance In
lurn In ErlichGinsburghHeydenmiddot (1987) foreign trade Is generaled according to an
intertemporal Irade welfare lunction with constant import and export elasticities In the shortmiddotrun
they allow inlernational trade inbalances which generate capilal ftows to the domestic households In
the long run however trade balance is assumed
35 The need for financial markets
A dynamic economic Wucture not only provides the ideal environment 10 model many features of
economic behavior it also permits one to Incorporate the financial side of the economy If the
government is allowed to run deficits the question 01 deficll financing automatically follows If
investment is optimally dewmined and returns to capital are different aClOSS sectors problems of
investment financing arise If there are seeral final1cfal assets In the economymiddot government
4-
I i
ibull
I I I I
I
j
22
bonds privpte bonds and equity (or simply physical capital installed inmiddot different sectors) the
problem of allocation of saving among assets with potentially different returns arises
The papers surveyed var greatly with respect to l~e extent of their attention to the financial
side of the economy At Clne extreme are the models in ~anardmiddotFuliertonmiddotShoven-Whalley (1985)
Andersson (1987) Ballltld (1983) Ballard-Goulder (1985) Bovenberg (1985 1986) and
Erlich-Gjnsburgh-Heyden C19B which are devoted exclusively to the real side of the economy
In turn Auerbach-KotliKoff (1984 1987) Feltenstein (1984 1986) and Goulde (1985)
allow for government debt In these models saving finances changes in government debt and physical
capital Private and publlc 2ssets are perceived by the households as perfect substitutes The
allocation of saving merely adjusts to the relative demands for funds
Feltenstein (1984 1986) is the only model surveyed which introduces money Government deficits are financed by issuing money and bonds according to an exogenously given rule Money is
demanded by consumers for transaction motives and an exogenously given fraction as a store of value
On the other hand government bonds and physical capital are the vehicles for the intertemporal
transfer of wealth
Summers (1985) and GOlldermiddotSummers (1987) introduce a whole menu of financial assets shy
firm specific equity capital Different assets earn different rates of retum However such rates
are equal up to constant and exogenous sector speCific risk premia Therefore the introduction of
constant exogenous risk premia as helpful as it may be in the context of calibration does not solve
the main issue of the non-optimality of the allocation of saving Also talking about risk premia in a
deterministic context is somewhat unsatisfactory
Pereira (1986b 1987d) also introduces a whole menu of assets private and public bonds and
firm specific equity However all the assets are expected to yield the same rate of return and
therefore perc~jved as perfect substitutes The different asset types allow consideration of
23
exogenOU$~8tUequjty triO dividendlrc~enlion rufes and therefore seve-ral sources of investment
financing bond equity and retained earnings
The nonmiddotcptimalHy of the allocation of saving and the absence or exogenelty of corporate
financial rules is _a reftectlon of the limitations of the determInistic approach Either in a context of
perfect anticipation oi the future prices Of in general within the t$atm of point price
expectations all the papers follow a deterministic approach Under such circumstances consumers
either expect diHerent rales of return (inclusive of risk premium) across assets in which case
they wilt buy only one asset (that with highest rate) Of they expect equal rat~s of relurn in which
case Ihey are indifferent about the asset composition of Iheir portfoliO There is no way of
Iradingmiddotoff rales of return and risks to obtain an optimal interiof solution to the problem of the
allocation of saving
The most advanced contribution in the modelling of saving allocation in a CGE setting is due to
Slemrod (1980 1983) in the context of a static one-period model In his model consumers act
according to a two stage separable decision process They Hrsl decide on how much to save Then
they decide on the allocation of saving according to an Indirect uttlity function dependent on the rates
of ratum and variances offered by the different assets in the eoonomy The sourCe of riskiness in the
ampConory comes from an uncertain marginal product of capital On the other hand aside from
portfolio decisions the res of the economy is insulated from uncertainty
The most complete contribution in terms of the treatment of the corporate financial rules is
FulienonmiddotGordon (1983) In a variant 01 Ihe BailardmiddotFulienon-ShovenmiddotWhelley (1985) model
Fullerton and Gordon have capital intensity and optimal financial decisions Jointly deter~ined
through a two siage process The cosl of financing capital is minimized by trading off the tax
advantages of debt against the epected real bankrupcy coSls inherenl with high debVequity ratios
Given the optimal debtequity ratio the level of inveSlment Is chosen such that at Ihe margin the
24 l
bull
rerurn pn ~uity equas the return on bonds plus an exogenous risk premium
36 Market assumptions equilibrium and expectations
Virtually all of the papers arc characterized by Walrasian market clearing assumptions All
markets are perlect1y competi1ive Atomistic competition among agents 1s assumed even thoughmiddot only
a flnile number of agents is considered Virtually no mark~t disequilibria or price stickiness afe
considered The only exception is ErlichmiddotGinsburghmiddotHeyden (1987) bull In thei model of the Belgium
economy I the _wage rate is fixed in the short-run Therefore in the shOrlwrun desequilibrium in
the labor market will generate endogenous unemployment However in lh$ long-run all prices
Including the wage rata are Ilexible and accordingly ali markets ciear
Given the dynamic nature of behavlor in the economy markelmiddotcfearing prices in each period
depend on expectations of fture prices and on tax variables In the economy There are essentially
two ways at interpreting the economic equilibrium in such a dynamic conte~bull If future prices are
perfectly anticipated (io expectations are self-fulfilling) a perfect foresight equilibrium
prevaHs Then future actions are merely the implementation of current decisions for future
periods However jf ptica expecla110ns are not perfect (Le agents make mistakes w1th re$pect 10
future prices) then a temporary or shortrun equilibrium prevails Markets ciear and clearing
prices depend on fu1ure price expectations Current plans about the future are typically not
precisely implemented They will be revised as more or better information becomes available to the
economic agents See Grandmant (1982) for a comprehensive survey 01 the temporary equilibrium
literature
Wilh the exception of Gould (1985) and Pereira (19S6b 19S7d) all the models surveyed
adopt th~ concept of a perfect foresight equilibrium In turn BaliardmiddotGouldr (1965) considers a
J
2S
flexible am9unt of foresight In terms of the number of years over which price movements are ~
foreseen Pereiras model (1986b 1987b)) is flexible in a somewhat different way in that it can
indude any range of foresight from myopia to perfect foresight The choice between the perfct foresight and lemporary equilibrium is ultimately to be made on
philosophic grounds It can be argued that less than perfect expectations Imply that agents are
irralional in some way (see Auerbach-Kctiikoff (1987 p 10) However the reverse argument can
be made One can question wiether agents are reaHy rationar and perfectly knowledgable about
future prices
Recent evidene of Balird (1987) Ballard-Goulder (19851 Goulder (19851 and Pereira
(1986b 1987d) suggests that the choice in modemng expectations is an important one They show
that the degree of foresight inl0 thc future (ranging from perfect foresight to-myopic expectations)
may have dramatic impacts on the pollcy conclusions of Ihe model Accordingly the best research
strategy may b to design models which are flexible enough 10 allow for different rules regarding the
formation of expectations With the exception of the articles just mentioned sensitivity analysis
bullhave previously not been performed along this dimension
In terms of implementation the two concepts of equilibrium - perfect foresight and temporary
equilibrium have different implications The dimensionality of the equ1ibrium soll1ion algorIthm
is involved Suppose we have a model with ten markets to be run for a period of 50 years Aside
from normalization a perfect foresight model implies computing prices In 500 dimensions while a
-t~mporary equilibrium model requires sorving SO equilibria each in ten dimensions Given that
computational speed often varies with the cube of the number of dimensions the lemporary
equilibrium formulation is potentially strikingly more feasible However Ballard (1987) and
Goulder-Ballard (1985) have devoloped techniques to greatly speed the computation of a perfect
foresight equilibrium
26
j
I
4 Jrnplemenlalion and Issues on policy evaluallon
41 Calibration and equilibrium comparlslon
CGE models at typically parameterized by the use of a calibration procedure Some parameters
are exogenously given However some crucial parameters are determined in such a way that the
model replicates the data for a given base year See MansurmiddotWhalleY 11984) for an eXlensive
discussion of 1~i$ issue
CalibratIon in a dynamic context is generaliy interpreted as requiring two properties First
replication of base year data is required Second the model is parameterized to simulate an
intertemporal balanced growth path when Ihe base policy is maintaineq This Is the approach
followed by Ballard (1983) BallardmiddotGoulder (1985) Goulder (1985) and SummersmiddotGoulde
(1986) It follows the practice of BallardmiddotFuliertonmiddotShovenmiddotWhalley (1985) with their
sequential cquilibrium dynamics
Other authors Auerbach-Kotlikof (1983 1984 1987) Boyenberg (1984 1985 1986)
and Pereira (198Gb 1987d) follow whal we cali a qualitalive calibration The slructural
palametars are exogenously chosen so that lh economy follows a reasonable path into Ihe fulure
ThaI has 10 do wllh the fact thaI given the recursive nature 01 the dynamic economy and aside tom
such structural parameters only iniUa stock values are needed to run these models Given initial
conditions on the stocks of say private wealth capltal and government debt agents will optimize
and Ihereby generate a lirst round of net demands and equilibrium conditions In lurn the
equilibrium prices will determine the evolution of the stock variables into the next perIod
There are several potential problems with calibration in the conlext of dynamic models The
assumption of a steadymiddotstate growth ~ath in the base case can be questioned First while
27 I Ibullbull
steadymiddotstaU is a possibity it certainly is not the only meaningfull solution to dynamic models
Even in the case of a perfect foresight equilibrium the model implies an equilibrium path which i may or may not involve colanced growth The model not the modeller should dictate the nature 01 I themiddotbase case path Second in the context of a temporarY equilibrium path a steady state solutlcn is Inol a likely model QutcOn1tl In fact unlessmiddot expectations are stattc short run behavior consistent
with a steady-state evolution will in general not be generated On the other hand if static I expectations are self~fufjili[1g we have in fact a perfect foresight model Third even the base year
replication requirement may cause problems In Ihe cOntexl of temporary equilibrium Any
calibration parameter would be conditional on expectation rules which is probably an undesirable
feature
The nalur of the two-requirement calibration strategy is very muchin the spirit of the bull
traditional design of the comparision of alternative equilibria comparis_ion between a steady-state
base case on one hand and alternative paths including a transition period and a finat sleadyyenstate on
the other hand Pereira (198Gb 1987d) compare different (not necessarily steadystate)
equilibrium paths The arguments against such a procedure are based on the idea that the impact of
policy changes can be observed most easily since all departures from the steadymiddotstate can be
attributed to the altemalive polioy On the other hand the base case so dEfined as a steadymiddotstate is
consist~nl with previous work in a less dynamtc setting and Iherefore allows a common standard
for comparing model results
42 Computation algorithms
We are still al a stage in which basically each author uses a different computational technique
Th development of dynamic models - in parlicular wilh adjustment costs andior perfect foresight
bull
28
has corresponded with the decline in the use of fixed point algorithms In -fact given the relative
arge dimensions inevitably involved such algorithms lend to be very inefficient at the best and
chen prohibitively stow See Stone (1985) and Preckel (1965) for a comparative assessment of
different computation techniques Among the models surveyed only Ball1rdFulle(ton~
-ShavenmiddotWhalley (1985) and Feltensteln 1985 use Merrills variant of the fixed point
algorithm 1echnique
AuerbachmiddotKoHikoff (1963 1984 1967) follow a three stage procedure They first compute a
base case steady-state then a revised case steady state and finally transition path for the economy
berveen these two steady-states tn all stages a Gauss-Seidel iterative procedure is used
Ballard (1982) Ballard-Goulder (1965) Goulder (HiSS) and Summers-Goulder (1986)
~
use a method developed by Ballard and Goutder which is similar in many awects to the FairmiddotTaylQr bull
(1983) algorl1hm Short-run equilibria are calculated (using Merrills algorithm) parametric on
nrice expeC~lions The model is then iterated to generate self-fulfilling intertemporaf expectations
and the corresponding perfect foresight equilibrium In a relatively similar approach Andersson
1987 uses a simulation program SIMNON developed In the University of Lund This program
can handle two-paint boundary problems In a fashion consistent with the multiple shooting
algcrlthm (see Lipton-Poterba-Sachs-Eummers (1982))
Boyenbergs computational approach (1985 1985) differs from the other models surveyed in
that he relies heavily on analytical techniques Computations are done by using a dynamic version of
Johansons linearization method Being essentially determined by the continuous time nature of the
I I
f
model thiS linearization model has the disadvantage of confining the analysls to infinltesinal changes
around the base case equilibrium (see Bovenberg (1985) p 53) Pereira (1986b 1987d) uses an optimization algorithm NPSOL - developped by
rGillMurray-Saunders-Wrlght (1986)) This algorithm is used 10 compute the sequence of
I
29
Shortmiddotrun temporary equilibrium which makes up the intertemporal equilibrium path The
equilibrium conditions are seen as nonlinear equality constraints in the minimization of an artlcial
objective function The prices are normalized to the unit simplex by an additional linear equality I I 1 bull constraint
II ErlichmiddotGinsburgh-Heyden (1987) follow a unique approach in that they use a variant of the
optimization technique introduced by Negishl (1960) The economic equilibrium can be generated i
as a solution of a mathemallcal program Ihe objective function of which is a weighted sum of the
utility functions of the various agents while tho constraints set consists of the market dearing 1
conditions Ginsburgh-Heyden (1985) have extended Negishis resulllO tho case of downward price
lrigidities I
Finally the paper by Jorgonson-Yun (1984) is also unique in tIlat it is the only I
econometrically estimated model Different blocks for the consumption and production Side of the
-
model are separately estimated to provide the necessary structural parameters i
I The diversity of computation techniques is yet another indicator of the exploratory nature of the
body of literature surveyed in this article
43 Equal yield comparisions
The link between a base case and counteriactual simulations Is usually provided by the concept of
equal yield Shaven-Whaley (1977) discuss the meaning 01 equal yield in a general equilibrium
context when government is not allowed to run defICits Equal yield is interpreted to moan constant
public utility Government base case utility is maintained in the counferfactual experiments With
balanced budgets the concepl of equal yield is unambiguous The new equUibrium prices and the
balanced budge condition will del ermine the minimal expendilure and taxes needed to maintain base
30
case publi~ utility Accordingly in general equal yield is inconsistent with equal nominal tax
revenue Some change in tax revenue Is necessary Different tax replacement schemes ate
considered to assure that enough tax revenue is collected This is the approach essentially followed
by most 01 the papers surveyed Only Feltenstein (1985) Goulder (198) and Jorgenson-Yun
(1984) chose not to follow an equal yield strategy
The question is of how to Interprete lila concept of equal yield when the government is alowed to
fUn deflcLts The optimal leve of expenditure for base ease public trliUty can now be taAinanced
bond financed or financed by a mix of bonds and taxation We have several versions of equal yield In
particular equal yield may now be consistent with equal tax revenUE Furthermore some meaSure
ltgtf financial crowding out effects of government deficns can be inferred from the comparislon of the
several equal yield alternatives These issues are extensively discussed ~Pereira (1987b)
44 Price expectations and the dynamic generalization of compensation
Indicators
The sum of equivalent and compensation variations over households is the most widely used
aggregated measure of efficiency gains or losses In the context of $teady~state eomparisions andor
when complete future markets exist andlor perfect foresighl is assumed there are no difficulties
associated with the use of the standard Hicksian indicators In fact correct future prices are known
in these cases
If expectations are not selfmiddot fulfilling andlor future markels are not available a dynamic
generalization is necessary 8allardmiddotGoulder (1985) provide some steps in that direction by
defining an indicator that accomodates periods far into the future when households do not have
perfect foresight but a steady~state prevails On he other hand this issue is more fundamental in
i
[
I
i I I I I [
I I
bullI
31
the contex~ of a Gner1 lemporary equilibrium ftamawor~ In such Circumstances Pereira
(1986a) dev~lops a cnamlc generalization which is obtained as the present discounted value of a i i
sequence of short-fUn optimal expenditure functions consistent with a base case expected future I stream of IlWities I
32
5 Empirical evidenee from selected policy Issues
Dynamic tax models have generated several important results which escaped Sialic modellers
The following is a seleCpoundd set of issuos whIch stress the marginal benefits of dynam~ modelHng of
economic behavior in innovltlttive areas of analysis The discussion of a consumption tax emphasizes
the benefits of dynamic tusehold behavior and intergenerational aspects The study of the
flliminatlon or the re~intToduction of the investsnemt tax credits is made meaningful by the dynamic
modelling of production betlavor Modelling governmenl in a dynamic conlexts allows the modeller
to study the irrrpact of government deflcits Finally a dynamic framework lets us appreCiate the I
relevance of consumer expectations in terms of the evaluation of policy alternatives
i
i
51 Consumption Tax
t
By Its very nature a consumption tax can not be adequately investigated with a static model bull
Fullerion-Shoven-Whaicy (1983) use Ihe model of the US economy s described in
Ballard-Fullerton-Shoven-Whalley (1935) to evaluate the movement from the currenl US tax Isys1em to a progressive consumption tax Since their modeJ Incorporates a Jaborleisure choice
where leisure is an untaxed commodlty their results reflect the fact that both the consumption tax
and Ihe present system are dslortionary
Concentrating on the rntertemporaf distortions they show that sheltering more saving from lhe
current US income tax could improve econornic efficiency even if marginal tax rate increases are
necessary in order to maintain government fevenue~ At first you have a revenue shortfall
However the economy moves to a higher sustatned growth path and uitimatefy lower tax rates
cangenerate the same revenue path Also wages increase in the long run with this policy The
bull
bullbull ~
33 shyt
i
I
I
present value of welfare gi[lS for a potiey of complete savin9 deduction wilh marginal tate
adjustments (consumption lax) is simulated to be around $500 billion to $600 billion 01 1973
dollars
They also investiga1e th 1ngth of ttme it takes th~ economy to adjust to these policy changes
Roughly they estimate the long run to b thirty years although this figure Is very sensitive to the
specification of the savings elasticity - a crucial parameler in this model
52 Investment T~x Credit
Goulder-$ummers (1 S87) address the impact of eliminating the lnvestmenl Tax Cradit (ITC) on
intersectoral capital formation and on economic growth Their model isYa1ticularly adequate to
addies$ this issu~ in that is postulates a forward looking investment m9we with adjustment costs
They show that the eliminating the ITC causes a reduction in the rale of investment Inveslment is
estimated to fall by about 7 in the short-run and by about 12 in Ihe new long run steady state
In lurn a previous policy anncuncement lowers Ihe overall al1ractiveness of Investment and leads 10
a downward shift of the Investrrent profile On the other hand the combined effect of a revenue
neutral simultaneous efimiraron of the ITC and reduction of the corporate tax rates is a long run
reduction ef the capital Sieck by 35 This pattern suggests that a revenue neutral increament in
both Ihe corporale tax rates and the inveSlment lax credits would be preferable 10 the formula
adopted in Ihe US Tax Reform Act of 1986
Pereira (1987d) addresses the impact of the ITe from Ihe oland point of Ihe curTenllax syslem
under the Tax Reform Act of 1986 Given Ihe concern over the potential depressiVe effects upon
savings and investment of the new tax system in conjunction with deficit reduC1ion mechanisms of
the GrammmiddotRudman type a pertinent question is should Ih lTC be e-introduced In the context
bull
34
-of his mod~1 of the US economy Pereira focuses on the trade-off between the distortion in the
intersectoral allocation of capital induced by the lTC the lower tax revenues it generates and
potential financial crowding-out effects on one hand and the positive effects of lowering the relative
price of new capital goods on the other hand Preliminary results confirm the qualitative results in
Goulder-Summers (1987) suggestting a welfare gain of about 2 of the present value of GNP in
the best scenario of absence of replacement taxes Furthermore preliminary results show an
important time pattern to welfare gains with the average benefits increasing the further you look
into the future This pattern is due to the presence of constraints to intersectoral mobility of capital
and inertia in the adjustment towards the optimal capital levels as reflected by the presence of
adjustment costs
I
I
bullAuerbach-Katlikaff (1gB) in the context of their intergenerational model of the US economy
consider the impact on savings capital formation and interest rates of deficit policies and balanced
budget increases in government consumption They conclude that deficit finance and government
consumption can significa-ntly crowd out capital formation and lower the welfare of future
generations However crowding out from deficit finance is a very slow process because it results
from increased government spending over potentially long horizons Also deficit policies may
substantially influence the longmiddotrun interest rates while leaving the short-run rates essentially
unaffected Finally and in opposition to the central role adjustment costs seem to play in both
53 Financial crowding outeff~cts of government deficits
Goulder-Summers and Pereiras models Auerbach-Kotlikoff suggest that the time path of interest
rates induced by a policy of deficit financing seems to be insensitive to the presence of adjustment
costs
bull
35
54 The role 01 consumer expectatlons
The expectations analysis has uncovered one of the benefits of dynamic modelling
Ballard-Goulder (1985) find that the appeal of adopting a consumption tax depends on the level of
foresight possessed by the consumers Furthermore additional foresight may be welfare worsening
This is a second best type of result This tends 10 occur under policies that lead to capital deepening
and declining rate of return to capital over time To the extent that consumers have more foresight
they will be beller equippod to anticipate the fall in the rental price of capital Consequently if a
saving incentive is enacted people save more with myopia than with perfect foresight Given the
exi~tence of taxes on capita and the discrepancy between the private an~--sccial returns to capital
the greater savings with myopia is beller socially Ballard-Goulder find that the welfare gains
from a consumption tax are reduced by about 10 when we move from myopia to farleet foresight
Therefore the attractivenness of adopting a consumption tax seems fairly [obust across these
foresight specifications
I
bull
36
6 Tbe power and weaknesses 01 dynamic modelling the example 01 corporale
tax Integrallon
The Issue of corporate tax integration allows us to sh~w the stre-ngths and weaknesses or the
dynamic approach Empirical evidence using CGE techniques indicates that depending on the precise
scheme of integration and the 1ax replacement methods integratiOn may have substantial effects In
Ihe work of FuliertonmiddotKingmiddotShoven-Whalley (1980 1981) total integration was found to yield an bull
annual static eHielency gain of $4 to $8 billion in 1973 dollars Simulated dynamic gains may be as
large as $695 billion or abcu 14 of the present value of tulUre consumption and leisure in the
US economy These gains result primarily from interindustry reallocalions of Investment and an
improved inter~emporal allocation of consumption
Mora recent work emphasizes hOw consumers asset portfolio decisions and firms finanCial
decisions affect the efficiency gains from integration Slemrod (1980) focusing on consumers
asset ponfoio deCisions finds static efficiency gains which are about twice -as Iarge as those
repOrted iiy Fulierton-King-ShovenWhalley (1961) FuUrlon-Gurdon (1983) focus on the
firms financial decisions They roporl efficiency gains of 6 of GNP from the elimination of the
tax distor1ions favoring debt However when they eliminate the corporate tax and repiace it with
increaSed persona income taxes addlHonal dls10t1ions are created in the optimal labot~leisure
decisions These distortions tend to dominate the analysis slJ(h that the overall effects of complete
integration are very modest Galpr-LuckmiddotTodr (1986) address simultaneously consumers
asset pOrlfollo decisions and the firms financial decisions They also find very modest efficiency
gains from integralion
The above results are important but may be severely biased There are severa aspects of
economic behavior and modelling crucial for the study of Income tax Integration whiCh have no been
i 37
captured innny of the above papers One aspect is the absence of government deficitsmiddot a balanced
budget is assumed It is true Ihal resource crowding OUI is caplured In these mOdels but financial
crowding out induced by government spending is nOI Second investment is not derived from
optimtztlcn behavior Investment behavior passively accomodates endogenous saving decisions As
a consequence the differential impacts of policies in the incentives to save and to invest are not
captured Aso full capital mobility across secors is assumed with instantaneous capital
adjustmenlS towards optimal levels This assumption rules qut different costs of capital across
seclors and therefore differentiated reactions to tax policies changes Finallyhe modelling of both
gov9rnmerlt deficits and of endogenous real and financial investment decisions necessitates the
~nsideraiion of a dynamilt framework and the introduction of financiaJ assets govemment bonds and
iprivate financial assets A dynamic framework also highlights the efficiency effects of Integration on r
ithe ptimal intertemporal decisions The above models are either sIalic (Slemrods and
GalpermiddotltJckemiddotTode(sj or a dynamic sequ~nce of otherwise Slatic model bullbull I Pereir a (1986b) develops a dynamic applied general equilibrium model bull to study the tbull
efficiency effects of integration on the growth and allocation of investmenl across sectors Special
Iattention is paid to the structural affeelS of resource and financial crowding oul induced by
gove mect spending and deficits and to the real and financial Investment reactions across industries
to both tax integration and flnancial crowding out His mode departs from previous work on income
lax Integraticn and for that matler from most 01 the CGIE literalure for tax policy evaluation in
several direcllons It encompasses an endogenous sequential equilibrium struelure founded on
dynamic behavior with flexible expectations Government deficits are optlmaUy determined
Investment decisions are tward looking and the resull of optimizing behavior Several financial
assetsmiddot public and privatemiddot are considered
Simulation resulls Indicale that the welfare gains from Integration under large deficits are at
32
best lJlodest when measured in terms of the GNP The elimination of the corporate tax and its I replacemenl by increased income lax rates yields long run benefits which are never larger than
2 of the presenl yalue ct telr consumption and leisure onder the previous tax regime and 1
under the current tax regime However the short run w~lare effects of full integration tend 10 be
very low and in in some scenarios even negative This is a flew intertemporar pattern of efficiency Ieffects which reflects an adjJsiment lag in the interindustry investment decisions due to the
existence of costs of adjustment On the of her hand partial integration achieved by excluding I dividends from the corporale lax base syslematically yields negalive efficlencyeffeclS In Ihe long Inm these can be as high as 3 of present value of the future value of consumption and leisure This
is a new second best effect suggesllng Ihat less than complete integration may have perverse I efficiency effects
The dynamic slructure of PerclrBs model snowing for an improved freatment of real investment
declsions and government deficts provides a potential setting for a more accurate measure of the
costs and benefits of inlegration The simulation resullS suggest lower benefits and an intertempora Ibull
pattern of growing benefits
Simulation results also clearly suggest robust negative effects from partial integration How
reliable is this result If dJvidends were deductible from the corporate lax base (as are interest
payments) the preferenlial Ireatmant of debl over equily would be ellminaled Corporallons should
be expecled 10 react by decreasing the optimal deWeqully rallo However corporale financial
gecisions are exogenously set In Pereiras model as in aft of the other models surveyed that go as far
as dealing with the issue
Also withdrawing dividends from Ihe corporate tax base is a way of eliminating Ihe double
taxation of dividends at belh Ihe corporate and personnal Income levels How will the optimal
aliocellon of household saving accomodate that change in Ihe relalive return to corporale equity
39 I
t However hOusehold portfolio decisions are exogenouly set in Pereiras model as in all of the other
models surveyeltj go as far as dealing with the issue
IIt is sate to say that the evaluation 01 the benefits 01 partial Integration as defined above are
sevrey underestima1eO Meanwhile the distortionSoenerated by the increase in the marginal
personal tax rates designed to ralse the revenue foregone by the dividend exclusion are fully
aCCltlunted for A good measure of the costs of integration together With a poor evaluation of the
benefils may very well be the reason why partial integration is Simulated 10 yield neg alive
eHikiency gains
On a different vein as most of Ihe model surveyed here Pereiras is a closed economy model It
is legitimate to wonder how the resullS would change If international capital flows were anowed
IThis almost certainly would affect financial crowding out since a good d~al of the capital inflows
could be used to finance public debt
For these various reasons the results should be treated as preliminary but they strongly I suggosl that the dynamic structure provides valuable new insights into the corporate tax integration SSU2
40
7 Concluding remarks I I I i
What has been acccmplished Ihis far The success of Ihe CGE research in laelding the challenge
of dynamic modelling can not be denied
I iGreat progress in the modelling of household behavior has been achieved Promising
developm~nts in the modeUing of production and government behavior have also been accomplished
Interesting innovations in the concept of equilibrium and computation techniques were discussed
The policy analysis was greatly enriched by considering transitional effects together with longer
run steadymiddotstate equilibrium paths generalized equal yield strategies accomodating different
financial crowding out impacts and generalized dynamic policy evaluation indicators
An important set of issues have been addresed by the models surveyed Traditional issues
focusing on the effects of capital taxation and consumption taxes have been pursued In terms of
capital taxation for example the intertemporal nature of the issue was enhanced by allowing an
optimal evolution of capital stock in the economy and forward looking optimal investment decisions
In Andersson (1987) GouldermiddotSummers (1987) and Pereira (1986b 1987d) this is coupled
with a improved treatment of several tax provisions like investment tax credits and depreciation
allowances In terms of the consumption taxes developments in the specification of Intertemporal
household behavior the introduction of overlapping generations and the modelling of
intergenerational links via bequest motives as well as a closer attention to demographic evolution
provided a much improved economic setting for the understanding of the several aspects of the
problem
Furthermore dynamic modelling has permitted the addressing of a set of new issues Policy
Issues ranging from dynamic tax policy analysis to the evaluation of exogenous changes in
government expenditures to the impact of different methods of financing government expenditure to
4 1
the importance of tinanciai croHrii1g out to 1he relevance of consumorS expectations in tne policy
results have been addressed
Despile all of the progress and in part due to such progress several avenues are wide open for
much neded additional research The following seem to be the most Interesting and promising
areasmiddot First the specincation of liquidity constraints to household behavior may help to obtain a
better understanding of policy chenges that affect directly the interest rates in the economy Second
major attention needs to be paid to the incorporafion of a roraign sector (with open capital markets)
Third some efforts are needed in terms of finding adequate computation techniques now that
dynamic modelling eally makes the curse of dimensionality worse than eVOrbull
The modelling of financial markets is the single most unsatisfactory speet of the dynamic
models surveyed here~ The problems associated with provlding a mo~ adequate treatment of
financial markets and in general endogenous and optimal financial deci$ions bull both at the corporate
and at the personal levelsmiddot have 10 be addressed That necessarily requires Introducing uncertainty
into the models To b adequate this must go well beyond the mere assumption of point expeetalions
with the whole array of modelling implementalion and evaluation problems thai generates
bull I i
42
REFERENCES
Abel A and O Blanchard 1983 An Intertemporal Model of Saving and Interest Econometrica
51 pp 675-92
~ Andersson K 1987 Tar-ation of Capital in Sweden - A General Equilibrium Model Ph D
Dissenation UnIversity of Lund Sweder
Auerbach A L KotlikoH 1983 National Savings Economic Welfare and the Structure of
Taxation in M Feldstein ed Behavioral Simulation Methods in Tax Policy- Analysis Chicago
University of Chicago Press
Auerbach A l Kotlikoff 1984 Social Security and the Economics of Demographic-
Transition in H Aaron and G Burtless eds Retirement and Economic Behavior Washington DC
Brookings Institution
Auerbach A L Kotlikoff 1987 Dynamic Fiscal Policy Cambridge Cambridge University
Press
Auerbach A L Kotlikof and J Skinner 1983 The Efficiency Gains from Dynamic Tax
Reform Internatiooaj Economic Reyiew 24 pp 81middot100
Ballard C 1983 Evaluation of the Consumption Tax with Dynamic General Equilibrium
Models Ph D Dissertation Stanford University
Ballard C 1987 Tax Policy and Consumer Foresight A General Equilibrium Simulation
Study Economic Inquiry 25 pp 267-284
Ballard C and L Goulder 1985 Consumption Taxes ForeSight and Welfare a Computatonal
General Equilibrium Analysis in J Piggott and J Whalley eds New Deyelooments in Apolied
General Equilibrium Analysis Cambridge Cambridge University Press
Ballard C D Fullerton J Shoven and J Whalley 1985 A General Equilibrium Model for
t 43
Tax Eoliey Eyalua~iQO NBER University or Chicago Press
Borges A (1987) bull A Survey of Compula1ional General Equilibnum Models CECD Economic
Slud~ 8
Bovenbig L 19B3 imperfect Capital Mobility in a Perfect Foresight Herberger Model
Mimea UC Br~eley
Bovenberg L 1984 Capitel Accumulation and Capital Immobility Qmiddottheory In a Dynamic General Equilibrium Framework PhD dissertation University of California at Berkeley t
Bovenberg L 19850 Dynamic General Equilibrium Tax Models with Adjustment Costs in A iManne ed bull 1985 EtQnomrc ECUjilbrium Model Formulation and Solution Mathematical
lProgramming Siudy 23 Arnsterdam NorthmiddotHoliand
Bovenberg L t 1986 Capital Income Taxation in Growing Open Eeondmies Journa of Pubtic
Eamp(omlcs 3t pp 347middot376 IBovenberg L ald W Keller 1981 Dynamics and Nonlinearlties in Applied General
Equilibrium Models Internal Report Mlmiddot81middot208 Department for Statistical Methods CBS bull
Vooiburg The Netherlands
Charnley C 1981 The Welfare Cost of C~pltal income Taxation In a Growing Economy
IJournal of Political ECQnQm~ 89 3 pp 468middot496 I
Charnley C 19810 Efficient Tax Reform in a Dynamic Model of General Equilibrium I Mimeo Yale University I Charnley C 1982 On the Optimal Taxation In Stylized Models 01 Inlartemporal General
Equilibrium Mimeo Yale University I Decaluw B and A Manens 1985 Pays en Oeveloppement at Modeles Calculables OEquilibre
General - Une Revue de al litterature empirique Monographie Sene 9 Volume 3 Centre de
Recherche et Developpemont Economlque Unrslt de Montreal
44
bull
I I
I
I
1
Erlich $ V Ginsburgh and L van der Heyden 1987 Wher~ do Real Wage Policies Lead
Belgium bull A General Equilibrium Analysis European Economic Reyiew (forthcoming)
air R and J Taylor 1983 Solution and Maximum Likelihood Estimation of Dynamic
Nonlinear Ratioant Expeclations Models Ecooometrica 51 pp 1169middot1186
Feltenstein A 1984 Money and Bonds In a Disaggregated Ec~nonYin Searl 1-1 J Shoven
Eds Aoolied ~eoeral EqJiliDCiulD tioalYsis Cambridge University Press Cambridge
Feltenstein A 19B6 A Dynamic General Equllibrlu Analysis of Financial Crowding Out
Theory with an Application to Australia Journal of Public Economics 31 No1
Fullerlon D R Gorden 1983 A Reexamination of Tax Distorllons in General Equilibrium
Models in M Feldstein Ed Behayora Simulation Methods in Tax pOlicy Analysis Chicago
University of Chipgo Press
Fvllerlon D J Shaven and J Whalley 1983 Replacing the US Income Tax with a
Progressive Consumption Tax JoulOal of Public ECQnomics 20 pp3middot23
Fullerton D Y Henderson J Shaven 1984 A COmparision of Methodologies in Empirical
General Equilibrium Models of Taxation in Searl H J Shoven Eds ~pQlied Geoaral Jguilibrium
lIoalyss Cambridge University Press Cambridge
Fullerton 0 A King J Shaven and J Whaney 1981 Corporate Tax Integration in the United
Statesmiddot a General Equiiibdufi1 Approach American Economic Review 71 pp~ 677~691
Galper H R Lucke E Tader 1985 Taxation portlor Choice and the Allocation of Capital A
General Equilibrium Approach Mimeo
Gill p W Murray M Saunders M Wright 1986 Users Guide for NPSOL (Version 40) A
Fortran Package for Nonlinear Programming T echnicsl Report SOL 86middot2 Deparlment of
Operations Research Stanford University
Glnsbur~h V and L van der Heyden 1985 General Equilibrium with Wage Rigidities An
45
AppJication to Belgium in A Manne ed t Economic EQumprit1m~ Made Formulation and $oJioo
Mathematical Programming Study 23 Amsterdam NorthmiddotHolland ~
Goulet J 1968 1 Adjustrent CO$1S in the Theory o(the Firm Reyiew of EconomiC Studies 35 bull
pp47middot55
Goulder L 1985 Tax-Financed versus BondmiddotFinancedGhanges in Governent Spending
bull Mimeo Harva(d UniversilYmiddot
Goulder l J Shaven and J Whalley 1983 Domestic Tax Policy and the Foreign Sector The
importance of altemative foreign policy formulations to results from a general equilibrium tax
analysis model In ~n methods in Ibe antico of income from caoia1 ed Martin
Feldstein Chicago University of Chicago riSS
Goulder l L Summers 1987 Tax Policy Asset Prices and 3rowtly ~
A General Equilibrium Analysis NBER Working Pper No 212B
Grandmont J 1982 Temporary General Equilibrium Theory Ch 20 in Handbook 01
Mathemalical Eccnornics Vol II NorthmiddotHoliand
Harberger A 1959 The Corporate Income Tax An empirical Appraisal In lax Revision
Comoendjum Voll House Cornmitee on Ways and Means Washington DC Govemment Printing
Office
Harberger A 1962 The Incidence of the Corporate Income Tax JOULn1 of Political
Economy 70 pp 215middot240
Harberger A 1964 Taxation Resource Aliocalrlon and Welfare fn lb Role of Dir bullbulll and
Indjresl Taxes io tbe Fede[al Reserve s~I~Ill Princeton Princeton Unlveresity Press
James J 1985 New Developments in the Applications of General Equilibrium models to
Economic History in Piggott aand Whalley Eds
Jorgenson D and K Yun 1984 Tax Policy and Capital Allocation Harvard Inslitute of
bull
I
I f
Economic Research WP 1107
Judd K 1965 ShonmiddotRun Analysis of Fiscal Policy in a Simple Perfect Foresight Model
Journal of poJiHcal 5CCJiW1l pp 298middot319 r
Upton D J Potbc J Sachsa nd L Summers 1982 Multiple Shooting in Rational
Expectations Models fuoometrlc 50 1329middot1333
Manne A ed 1985 fconQrnjc Eoullibrium Model FOrIDujaljQo and Solution Mathematical
Pogrammlng Study 23 Amsterdam NorthmiddotHolland
Manne A 1985 Or the Formulation and Solution 01 Economic Equilibrium Models In A
Manne ed ECQOQlli-Jr~ibmiddotum ~tQdel EQlJulailon and Solution Malhematical Programming
Study 23 Amsterdam NorthmiddotHolland
Mansur A and J Whalley 1984 Numerical Specification of Applied General Equilibrium
Models in Scarf H J Shoven Eds Aoplied General Eouillbrium AnaiysectIsect Cambridge Universily
Press Cambridge
Merrill 0 1972 Applications and Extensions of an Algorithm thaI Computes FixedmiddotPoints of
Cenain Upper Semimiddotcominuou Point to Set Mappings PhD Disserlation Unlvel$ity 01 Michigan
Negishi T 1960 Welfar Economics and Exislence I an Equnlbrlum for a Competitive
Economy MelroeCOOQ11ka 12 92middot97
Pereira A 1985 On the Existence and Uniqueness of Optimal Ou1put Path lor CRTS Firms in
Adjustment Costs Technologies Mimeo Stanford University
Pereira A 1985a Consumer Expectations and Ihe COSI 01 UllUty In an Intertemporat
Framework Mlmeo Stanford University
Pereira A 1985b Corporale Tax Integrallon and Government Deficits A Dynamic General
Equilibrium Anaiysls Mlmeo Stanford University
Pereir A 1987 On Ihe Computation of the Users Costs of Stocks Ecooomi (forthcoming)
47
Pereira A bull 1987b Equal Yield Alternatives and Government Oeficlts Mimeof Stanford
University
Pereira A 1987c A Dynamic Applied General ~uHlbrium Model for Tax Policy Evaluation
Comlletamp Docurnent2tion Mimeo University of California San Diego (in progress)
Pereira A 1987d Should the investment Tax Credit be Rei(ltroduced Mlmeo University
of California San Diego (ir progreSs)
Piggott J and J Whalley eds 1985 tJew QeyeQpments in Applied General EQuilibrium
bullenalysjs Cambridge Cambridge University Press
Preckel Pbull 1985 Alternative algorithms for computing economic equilibria In A Manne Ed
Radnor R 19amp3 Equllibrium under Uncertainty In K Arrow M Intrilligaor ods
I ibull
Robinson S 1 g86 Mullisectorai Models of Oeveloping CountrieS a Survey Oivision of
Agricullure and Resources UC Borkeley WP 401
Scarf H 1967 The approximation of fixed points of a continuous mapping ~MM Jouwal of
epplid Mathornallcs 15 pp 1328middot43
Scarf H with T Hansen 1973 00 the Computation of EconomiC Eguilibria New Haven Yale
University Press
Scarf H J Shoven Eds 1984 61l1llied Genera Elujlibrium 6nalyssect Cambridge University
~ Press Cambridge
Shoven J 1976 The Incidence and Efficiency Effects of Taxe on Income from Capital
Jurnal of ramie1 Economy 84 pp 1261-83
Shoven J 1983 Applied General Equilibrium Tax Modemng IMF Slaff PaoerS 30
Shoven J and LB Simon 1986 Share Repurchases and Acquisitions An Analysis of which
Firms Participate NBER Working Paper No 2243
bull
48
IShaven J and J Whalley 1972 ft General Equmbrium Calculation of lhe Effects of
Differential Taxation of Incoma from Capital in the US Jauroe of public Econgmics 1
pp281middot321 (
Shoven J bull and J Whalley 1973 General Equilibrium with Tax$s a computation procedure
and an existence pr00f ReviEhi of Economic Studies 40 pp 475middot490
Shoven J J Whalley 1977 Equal Yield Tax Alternatives A General Equilibrium
Computa~ui0nal Tfchnique J~HJrnal of PUblic EconomIcs 8 pp 211~224 bull
Shaven J J Whalley 1984 Applied GeneralmiddotEquilibrlum MOdels of Taxalion and
International Tr8de An Introduction and Survey Journal of Economic literature 23t pp
1007middot1051
Slemrod J 1980 A General Equilibrium Model of Capital Income Taxation Ph D bullDissertation Harvard Univnrsity
Slemrod J 1983 A General Equilibrium Model of Taxation with Endogenous Financial
Behavior in M Feldstein edbull Behavioral Simulation Methods In Tax Policy Analysis Chicago
UniVerSity of Chicago Press
Stone J 1985 Sequential Optimization and Complementarily Techniques for computing
Economic equilibrium in A Manne Ed
Summers L 1981 Taxation and Corporate Investment a q-Theory Approach BroQkinas
Summers L 1985 Taxation and the Size and ComposHion 01 the Capital Stock An Asset Price
Approach Mlmeo Harvard University
Whalley J 1975 A Geneal Equiloibrium Assessment of the 1973 United Kingdom tax
reform Economica 42 pp 139middot61
1JLTIllOS JIQll~)ltG PAPERS PUBLICADOS bull
I
I I
j
112 70 - LUCENA iJiCIIJo ~ IIEnvironrilent Honitoring and Orqanization Structure T (Setellibro 1987)
nI 71 - L~ENA Diogo A Note on the Representation of Information Stxueture (SeteJObro 1987)
nQ 72 - LUCENA D1CgO Envuorment 111)nitoring and Organization
Structure II (Setelilhro 1987)
ril 73 - AlmmE3 Antonio Pais e CAsPAR Yitor bull Tributacao Inoentivos e InvesUlllentos Analise Qualitativa noYellibro 19S7)
n2 74 - liAlltOSA rose Pedro Opt1lllal lIage R1gidity A SUggestedtlthodology to Teet tb Theory and an Application (Outubro 1987)
nI 75 - CABRAL L 11 B Three lIotes on SYJIIlletric Gemes witt itsYJIIlletric Equilibri (NoveJObro 1987)
n2 76 - LUCENlt D10gOTo Search or Not to Search (Fevereiro1967)
nQ 17 - SA Jorge asconcelos euroI A Theory of SynergyB (Fevereiro1987)
n2 76 - VILARES tanuei Jose Os Be IntenaMios IlIIportados ColIIO Factor de Produao (Julho 1987)
nQ 79 - SA J~rge VasCQncel03 e HoY to Copete and Co~unicate in tature Industrial Products_ (aio 1988)
n2 80 - ROIl Rafael Learning and Capacity Expansion in a New larket Under Uncertainty (Fevereiro 1988)
n2 81 PEREIRA Alfredo llarvilo Survey of Dynamic Computational General EquilibriUlll Hodel for Tax Policy Evaluation (Outubro 1987)
Qualquer in ormacao sabre o 1IIorlltlng Papers ja publlcados sera prestada pelo Secretariado de Apoio aos Docentes podendo os ~e3OSf
ser adquirido5 ne Secyo de Vlndas da Faculdade de EconollLia UNL na Travessa Esteviio Pinto Cltpolide - 1000 LISllOA
15 I I ~3 ~ssues in the modelling Of dynamic behavior
We will 100k firt at the speclficalion of economic behavior of consume(s~ producers
goyemment and Ihe r~sl of Ihe world The modemog of corporale and household financial decisions
I is then addr5sed Finally Ihe eencept of equilibrium is discussed
I
i
I 1
31 Consumers behavior
Early efforts 10 build dynamic features into the economic behavior 01 consumers are due to
Ballard (1963) and Auerbach-Kotllkoff (1983 1984) Their work has been closely followed by
several subsequent authors In faet dynamic household behavior Is the single most pervasive aspect
amohg the models surveyed
All the models Inc-O1orale some loro of lila-cycle behavior Household behavior is dewmloed
by the maximization of an addWVely separable time invariant intertemporal utillty function The
utility lune1ion is defined oyer the domain of the consumption goods In the economy In most of the
models leisure is also an argument in utility so that labor supply is optimally determined
Utility maximi4ation is subject to a lifetime intertemporal budget constraint which equalizes
the presen1 value of consumers income and expenditure More reeentiy in Andersson (1987) bull
Bovenberg (1985) and Pereira (1986b 1987d) the constraint i$ defined as a sequence 01
recursive equations of motion on wealth This has the potential advantage 01 accomodating liquidity
constraints However It should be recognized that in the absence 01 liquidity constraints (ie when
consumers are free to completely borrow against future income) the two specificallons of the
househofd ccnstraint are essentiay equivatent Furthermore tn both versions saving is optimally
determined as a way of translaring wealth intertemporally
1 6
Some lodels now have a sophisticated dynamic specification of hoosehokl In Ballard (1983) -
BallardmiddotGoulder (1985) and Auerbach-Kotlikof (1983 1984 1987) consumers behavio is
embedded in an iterQEneraional setting Each of them has S5 age cohorts simulataneously alive
Som-e measure of jntergener1ional altruism is considered in the form of bequest motives In Ballard
(1983) and Ballard-Gouldr (1985) consumers derive utility directly from bequthing parl of
their wealth Therefore bequests assume 1he form of a scrap value of terminal wealth of a
generation In AuerbachmiddotKotlikof (1983 1984 1987) each generation empathyz bullbull with Ihe
utility of future generations ovar bequethd wealth In addition Ballard (1ge3) includes detailed
demographic poiections wfill the intent of incorporating in the policy analysis the effects of the
bull Post World War tI baby-boom
In the real world household decisions also include the optimat allo~atiOn of saving among
shyalternative physical or financial assets Theories of household portfolio behavior are relatively
Iwen~establshed However they involve uncertainty as a crucial element Accordingly they have
middotnot been incorpora~ed in the context of the determinIstic dynamic models we sutyey here We will 1 Come backt-O this Issue below
32 Producer behavior
The efforts to build dynamic features into the economic behavior of producers are more recellt
and less widely adopted The first attempts are due to Bovnberg (1984 1985) and Summers
(1965) Dynamic behavior has been more fully incorporated In the recent models of Andersson
(1987) Auerbach-Kotlikoff (19B7) GouldermiddotSummers (1987) and Pereira (1986b 1987d)
Part of the reason why production side dinamlcs has been more slowty adopted is the weak
supply of accepted theories referring to the dynamic behavior of the firms In the models referred
I 17
to above gynamlc producion and investment b~havior are middotinduced by the existence of capital
adjustment costs and linked to Tobins q theory Adjuslment costs are designed to capture both the
incomplete mobility of capital aoross industries and installation costs ~e the costs of adjusting
capital towards its cpHmai level)
Adjllstment costs can be conceived as internal to the firm and measured in terms of foregone
outpul along the lines of Lucas (1967) Alternatively adjustment costs can be viewed as actual
external cests incurred together with Ihe purchase costs along the Unes of Gould (1969) With the
exception of Pereira (198Gb 1937d) all of the CGE authors follow the former approach
The firms in the economy maximize their market value as the present discountad value of the
future stream of dividends In Andersson (1981) and Pereira (198Gb 1987d) firms are seen as
maximizing the present discounted value of net cash flow Maximization is constrained by Ihe
adjustmenl cosl lechnology and ~n equation of motion describing Ihe evolution of the capital stock
It should be noted that undereogenous divldendrelention rules the two problems are equivalent II
should also be nOled Ihat a salisfectory economic rational for th bullbullistoo of dMdonds with the
present tax code is missing in the profession (Shoven (1986))
The models with static formulation of producers behavior are characterized by passive
investment behavior Investment merely accomodates to saving in the eccnomy With a dynamic
formulation induced b adjustment costs real investmen decisions are forward looking Investmenl
Is endogenously and optimally determined by the firms A fundamental difference between the
shonmiddotrun in which capital stock is given and longmiddotrun in which Ihe level of capital is allowed to be
optimally determined is emphasized
This extra richness of produc1ion dynamics is not without costs A careful look at the summary
tables will clearly show an inverse relation between the adoptlon of dynamic features in production
and the level of disagreggation of the production side of the economy In fact with production
bull
j
1 6
dynamics $e dimension of the problems is immensly increased Let IS be more specific
In a static framework with constant returns to scale production technology the output level ismiddot
indeterminate The optimal aliocation of inputs can be obtained by cost minimization with any
feasible output level generatillg zero profits Such zero profit conditions are used to solve for the
output prices in terms of the factor prices and hencemiddot to reduce the dlmensionality of the problem
from the -number of commodities and factors to the number of factors Thus even if the model deals
with 30 production sectors the computation of an economic equilibrium can take place using only
the dimensionality of the primary inputs in the economy
Now under certain regularity conditions on the production and adjustment costs technologies
leading to enough concavity of the optimality objective the intertemporal output path for the firm is
endogenously optimally and uniquely determined even with constant relums to scale technologies
(see Pereira (1985 1987a) Accordin~y with adjustment costs in the model optimal profits
will in general be non-zero This result has crucial implications for the computatiQn of equilibrium
in the model~ with production dynamics With adjustment costs no reduction of dimensionality is
possible The curse of dimensionality returns as a binding constraint
The introduction of adjustment costs adds another significant complication to the model With
capital being less that perfectly mobile in the economy different rates of return on capital will
exist in different sectors This is a difficult problem to tackle conceptually in the absence of
uncertainty Also in the real world producer maximization choices include also its choice of
financial ratios (debVequity) and payout rates (dividendretained earnings) These financial
subjects are difficult and much Clf this behavior is still taken as exogenous by the modellers We
will come back to these issues below
middot i
19
bull 33 government behavior
middot
Two issues dominate the modelling of government behavIor First govemment behavior has
Ibeen typically seen in Ihe CGE liIerature for lax poliCY evaluation as constrained by yearly balanced
I budsets (see for example 8allard-Fullrton-S~ovn-Whafley 098Sraquo The analysis of
govemment defidlS and public debl in a CelE context requires a dynamic selling Second the level of
government expenditures is either exegenouty given as in Auerbach-KoUikofi (1984 1987)
6ovenbrg (1984 1985) Feltenstein (1984 1986) and Jorgenson-Yun (1984) or
endogenously (but no optimally) determined by the balanced budget conditions as in
6allard-Fullerton-Shoven-Whalley (1985) Andersson (1987) Erllch-Ginshurgh-Heyden
~
(1987) Gouder (19aS) and Gouder-Summers (1987) In the second case the composition of
bull public expenditures is often optimally determined However the leve~ of government expendi1ures
can only 00 endogenously and optimally determined if the government is seen as an optimizing agent
and is allowed to run deficits
The first attempts to deal wiU the government defiCits in CGE tax models are due to
Auerbach-Kollilltoff (1984) and Feensain (1984) In Auerbach-Kotlikof (1984) government
expenditures afe exogenous They grow at the rate of growth of population However given 1he tax
structure and tax revenues yearly deficits and surpluses are atTowed subject to an in1ertempond
constraint that the present value of future tax revenues equals the present value of future
expenditures
In Fellenstein (1984 1986) government expenditures are also exogenously given He also
allows the government to run deficits to finance expenditures in excess to 1ax revenues However
surpluses are returned to CQnsumars in the form cf transfers Accordingly government ts not
subject to any constraint regarding the future repayment of public debt
In GouJder (19851 the 9Vernmenl maximizes a static social welfare ~nctio subject to a-
bull
20 I i
I
I I I
I bull
balanced budget conSirtlL This follows the optimal alocatkm of government expenditures alpng
the lines of Baliard-FHeiton-Shoven-Whalley (1985) Ttle model Is generallzea to allow for
exogenous changes in HIe time path of government expenditures Two financing alternatives are
consjdeft~d and contr8s1ed EldditionaJ tax revenues and bond issuance
Pereira (198Gb 1137d) auempts to address both the incorporation of deficits and the
determination of govement expenditures The path of government expenditures and the path of
delicilSlsurpluses (ano therefore the path lor debt) are endogenously and optimally determined The
government is seen as mltJximizing n intertemporal social welfare function given the 13x structure
Optimization is subject to a sequence of recursive equations of motion reflecting the evolution of the
ptlbllc debt allowing fDr government budget imbalances It should be stressathat this specification
01 the government cOMtcalnt is equivalent to the specificaUon in Auerbaeh-Kotlikoff (1983 1987)
when no liquidity constraints affect -gove~nment behavior
The extra richness in the treatment of government behavior allows tfte examination of
government debt poices in a truly dynamic setting Also financial crowding out effects induced by
government deficits can be analyzed (see Auerbach-Kottlikoff (1987) Feltenstein (1986) and
Perer (198Gb 1987e))
34 Foreign sector
Most of the open economy models follow the assumptions of balanced trade with import and
export net demands characterized by constant elasticities along the line of Ballardmiddot
-Fullerton-Shaven-Whalley (1965) Such is the case of Salfard (1963) Ballard-Goulder
(19S5) and Goulder-Sllmmers (1987)_ In Feltenstin (1985) the rest of the world is treated as
21
an addjlion~ consumer group
Bovcoberg (1986) develops a model In which two economies lIr considered each following
intertempoml perlect foresight paths These economies meet in the international forum Their
trade relaroMhips are characterized by yearly balanced trade accounts
None of the new generation of dynamic cae tax models has yet IncorPorated the lnternational
capital flows as done in aoulder-Shoven-Whalley (1983) for the earlier Ballarrshy
-Fullertonmiddot Shaven-Whalley (1985) model This is unfortunate as there is an important
Intertmporal aspect to international lending The first al1empts along Ihese lines are due to
Andersson (1987) and Erlich-Glnsburgh-Heyden (1987) Andersson (1987) in his model of the
Swedish economy adopts a closed economy appraoch in which rates of return in the domesJic
economy are largely determined by the international capital markets Fidinterest rate induce
Intlnatlonal capital flows which determine and finance the International lrade imbalance In
lurn In ErlichGinsburghHeydenmiddot (1987) foreign trade Is generaled according to an
intertemporal Irade welfare lunction with constant import and export elasticities In the shortmiddotrun
they allow inlernational trade inbalances which generate capilal ftows to the domestic households In
the long run however trade balance is assumed
35 The need for financial markets
A dynamic economic Wucture not only provides the ideal environment 10 model many features of
economic behavior it also permits one to Incorporate the financial side of the economy If the
government is allowed to run deficits the question 01 deficll financing automatically follows If
investment is optimally dewmined and returns to capital are different aClOSS sectors problems of
investment financing arise If there are seeral final1cfal assets In the economymiddot government
4-
I i
ibull
I I I I
I
j
22
bonds privpte bonds and equity (or simply physical capital installed inmiddot different sectors) the
problem of allocation of saving among assets with potentially different returns arises
The papers surveyed var greatly with respect to l~e extent of their attention to the financial
side of the economy At Clne extreme are the models in ~anardmiddotFuliertonmiddotShoven-Whalley (1985)
Andersson (1987) Ballltld (1983) Ballard-Goulder (1985) Bovenberg (1985 1986) and
Erlich-Gjnsburgh-Heyden C19B which are devoted exclusively to the real side of the economy
In turn Auerbach-KotliKoff (1984 1987) Feltenstein (1984 1986) and Goulde (1985)
allow for government debt In these models saving finances changes in government debt and physical
capital Private and publlc 2ssets are perceived by the households as perfect substitutes The
allocation of saving merely adjusts to the relative demands for funds
Feltenstein (1984 1986) is the only model surveyed which introduces money Government deficits are financed by issuing money and bonds according to an exogenously given rule Money is
demanded by consumers for transaction motives and an exogenously given fraction as a store of value
On the other hand government bonds and physical capital are the vehicles for the intertemporal
transfer of wealth
Summers (1985) and GOlldermiddotSummers (1987) introduce a whole menu of financial assets shy
firm specific equity capital Different assets earn different rates of retum However such rates
are equal up to constant and exogenous sector speCific risk premia Therefore the introduction of
constant exogenous risk premia as helpful as it may be in the context of calibration does not solve
the main issue of the non-optimality of the allocation of saving Also talking about risk premia in a
deterministic context is somewhat unsatisfactory
Pereira (1986b 1987d) also introduces a whole menu of assets private and public bonds and
firm specific equity However all the assets are expected to yield the same rate of return and
therefore perc~jved as perfect substitutes The different asset types allow consideration of
23
exogenOU$~8tUequjty triO dividendlrc~enlion rufes and therefore seve-ral sources of investment
financing bond equity and retained earnings
The nonmiddotcptimalHy of the allocation of saving and the absence or exogenelty of corporate
financial rules is _a reftectlon of the limitations of the determInistic approach Either in a context of
perfect anticipation oi the future prices Of in general within the t$atm of point price
expectations all the papers follow a deterministic approach Under such circumstances consumers
either expect diHerent rales of return (inclusive of risk premium) across assets in which case
they wilt buy only one asset (that with highest rate) Of they expect equal rat~s of relurn in which
case Ihey are indifferent about the asset composition of Iheir portfoliO There is no way of
Iradingmiddotoff rales of return and risks to obtain an optimal interiof solution to the problem of the
allocation of saving
The most advanced contribution in the modelling of saving allocation in a CGE setting is due to
Slemrod (1980 1983) in the context of a static one-period model In his model consumers act
according to a two stage separable decision process They Hrsl decide on how much to save Then
they decide on the allocation of saving according to an Indirect uttlity function dependent on the rates
of ratum and variances offered by the different assets in the eoonomy The sourCe of riskiness in the
ampConory comes from an uncertain marginal product of capital On the other hand aside from
portfolio decisions the res of the economy is insulated from uncertainty
The most complete contribution in terms of the treatment of the corporate financial rules is
FulienonmiddotGordon (1983) In a variant 01 Ihe BailardmiddotFulienon-ShovenmiddotWhelley (1985) model
Fullerton and Gordon have capital intensity and optimal financial decisions Jointly deter~ined
through a two siage process The cosl of financing capital is minimized by trading off the tax
advantages of debt against the epected real bankrupcy coSls inherenl with high debVequity ratios
Given the optimal debtequity ratio the level of inveSlment Is chosen such that at Ihe margin the
24 l
bull
rerurn pn ~uity equas the return on bonds plus an exogenous risk premium
36 Market assumptions equilibrium and expectations
Virtually all of the papers arc characterized by Walrasian market clearing assumptions All
markets are perlect1y competi1ive Atomistic competition among agents 1s assumed even thoughmiddot only
a flnile number of agents is considered Virtually no mark~t disequilibria or price stickiness afe
considered The only exception is ErlichmiddotGinsburghmiddotHeyden (1987) bull In thei model of the Belgium
economy I the _wage rate is fixed in the short-run Therefore in the shOrlwrun desequilibrium in
the labor market will generate endogenous unemployment However in lh$ long-run all prices
Including the wage rata are Ilexible and accordingly ali markets ciear
Given the dynamic nature of behavlor in the economy markelmiddotcfearing prices in each period
depend on expectations of fture prices and on tax variables In the economy There are essentially
two ways at interpreting the economic equilibrium in such a dynamic conte~bull If future prices are
perfectly anticipated (io expectations are self-fulfilling) a perfect foresight equilibrium
prevaHs Then future actions are merely the implementation of current decisions for future
periods However jf ptica expecla110ns are not perfect (Le agents make mistakes w1th re$pect 10
future prices) then a temporary or shortrun equilibrium prevails Markets ciear and clearing
prices depend on fu1ure price expectations Current plans about the future are typically not
precisely implemented They will be revised as more or better information becomes available to the
economic agents See Grandmant (1982) for a comprehensive survey 01 the temporary equilibrium
literature
Wilh the exception of Gould (1985) and Pereira (19S6b 19S7d) all the models surveyed
adopt th~ concept of a perfect foresight equilibrium In turn BaliardmiddotGouldr (1965) considers a
J
2S
flexible am9unt of foresight In terms of the number of years over which price movements are ~
foreseen Pereiras model (1986b 1987b)) is flexible in a somewhat different way in that it can
indude any range of foresight from myopia to perfect foresight The choice between the perfct foresight and lemporary equilibrium is ultimately to be made on
philosophic grounds It can be argued that less than perfect expectations Imply that agents are
irralional in some way (see Auerbach-Kctiikoff (1987 p 10) However the reverse argument can
be made One can question wiether agents are reaHy rationar and perfectly knowledgable about
future prices
Recent evidene of Balird (1987) Ballard-Goulder (19851 Goulder (19851 and Pereira
(1986b 1987d) suggests that the choice in modemng expectations is an important one They show
that the degree of foresight inl0 thc future (ranging from perfect foresight to-myopic expectations)
may have dramatic impacts on the pollcy conclusions of Ihe model Accordingly the best research
strategy may b to design models which are flexible enough 10 allow for different rules regarding the
formation of expectations With the exception of the articles just mentioned sensitivity analysis
bullhave previously not been performed along this dimension
In terms of implementation the two concepts of equilibrium - perfect foresight and temporary
equilibrium have different implications The dimensionality of the equ1ibrium soll1ion algorIthm
is involved Suppose we have a model with ten markets to be run for a period of 50 years Aside
from normalization a perfect foresight model implies computing prices In 500 dimensions while a
-t~mporary equilibrium model requires sorving SO equilibria each in ten dimensions Given that
computational speed often varies with the cube of the number of dimensions the lemporary
equilibrium formulation is potentially strikingly more feasible However Ballard (1987) and
Goulder-Ballard (1985) have devoloped techniques to greatly speed the computation of a perfect
foresight equilibrium
26
j
I
4 Jrnplemenlalion and Issues on policy evaluallon
41 Calibration and equilibrium comparlslon
CGE models at typically parameterized by the use of a calibration procedure Some parameters
are exogenously given However some crucial parameters are determined in such a way that the
model replicates the data for a given base year See MansurmiddotWhalleY 11984) for an eXlensive
discussion of 1~i$ issue
CalibratIon in a dynamic context is generaliy interpreted as requiring two properties First
replication of base year data is required Second the model is parameterized to simulate an
intertemporal balanced growth path when Ihe base policy is maintaineq This Is the approach
followed by Ballard (1983) BallardmiddotGoulder (1985) Goulder (1985) and SummersmiddotGoulde
(1986) It follows the practice of BallardmiddotFuliertonmiddotShovenmiddotWhalley (1985) with their
sequential cquilibrium dynamics
Other authors Auerbach-Kotlikof (1983 1984 1987) Boyenberg (1984 1985 1986)
and Pereira (198Gb 1987d) follow whal we cali a qualitalive calibration The slructural
palametars are exogenously chosen so that lh economy follows a reasonable path into Ihe fulure
ThaI has 10 do wllh the fact thaI given the recursive nature 01 the dynamic economy and aside tom
such structural parameters only iniUa stock values are needed to run these models Given initial
conditions on the stocks of say private wealth capltal and government debt agents will optimize
and Ihereby generate a lirst round of net demands and equilibrium conditions In lurn the
equilibrium prices will determine the evolution of the stock variables into the next perIod
There are several potential problems with calibration in the conlext of dynamic models The
assumption of a steadymiddotstate growth ~ath in the base case can be questioned First while
27 I Ibullbull
steadymiddotstaU is a possibity it certainly is not the only meaningfull solution to dynamic models
Even in the case of a perfect foresight equilibrium the model implies an equilibrium path which i may or may not involve colanced growth The model not the modeller should dictate the nature 01 I themiddotbase case path Second in the context of a temporarY equilibrium path a steady state solutlcn is Inol a likely model QutcOn1tl In fact unlessmiddot expectations are stattc short run behavior consistent
with a steady-state evolution will in general not be generated On the other hand if static I expectations are self~fufjili[1g we have in fact a perfect foresight model Third even the base year
replication requirement may cause problems In Ihe cOntexl of temporary equilibrium Any
calibration parameter would be conditional on expectation rules which is probably an undesirable
feature
The nalur of the two-requirement calibration strategy is very muchin the spirit of the bull
traditional design of the comparision of alternative equilibria comparis_ion between a steady-state
base case on one hand and alternative paths including a transition period and a finat sleadyyenstate on
the other hand Pereira (198Gb 1987d) compare different (not necessarily steadystate)
equilibrium paths The arguments against such a procedure are based on the idea that the impact of
policy changes can be observed most easily since all departures from the steadymiddotstate can be
attributed to the altemalive polioy On the other hand the base case so dEfined as a steadymiddotstate is
consist~nl with previous work in a less dynamtc setting and Iherefore allows a common standard
for comparing model results
42 Computation algorithms
We are still al a stage in which basically each author uses a different computational technique
Th development of dynamic models - in parlicular wilh adjustment costs andior perfect foresight
bull
28
has corresponded with the decline in the use of fixed point algorithms In -fact given the relative
arge dimensions inevitably involved such algorithms lend to be very inefficient at the best and
chen prohibitively stow See Stone (1985) and Preckel (1965) for a comparative assessment of
different computation techniques Among the models surveyed only Ball1rdFulle(ton~
-ShavenmiddotWhalley (1985) and Feltensteln 1985 use Merrills variant of the fixed point
algorithm 1echnique
AuerbachmiddotKoHikoff (1963 1984 1967) follow a three stage procedure They first compute a
base case steady-state then a revised case steady state and finally transition path for the economy
berveen these two steady-states tn all stages a Gauss-Seidel iterative procedure is used
Ballard (1982) Ballard-Goulder (1965) Goulder (HiSS) and Summers-Goulder (1986)
~
use a method developed by Ballard and Goutder which is similar in many awects to the FairmiddotTaylQr bull
(1983) algorl1hm Short-run equilibria are calculated (using Merrills algorithm) parametric on
nrice expeC~lions The model is then iterated to generate self-fulfilling intertemporaf expectations
and the corresponding perfect foresight equilibrium In a relatively similar approach Andersson
1987 uses a simulation program SIMNON developed In the University of Lund This program
can handle two-paint boundary problems In a fashion consistent with the multiple shooting
algcrlthm (see Lipton-Poterba-Sachs-Eummers (1982))
Boyenbergs computational approach (1985 1985) differs from the other models surveyed in
that he relies heavily on analytical techniques Computations are done by using a dynamic version of
Johansons linearization method Being essentially determined by the continuous time nature of the
I I
f
model thiS linearization model has the disadvantage of confining the analysls to infinltesinal changes
around the base case equilibrium (see Bovenberg (1985) p 53) Pereira (1986b 1987d) uses an optimization algorithm NPSOL - developped by
rGillMurray-Saunders-Wrlght (1986)) This algorithm is used 10 compute the sequence of
I
29
Shortmiddotrun temporary equilibrium which makes up the intertemporal equilibrium path The
equilibrium conditions are seen as nonlinear equality constraints in the minimization of an artlcial
objective function The prices are normalized to the unit simplex by an additional linear equality I I 1 bull constraint
II ErlichmiddotGinsburgh-Heyden (1987) follow a unique approach in that they use a variant of the
optimization technique introduced by Negishl (1960) The economic equilibrium can be generated i
as a solution of a mathemallcal program Ihe objective function of which is a weighted sum of the
utility functions of the various agents while tho constraints set consists of the market dearing 1
conditions Ginsburgh-Heyden (1985) have extended Negishis resulllO tho case of downward price
lrigidities I
Finally the paper by Jorgonson-Yun (1984) is also unique in tIlat it is the only I
econometrically estimated model Different blocks for the consumption and production Side of the
-
model are separately estimated to provide the necessary structural parameters i
I The diversity of computation techniques is yet another indicator of the exploratory nature of the
body of literature surveyed in this article
43 Equal yield comparisions
The link between a base case and counteriactual simulations Is usually provided by the concept of
equal yield Shaven-Whaley (1977) discuss the meaning 01 equal yield in a general equilibrium
context when government is not allowed to run defICits Equal yield is interpreted to moan constant
public utility Government base case utility is maintained in the counferfactual experiments With
balanced budgets the concepl of equal yield is unambiguous The new equUibrium prices and the
balanced budge condition will del ermine the minimal expendilure and taxes needed to maintain base
30
case publi~ utility Accordingly in general equal yield is inconsistent with equal nominal tax
revenue Some change in tax revenue Is necessary Different tax replacement schemes ate
considered to assure that enough tax revenue is collected This is the approach essentially followed
by most 01 the papers surveyed Only Feltenstein (1985) Goulder (198) and Jorgenson-Yun
(1984) chose not to follow an equal yield strategy
The question is of how to Interprete lila concept of equal yield when the government is alowed to
fUn deflcLts The optimal leve of expenditure for base ease public trliUty can now be taAinanced
bond financed or financed by a mix of bonds and taxation We have several versions of equal yield In
particular equal yield may now be consistent with equal tax revenUE Furthermore some meaSure
ltgtf financial crowding out effects of government deficns can be inferred from the comparislon of the
several equal yield alternatives These issues are extensively discussed ~Pereira (1987b)
44 Price expectations and the dynamic generalization of compensation
Indicators
The sum of equivalent and compensation variations over households is the most widely used
aggregated measure of efficiency gains or losses In the context of $teady~state eomparisions andor
when complete future markets exist andlor perfect foresighl is assumed there are no difficulties
associated with the use of the standard Hicksian indicators In fact correct future prices are known
in these cases
If expectations are not selfmiddot fulfilling andlor future markels are not available a dynamic
generalization is necessary 8allardmiddotGoulder (1985) provide some steps in that direction by
defining an indicator that accomodates periods far into the future when households do not have
perfect foresight but a steady~state prevails On he other hand this issue is more fundamental in
i
[
I
i I I I I [
I I
bullI
31
the contex~ of a Gner1 lemporary equilibrium ftamawor~ In such Circumstances Pereira
(1986a) dev~lops a cnamlc generalization which is obtained as the present discounted value of a i i
sequence of short-fUn optimal expenditure functions consistent with a base case expected future I stream of IlWities I
32
5 Empirical evidenee from selected policy Issues
Dynamic tax models have generated several important results which escaped Sialic modellers
The following is a seleCpoundd set of issuos whIch stress the marginal benefits of dynam~ modelHng of
economic behavior in innovltlttive areas of analysis The discussion of a consumption tax emphasizes
the benefits of dynamic tusehold behavior and intergenerational aspects The study of the
flliminatlon or the re~intToduction of the investsnemt tax credits is made meaningful by the dynamic
modelling of production betlavor Modelling governmenl in a dynamic conlexts allows the modeller
to study the irrrpact of government deflcits Finally a dynamic framework lets us appreCiate the I
relevance of consumer expectations in terms of the evaluation of policy alternatives
i
i
51 Consumption Tax
t
By Its very nature a consumption tax can not be adequately investigated with a static model bull
Fullerion-Shoven-Whaicy (1983) use Ihe model of the US economy s described in
Ballard-Fullerton-Shoven-Whalley (1935) to evaluate the movement from the currenl US tax Isys1em to a progressive consumption tax Since their modeJ Incorporates a Jaborleisure choice
where leisure is an untaxed commodlty their results reflect the fact that both the consumption tax
and Ihe present system are dslortionary
Concentrating on the rntertemporaf distortions they show that sheltering more saving from lhe
current US income tax could improve econornic efficiency even if marginal tax rate increases are
necessary in order to maintain government fevenue~ At first you have a revenue shortfall
However the economy moves to a higher sustatned growth path and uitimatefy lower tax rates
cangenerate the same revenue path Also wages increase in the long run with this policy The
bull
bullbull ~
33 shyt
i
I
I
present value of welfare gi[lS for a potiey of complete savin9 deduction wilh marginal tate
adjustments (consumption lax) is simulated to be around $500 billion to $600 billion 01 1973
dollars
They also investiga1e th 1ngth of ttme it takes th~ economy to adjust to these policy changes
Roughly they estimate the long run to b thirty years although this figure Is very sensitive to the
specification of the savings elasticity - a crucial parameler in this model
52 Investment T~x Credit
Goulder-$ummers (1 S87) address the impact of eliminating the lnvestmenl Tax Cradit (ITC) on
intersectoral capital formation and on economic growth Their model isYa1ticularly adequate to
addies$ this issu~ in that is postulates a forward looking investment m9we with adjustment costs
They show that the eliminating the ITC causes a reduction in the rale of investment Inveslment is
estimated to fall by about 7 in the short-run and by about 12 in Ihe new long run steady state
In lurn a previous policy anncuncement lowers Ihe overall al1ractiveness of Investment and leads 10
a downward shift of the Investrrent profile On the other hand the combined effect of a revenue
neutral simultaneous efimiraron of the ITC and reduction of the corporate tax rates is a long run
reduction ef the capital Sieck by 35 This pattern suggests that a revenue neutral increament in
both Ihe corporale tax rates and the inveSlment lax credits would be preferable 10 the formula
adopted in Ihe US Tax Reform Act of 1986
Pereira (1987d) addresses the impact of the ITe from Ihe oland point of Ihe curTenllax syslem
under the Tax Reform Act of 1986 Given Ihe concern over the potential depressiVe effects upon
savings and investment of the new tax system in conjunction with deficit reduC1ion mechanisms of
the GrammmiddotRudman type a pertinent question is should Ih lTC be e-introduced In the context
bull
34
-of his mod~1 of the US economy Pereira focuses on the trade-off between the distortion in the
intersectoral allocation of capital induced by the lTC the lower tax revenues it generates and
potential financial crowding-out effects on one hand and the positive effects of lowering the relative
price of new capital goods on the other hand Preliminary results confirm the qualitative results in
Goulder-Summers (1987) suggestting a welfare gain of about 2 of the present value of GNP in
the best scenario of absence of replacement taxes Furthermore preliminary results show an
important time pattern to welfare gains with the average benefits increasing the further you look
into the future This pattern is due to the presence of constraints to intersectoral mobility of capital
and inertia in the adjustment towards the optimal capital levels as reflected by the presence of
adjustment costs
I
I
bullAuerbach-Katlikaff (1gB) in the context of their intergenerational model of the US economy
consider the impact on savings capital formation and interest rates of deficit policies and balanced
budget increases in government consumption They conclude that deficit finance and government
consumption can significa-ntly crowd out capital formation and lower the welfare of future
generations However crowding out from deficit finance is a very slow process because it results
from increased government spending over potentially long horizons Also deficit policies may
substantially influence the longmiddotrun interest rates while leaving the short-run rates essentially
unaffected Finally and in opposition to the central role adjustment costs seem to play in both
53 Financial crowding outeff~cts of government deficits
Goulder-Summers and Pereiras models Auerbach-Kotlikoff suggest that the time path of interest
rates induced by a policy of deficit financing seems to be insensitive to the presence of adjustment
costs
bull
35
54 The role 01 consumer expectatlons
The expectations analysis has uncovered one of the benefits of dynamic modelling
Ballard-Goulder (1985) find that the appeal of adopting a consumption tax depends on the level of
foresight possessed by the consumers Furthermore additional foresight may be welfare worsening
This is a second best type of result This tends 10 occur under policies that lead to capital deepening
and declining rate of return to capital over time To the extent that consumers have more foresight
they will be beller equippod to anticipate the fall in the rental price of capital Consequently if a
saving incentive is enacted people save more with myopia than with perfect foresight Given the
exi~tence of taxes on capita and the discrepancy between the private an~--sccial returns to capital
the greater savings with myopia is beller socially Ballard-Goulder find that the welfare gains
from a consumption tax are reduced by about 10 when we move from myopia to farleet foresight
Therefore the attractivenness of adopting a consumption tax seems fairly [obust across these
foresight specifications
I
bull
36
6 Tbe power and weaknesses 01 dynamic modelling the example 01 corporale
tax Integrallon
The Issue of corporate tax integration allows us to sh~w the stre-ngths and weaknesses or the
dynamic approach Empirical evidence using CGE techniques indicates that depending on the precise
scheme of integration and the 1ax replacement methods integratiOn may have substantial effects In
Ihe work of FuliertonmiddotKingmiddotShoven-Whalley (1980 1981) total integration was found to yield an bull
annual static eHielency gain of $4 to $8 billion in 1973 dollars Simulated dynamic gains may be as
large as $695 billion or abcu 14 of the present value of tulUre consumption and leisure in the
US economy These gains result primarily from interindustry reallocalions of Investment and an
improved inter~emporal allocation of consumption
Mora recent work emphasizes hOw consumers asset portfolio decisions and firms finanCial
decisions affect the efficiency gains from integration Slemrod (1980) focusing on consumers
asset ponfoio deCisions finds static efficiency gains which are about twice -as Iarge as those
repOrted iiy Fulierton-King-ShovenWhalley (1961) FuUrlon-Gurdon (1983) focus on the
firms financial decisions They roporl efficiency gains of 6 of GNP from the elimination of the
tax distor1ions favoring debt However when they eliminate the corporate tax and repiace it with
increaSed persona income taxes addlHonal dls10t1ions are created in the optimal labot~leisure
decisions These distortions tend to dominate the analysis slJ(h that the overall effects of complete
integration are very modest Galpr-LuckmiddotTodr (1986) address simultaneously consumers
asset pOrlfollo decisions and the firms financial decisions They also find very modest efficiency
gains from integralion
The above results are important but may be severely biased There are severa aspects of
economic behavior and modelling crucial for the study of Income tax Integration whiCh have no been
i 37
captured innny of the above papers One aspect is the absence of government deficitsmiddot a balanced
budget is assumed It is true Ihal resource crowding OUI is caplured In these mOdels but financial
crowding out induced by government spending is nOI Second investment is not derived from
optimtztlcn behavior Investment behavior passively accomodates endogenous saving decisions As
a consequence the differential impacts of policies in the incentives to save and to invest are not
captured Aso full capital mobility across secors is assumed with instantaneous capital
adjustmenlS towards optimal levels This assumption rules qut different costs of capital across
seclors and therefore differentiated reactions to tax policies changes Finallyhe modelling of both
gov9rnmerlt deficits and of endogenous real and financial investment decisions necessitates the
~nsideraiion of a dynamilt framework and the introduction of financiaJ assets govemment bonds and
iprivate financial assets A dynamic framework also highlights the efficiency effects of Integration on r
ithe ptimal intertemporal decisions The above models are either sIalic (Slemrods and
GalpermiddotltJckemiddotTode(sj or a dynamic sequ~nce of otherwise Slatic model bullbull I Pereir a (1986b) develops a dynamic applied general equilibrium model bull to study the tbull
efficiency effects of integration on the growth and allocation of investmenl across sectors Special
Iattention is paid to the structural affeelS of resource and financial crowding oul induced by
gove mect spending and deficits and to the real and financial Investment reactions across industries
to both tax integration and flnancial crowding out His mode departs from previous work on income
lax Integraticn and for that matler from most 01 the CGIE literalure for tax policy evaluation in
several direcllons It encompasses an endogenous sequential equilibrium struelure founded on
dynamic behavior with flexible expectations Government deficits are optlmaUy determined
Investment decisions are tward looking and the resull of optimizing behavior Several financial
assetsmiddot public and privatemiddot are considered
Simulation resulls Indicale that the welfare gains from Integration under large deficits are at
32
best lJlodest when measured in terms of the GNP The elimination of the corporate tax and its I replacemenl by increased income lax rates yields long run benefits which are never larger than
2 of the presenl yalue ct telr consumption and leisure onder the previous tax regime and 1
under the current tax regime However the short run w~lare effects of full integration tend 10 be
very low and in in some scenarios even negative This is a flew intertemporar pattern of efficiency Ieffects which reflects an adjJsiment lag in the interindustry investment decisions due to the
existence of costs of adjustment On the of her hand partial integration achieved by excluding I dividends from the corporale lax base syslematically yields negalive efficlencyeffeclS In Ihe long Inm these can be as high as 3 of present value of the future value of consumption and leisure This
is a new second best effect suggesllng Ihat less than complete integration may have perverse I efficiency effects
The dynamic slructure of PerclrBs model snowing for an improved freatment of real investment
declsions and government deficts provides a potential setting for a more accurate measure of the
costs and benefits of inlegration The simulation resullS suggest lower benefits and an intertempora Ibull
pattern of growing benefits
Simulation results also clearly suggest robust negative effects from partial integration How
reliable is this result If dJvidends were deductible from the corporate lax base (as are interest
payments) the preferenlial Ireatmant of debl over equily would be ellminaled Corporallons should
be expecled 10 react by decreasing the optimal deWeqully rallo However corporale financial
gecisions are exogenously set In Pereiras model as in aft of the other models surveyed that go as far
as dealing with the issue
Also withdrawing dividends from Ihe corporate tax base is a way of eliminating Ihe double
taxation of dividends at belh Ihe corporate and personnal Income levels How will the optimal
aliocellon of household saving accomodate that change in Ihe relalive return to corporale equity
39 I
t However hOusehold portfolio decisions are exogenouly set in Pereiras model as in all of the other
models surveyeltj go as far as dealing with the issue
IIt is sate to say that the evaluation 01 the benefits 01 partial Integration as defined above are
sevrey underestima1eO Meanwhile the distortionSoenerated by the increase in the marginal
personal tax rates designed to ralse the revenue foregone by the dividend exclusion are fully
aCCltlunted for A good measure of the costs of integration together With a poor evaluation of the
benefils may very well be the reason why partial integration is Simulated 10 yield neg alive
eHikiency gains
On a different vein as most of Ihe model surveyed here Pereiras is a closed economy model It
is legitimate to wonder how the resullS would change If international capital flows were anowed
IThis almost certainly would affect financial crowding out since a good d~al of the capital inflows
could be used to finance public debt
For these various reasons the results should be treated as preliminary but they strongly I suggosl that the dynamic structure provides valuable new insights into the corporate tax integration SSU2
40
7 Concluding remarks I I I i
What has been acccmplished Ihis far The success of Ihe CGE research in laelding the challenge
of dynamic modelling can not be denied
I iGreat progress in the modelling of household behavior has been achieved Promising
developm~nts in the modeUing of production and government behavior have also been accomplished
Interesting innovations in the concept of equilibrium and computation techniques were discussed
The policy analysis was greatly enriched by considering transitional effects together with longer
run steadymiddotstate equilibrium paths generalized equal yield strategies accomodating different
financial crowding out impacts and generalized dynamic policy evaluation indicators
An important set of issues have been addresed by the models surveyed Traditional issues
focusing on the effects of capital taxation and consumption taxes have been pursued In terms of
capital taxation for example the intertemporal nature of the issue was enhanced by allowing an
optimal evolution of capital stock in the economy and forward looking optimal investment decisions
In Andersson (1987) GouldermiddotSummers (1987) and Pereira (1986b 1987d) this is coupled
with a improved treatment of several tax provisions like investment tax credits and depreciation
allowances In terms of the consumption taxes developments in the specification of Intertemporal
household behavior the introduction of overlapping generations and the modelling of
intergenerational links via bequest motives as well as a closer attention to demographic evolution
provided a much improved economic setting for the understanding of the several aspects of the
problem
Furthermore dynamic modelling has permitted the addressing of a set of new issues Policy
Issues ranging from dynamic tax policy analysis to the evaluation of exogenous changes in
government expenditures to the impact of different methods of financing government expenditure to
4 1
the importance of tinanciai croHrii1g out to 1he relevance of consumorS expectations in tne policy
results have been addressed
Despile all of the progress and in part due to such progress several avenues are wide open for
much neded additional research The following seem to be the most Interesting and promising
areasmiddot First the specincation of liquidity constraints to household behavior may help to obtain a
better understanding of policy chenges that affect directly the interest rates in the economy Second
major attention needs to be paid to the incorporafion of a roraign sector (with open capital markets)
Third some efforts are needed in terms of finding adequate computation techniques now that
dynamic modelling eally makes the curse of dimensionality worse than eVOrbull
The modelling of financial markets is the single most unsatisfactory speet of the dynamic
models surveyed here~ The problems associated with provlding a mo~ adequate treatment of
financial markets and in general endogenous and optimal financial deci$ions bull both at the corporate
and at the personal levelsmiddot have 10 be addressed That necessarily requires Introducing uncertainty
into the models To b adequate this must go well beyond the mere assumption of point expeetalions
with the whole array of modelling implementalion and evaluation problems thai generates
bull I i
42
REFERENCES
Abel A and O Blanchard 1983 An Intertemporal Model of Saving and Interest Econometrica
51 pp 675-92
~ Andersson K 1987 Tar-ation of Capital in Sweden - A General Equilibrium Model Ph D
Dissenation UnIversity of Lund Sweder
Auerbach A L KotlikoH 1983 National Savings Economic Welfare and the Structure of
Taxation in M Feldstein ed Behavioral Simulation Methods in Tax Policy- Analysis Chicago
University of Chicago Press
Auerbach A l Kotlikoff 1984 Social Security and the Economics of Demographic-
Transition in H Aaron and G Burtless eds Retirement and Economic Behavior Washington DC
Brookings Institution
Auerbach A L Kotlikoff 1987 Dynamic Fiscal Policy Cambridge Cambridge University
Press
Auerbach A L Kotlikof and J Skinner 1983 The Efficiency Gains from Dynamic Tax
Reform Internatiooaj Economic Reyiew 24 pp 81middot100
Ballard C 1983 Evaluation of the Consumption Tax with Dynamic General Equilibrium
Models Ph D Dissertation Stanford University
Ballard C 1987 Tax Policy and Consumer Foresight A General Equilibrium Simulation
Study Economic Inquiry 25 pp 267-284
Ballard C and L Goulder 1985 Consumption Taxes ForeSight and Welfare a Computatonal
General Equilibrium Analysis in J Piggott and J Whalley eds New Deyelooments in Apolied
General Equilibrium Analysis Cambridge Cambridge University Press
Ballard C D Fullerton J Shoven and J Whalley 1985 A General Equilibrium Model for
t 43
Tax Eoliey Eyalua~iQO NBER University or Chicago Press
Borges A (1987) bull A Survey of Compula1ional General Equilibnum Models CECD Economic
Slud~ 8
Bovenbig L 19B3 imperfect Capital Mobility in a Perfect Foresight Herberger Model
Mimea UC Br~eley
Bovenberg L 1984 Capitel Accumulation and Capital Immobility Qmiddottheory In a Dynamic General Equilibrium Framework PhD dissertation University of California at Berkeley t
Bovenberg L 19850 Dynamic General Equilibrium Tax Models with Adjustment Costs in A iManne ed bull 1985 EtQnomrc ECUjilbrium Model Formulation and Solution Mathematical
lProgramming Siudy 23 Arnsterdam NorthmiddotHoliand
Bovenberg L t 1986 Capital Income Taxation in Growing Open Eeondmies Journa of Pubtic
Eamp(omlcs 3t pp 347middot376 IBovenberg L ald W Keller 1981 Dynamics and Nonlinearlties in Applied General
Equilibrium Models Internal Report Mlmiddot81middot208 Department for Statistical Methods CBS bull
Vooiburg The Netherlands
Charnley C 1981 The Welfare Cost of C~pltal income Taxation In a Growing Economy
IJournal of Political ECQnQm~ 89 3 pp 468middot496 I
Charnley C 19810 Efficient Tax Reform in a Dynamic Model of General Equilibrium I Mimeo Yale University I Charnley C 1982 On the Optimal Taxation In Stylized Models 01 Inlartemporal General
Equilibrium Mimeo Yale University I Decaluw B and A Manens 1985 Pays en Oeveloppement at Modeles Calculables OEquilibre
General - Une Revue de al litterature empirique Monographie Sene 9 Volume 3 Centre de
Recherche et Developpemont Economlque Unrslt de Montreal
44
bull
I I
I
I
1
Erlich $ V Ginsburgh and L van der Heyden 1987 Wher~ do Real Wage Policies Lead
Belgium bull A General Equilibrium Analysis European Economic Reyiew (forthcoming)
air R and J Taylor 1983 Solution and Maximum Likelihood Estimation of Dynamic
Nonlinear Ratioant Expeclations Models Ecooometrica 51 pp 1169middot1186
Feltenstein A 1984 Money and Bonds In a Disaggregated Ec~nonYin Searl 1-1 J Shoven
Eds Aoolied ~eoeral EqJiliDCiulD tioalYsis Cambridge University Press Cambridge
Feltenstein A 19B6 A Dynamic General Equllibrlu Analysis of Financial Crowding Out
Theory with an Application to Australia Journal of Public Economics 31 No1
Fullerlon D R Gorden 1983 A Reexamination of Tax Distorllons in General Equilibrium
Models in M Feldstein Ed Behayora Simulation Methods in Tax pOlicy Analysis Chicago
University of Chipgo Press
Fvllerlon D J Shaven and J Whalley 1983 Replacing the US Income Tax with a
Progressive Consumption Tax JoulOal of Public ECQnomics 20 pp3middot23
Fullerton D Y Henderson J Shaven 1984 A COmparision of Methodologies in Empirical
General Equilibrium Models of Taxation in Searl H J Shoven Eds ~pQlied Geoaral Jguilibrium
lIoalyss Cambridge University Press Cambridge
Fullerton 0 A King J Shaven and J Whaney 1981 Corporate Tax Integration in the United
Statesmiddot a General Equiiibdufi1 Approach American Economic Review 71 pp~ 677~691
Galper H R Lucke E Tader 1985 Taxation portlor Choice and the Allocation of Capital A
General Equilibrium Approach Mimeo
Gill p W Murray M Saunders M Wright 1986 Users Guide for NPSOL (Version 40) A
Fortran Package for Nonlinear Programming T echnicsl Report SOL 86middot2 Deparlment of
Operations Research Stanford University
Glnsbur~h V and L van der Heyden 1985 General Equilibrium with Wage Rigidities An
45
AppJication to Belgium in A Manne ed t Economic EQumprit1m~ Made Formulation and $oJioo
Mathematical Programming Study 23 Amsterdam NorthmiddotHolland ~
Goulet J 1968 1 Adjustrent CO$1S in the Theory o(the Firm Reyiew of EconomiC Studies 35 bull
pp47middot55
Goulder L 1985 Tax-Financed versus BondmiddotFinancedGhanges in Governent Spending
bull Mimeo Harva(d UniversilYmiddot
Goulder l J Shaven and J Whalley 1983 Domestic Tax Policy and the Foreign Sector The
importance of altemative foreign policy formulations to results from a general equilibrium tax
analysis model In ~n methods in Ibe antico of income from caoia1 ed Martin
Feldstein Chicago University of Chicago riSS
Goulder l L Summers 1987 Tax Policy Asset Prices and 3rowtly ~
A General Equilibrium Analysis NBER Working Pper No 212B
Grandmont J 1982 Temporary General Equilibrium Theory Ch 20 in Handbook 01
Mathemalical Eccnornics Vol II NorthmiddotHoliand
Harberger A 1959 The Corporate Income Tax An empirical Appraisal In lax Revision
Comoendjum Voll House Cornmitee on Ways and Means Washington DC Govemment Printing
Office
Harberger A 1962 The Incidence of the Corporate Income Tax JOULn1 of Political
Economy 70 pp 215middot240
Harberger A 1964 Taxation Resource Aliocalrlon and Welfare fn lb Role of Dir bullbulll and
Indjresl Taxes io tbe Fede[al Reserve s~I~Ill Princeton Princeton Unlveresity Press
James J 1985 New Developments in the Applications of General Equilibrium models to
Economic History in Piggott aand Whalley Eds
Jorgenson D and K Yun 1984 Tax Policy and Capital Allocation Harvard Inslitute of
bull
I
I f
Economic Research WP 1107
Judd K 1965 ShonmiddotRun Analysis of Fiscal Policy in a Simple Perfect Foresight Model
Journal of poJiHcal 5CCJiW1l pp 298middot319 r
Upton D J Potbc J Sachsa nd L Summers 1982 Multiple Shooting in Rational
Expectations Models fuoometrlc 50 1329middot1333
Manne A ed 1985 fconQrnjc Eoullibrium Model FOrIDujaljQo and Solution Mathematical
Pogrammlng Study 23 Amsterdam NorthmiddotHolland
Manne A 1985 Or the Formulation and Solution 01 Economic Equilibrium Models In A
Manne ed ECQOQlli-Jr~ibmiddotum ~tQdel EQlJulailon and Solution Malhematical Programming
Study 23 Amsterdam NorthmiddotHolland
Mansur A and J Whalley 1984 Numerical Specification of Applied General Equilibrium
Models in Scarf H J Shoven Eds Aoplied General Eouillbrium AnaiysectIsect Cambridge Universily
Press Cambridge
Merrill 0 1972 Applications and Extensions of an Algorithm thaI Computes FixedmiddotPoints of
Cenain Upper Semimiddotcominuou Point to Set Mappings PhD Disserlation Unlvel$ity 01 Michigan
Negishi T 1960 Welfar Economics and Exislence I an Equnlbrlum for a Competitive
Economy MelroeCOOQ11ka 12 92middot97
Pereira A 1985 On the Existence and Uniqueness of Optimal Ou1put Path lor CRTS Firms in
Adjustment Costs Technologies Mimeo Stanford University
Pereira A 1985a Consumer Expectations and Ihe COSI 01 UllUty In an Intertemporat
Framework Mlmeo Stanford University
Pereira A 1985b Corporale Tax Integrallon and Government Deficits A Dynamic General
Equilibrium Anaiysls Mlmeo Stanford University
Pereir A 1987 On Ihe Computation of the Users Costs of Stocks Ecooomi (forthcoming)
47
Pereira A bull 1987b Equal Yield Alternatives and Government Oeficlts Mimeof Stanford
University
Pereira A 1987c A Dynamic Applied General ~uHlbrium Model for Tax Policy Evaluation
Comlletamp Docurnent2tion Mimeo University of California San Diego (in progress)
Pereira A 1987d Should the investment Tax Credit be Rei(ltroduced Mlmeo University
of California San Diego (ir progreSs)
Piggott J and J Whalley eds 1985 tJew QeyeQpments in Applied General EQuilibrium
bullenalysjs Cambridge Cambridge University Press
Preckel Pbull 1985 Alternative algorithms for computing economic equilibria In A Manne Ed
Radnor R 19amp3 Equllibrium under Uncertainty In K Arrow M Intrilligaor ods
I ibull
Robinson S 1 g86 Mullisectorai Models of Oeveloping CountrieS a Survey Oivision of
Agricullure and Resources UC Borkeley WP 401
Scarf H 1967 The approximation of fixed points of a continuous mapping ~MM Jouwal of
epplid Mathornallcs 15 pp 1328middot43
Scarf H with T Hansen 1973 00 the Computation of EconomiC Eguilibria New Haven Yale
University Press
Scarf H J Shoven Eds 1984 61l1llied Genera Elujlibrium 6nalyssect Cambridge University
~ Press Cambridge
Shoven J 1976 The Incidence and Efficiency Effects of Taxe on Income from Capital
Jurnal of ramie1 Economy 84 pp 1261-83
Shoven J 1983 Applied General Equilibrium Tax Modemng IMF Slaff PaoerS 30
Shoven J and LB Simon 1986 Share Repurchases and Acquisitions An Analysis of which
Firms Participate NBER Working Paper No 2243
bull
48
IShaven J and J Whalley 1972 ft General Equmbrium Calculation of lhe Effects of
Differential Taxation of Incoma from Capital in the US Jauroe of public Econgmics 1
pp281middot321 (
Shoven J bull and J Whalley 1973 General Equilibrium with Tax$s a computation procedure
and an existence pr00f ReviEhi of Economic Studies 40 pp 475middot490
Shoven J J Whalley 1977 Equal Yield Tax Alternatives A General Equilibrium
Computa~ui0nal Tfchnique J~HJrnal of PUblic EconomIcs 8 pp 211~224 bull
Shaven J J Whalley 1984 Applied GeneralmiddotEquilibrlum MOdels of Taxalion and
International Tr8de An Introduction and Survey Journal of Economic literature 23t pp
1007middot1051
Slemrod J 1980 A General Equilibrium Model of Capital Income Taxation Ph D bullDissertation Harvard Univnrsity
Slemrod J 1983 A General Equilibrium Model of Taxation with Endogenous Financial
Behavior in M Feldstein edbull Behavioral Simulation Methods In Tax Policy Analysis Chicago
UniVerSity of Chicago Press
Stone J 1985 Sequential Optimization and Complementarily Techniques for computing
Economic equilibrium in A Manne Ed
Summers L 1981 Taxation and Corporate Investment a q-Theory Approach BroQkinas
Summers L 1985 Taxation and the Size and ComposHion 01 the Capital Stock An Asset Price
Approach Mlmeo Harvard University
Whalley J 1975 A Geneal Equiloibrium Assessment of the 1973 United Kingdom tax
reform Economica 42 pp 139middot61
1JLTIllOS JIQll~)ltG PAPERS PUBLICADOS bull
I
I I
j
112 70 - LUCENA iJiCIIJo ~ IIEnvironrilent Honitoring and Orqanization Structure T (Setellibro 1987)
nI 71 - L~ENA Diogo A Note on the Representation of Information Stxueture (SeteJObro 1987)
nQ 72 - LUCENA D1CgO Envuorment 111)nitoring and Organization
Structure II (Setelilhro 1987)
ril 73 - AlmmE3 Antonio Pais e CAsPAR Yitor bull Tributacao Inoentivos e InvesUlllentos Analise Qualitativa noYellibro 19S7)
n2 74 - liAlltOSA rose Pedro Opt1lllal lIage R1gidity A SUggestedtlthodology to Teet tb Theory and an Application (Outubro 1987)
nI 75 - CABRAL L 11 B Three lIotes on SYJIIlletric Gemes witt itsYJIIlletric Equilibri (NoveJObro 1987)
n2 76 - LUCENlt D10gOTo Search or Not to Search (Fevereiro1967)
nQ 17 - SA Jorge asconcelos euroI A Theory of SynergyB (Fevereiro1987)
n2 76 - VILARES tanuei Jose Os Be IntenaMios IlIIportados ColIIO Factor de Produao (Julho 1987)
nQ 79 - SA J~rge VasCQncel03 e HoY to Copete and Co~unicate in tature Industrial Products_ (aio 1988)
n2 80 - ROIl Rafael Learning and Capacity Expansion in a New larket Under Uncertainty (Fevereiro 1988)
n2 81 PEREIRA Alfredo llarvilo Survey of Dynamic Computational General EquilibriUlll Hodel for Tax Policy Evaluation (Outubro 1987)
Qualquer in ormacao sabre o 1IIorlltlng Papers ja publlcados sera prestada pelo Secretariado de Apoio aos Docentes podendo os ~e3OSf
ser adquirido5 ne Secyo de Vlndas da Faculdade de EconollLia UNL na Travessa Esteviio Pinto Cltpolide - 1000 LISllOA
1 6
Some lodels now have a sophisticated dynamic specification of hoosehokl In Ballard (1983) -
BallardmiddotGoulder (1985) and Auerbach-Kotlikof (1983 1984 1987) consumers behavio is
embedded in an iterQEneraional setting Each of them has S5 age cohorts simulataneously alive
Som-e measure of jntergener1ional altruism is considered in the form of bequest motives In Ballard
(1983) and Ballard-Gouldr (1985) consumers derive utility directly from bequthing parl of
their wealth Therefore bequests assume 1he form of a scrap value of terminal wealth of a
generation In AuerbachmiddotKotlikof (1983 1984 1987) each generation empathyz bullbull with Ihe
utility of future generations ovar bequethd wealth In addition Ballard (1ge3) includes detailed
demographic poiections wfill the intent of incorporating in the policy analysis the effects of the
bull Post World War tI baby-boom
In the real world household decisions also include the optimat allo~atiOn of saving among
shyalternative physical or financial assets Theories of household portfolio behavior are relatively
Iwen~establshed However they involve uncertainty as a crucial element Accordingly they have
middotnot been incorpora~ed in the context of the determinIstic dynamic models we sutyey here We will 1 Come backt-O this Issue below
32 Producer behavior
The efforts to build dynamic features into the economic behavior of producers are more recellt
and less widely adopted The first attempts are due to Bovnberg (1984 1985) and Summers
(1965) Dynamic behavior has been more fully incorporated In the recent models of Andersson
(1987) Auerbach-Kotlikoff (19B7) GouldermiddotSummers (1987) and Pereira (1986b 1987d)
Part of the reason why production side dinamlcs has been more slowty adopted is the weak
supply of accepted theories referring to the dynamic behavior of the firms In the models referred
I 17
to above gynamlc producion and investment b~havior are middotinduced by the existence of capital
adjustment costs and linked to Tobins q theory Adjuslment costs are designed to capture both the
incomplete mobility of capital aoross industries and installation costs ~e the costs of adjusting
capital towards its cpHmai level)
Adjllstment costs can be conceived as internal to the firm and measured in terms of foregone
outpul along the lines of Lucas (1967) Alternatively adjustment costs can be viewed as actual
external cests incurred together with Ihe purchase costs along the Unes of Gould (1969) With the
exception of Pereira (198Gb 1937d) all of the CGE authors follow the former approach
The firms in the economy maximize their market value as the present discountad value of the
future stream of dividends In Andersson (1981) and Pereira (198Gb 1987d) firms are seen as
maximizing the present discounted value of net cash flow Maximization is constrained by Ihe
adjustmenl cosl lechnology and ~n equation of motion describing Ihe evolution of the capital stock
It should be noted that undereogenous divldendrelention rules the two problems are equivalent II
should also be nOled Ihat a salisfectory economic rational for th bullbullistoo of dMdonds with the
present tax code is missing in the profession (Shoven (1986))
The models with static formulation of producers behavior are characterized by passive
investment behavior Investment merely accomodates to saving in the eccnomy With a dynamic
formulation induced b adjustment costs real investmen decisions are forward looking Investmenl
Is endogenously and optimally determined by the firms A fundamental difference between the
shonmiddotrun in which capital stock is given and longmiddotrun in which Ihe level of capital is allowed to be
optimally determined is emphasized
This extra richness of produc1ion dynamics is not without costs A careful look at the summary
tables will clearly show an inverse relation between the adoptlon of dynamic features in production
and the level of disagreggation of the production side of the economy In fact with production
bull
j
1 6
dynamics $e dimension of the problems is immensly increased Let IS be more specific
In a static framework with constant returns to scale production technology the output level ismiddot
indeterminate The optimal aliocation of inputs can be obtained by cost minimization with any
feasible output level generatillg zero profits Such zero profit conditions are used to solve for the
output prices in terms of the factor prices and hencemiddot to reduce the dlmensionality of the problem
from the -number of commodities and factors to the number of factors Thus even if the model deals
with 30 production sectors the computation of an economic equilibrium can take place using only
the dimensionality of the primary inputs in the economy
Now under certain regularity conditions on the production and adjustment costs technologies
leading to enough concavity of the optimality objective the intertemporal output path for the firm is
endogenously optimally and uniquely determined even with constant relums to scale technologies
(see Pereira (1985 1987a) Accordin~y with adjustment costs in the model optimal profits
will in general be non-zero This result has crucial implications for the computatiQn of equilibrium
in the model~ with production dynamics With adjustment costs no reduction of dimensionality is
possible The curse of dimensionality returns as a binding constraint
The introduction of adjustment costs adds another significant complication to the model With
capital being less that perfectly mobile in the economy different rates of return on capital will
exist in different sectors This is a difficult problem to tackle conceptually in the absence of
uncertainty Also in the real world producer maximization choices include also its choice of
financial ratios (debVequity) and payout rates (dividendretained earnings) These financial
subjects are difficult and much Clf this behavior is still taken as exogenous by the modellers We
will come back to these issues below
middot i
19
bull 33 government behavior
middot
Two issues dominate the modelling of government behavIor First govemment behavior has
Ibeen typically seen in Ihe CGE liIerature for lax poliCY evaluation as constrained by yearly balanced
I budsets (see for example 8allard-Fullrton-S~ovn-Whafley 098Sraquo The analysis of
govemment defidlS and public debl in a CelE context requires a dynamic selling Second the level of
government expenditures is either exegenouty given as in Auerbach-KoUikofi (1984 1987)
6ovenbrg (1984 1985) Feltenstein (1984 1986) and Jorgenson-Yun (1984) or
endogenously (but no optimally) determined by the balanced budget conditions as in
6allard-Fullerton-Shoven-Whalley (1985) Andersson (1987) Erllch-Ginshurgh-Heyden
~
(1987) Gouder (19aS) and Gouder-Summers (1987) In the second case the composition of
bull public expenditures is often optimally determined However the leve~ of government expendi1ures
can only 00 endogenously and optimally determined if the government is seen as an optimizing agent
and is allowed to run deficits
The first attempts to deal wiU the government defiCits in CGE tax models are due to
Auerbach-Kollilltoff (1984) and Feensain (1984) In Auerbach-Kotlikof (1984) government
expenditures afe exogenous They grow at the rate of growth of population However given 1he tax
structure and tax revenues yearly deficits and surpluses are atTowed subject to an in1ertempond
constraint that the present value of future tax revenues equals the present value of future
expenditures
In Fellenstein (1984 1986) government expenditures are also exogenously given He also
allows the government to run deficits to finance expenditures in excess to 1ax revenues However
surpluses are returned to CQnsumars in the form cf transfers Accordingly government ts not
subject to any constraint regarding the future repayment of public debt
In GouJder (19851 the 9Vernmenl maximizes a static social welfare ~nctio subject to a-
bull
20 I i
I
I I I
I bull
balanced budget conSirtlL This follows the optimal alocatkm of government expenditures alpng
the lines of Baliard-FHeiton-Shoven-Whalley (1985) Ttle model Is generallzea to allow for
exogenous changes in HIe time path of government expenditures Two financing alternatives are
consjdeft~d and contr8s1ed EldditionaJ tax revenues and bond issuance
Pereira (198Gb 1137d) auempts to address both the incorporation of deficits and the
determination of govement expenditures The path of government expenditures and the path of
delicilSlsurpluses (ano therefore the path lor debt) are endogenously and optimally determined The
government is seen as mltJximizing n intertemporal social welfare function given the 13x structure
Optimization is subject to a sequence of recursive equations of motion reflecting the evolution of the
ptlbllc debt allowing fDr government budget imbalances It should be stressathat this specification
01 the government cOMtcalnt is equivalent to the specificaUon in Auerbaeh-Kotlikoff (1983 1987)
when no liquidity constraints affect -gove~nment behavior
The extra richness in the treatment of government behavior allows tfte examination of
government debt poices in a truly dynamic setting Also financial crowding out effects induced by
government deficits can be analyzed (see Auerbach-Kottlikoff (1987) Feltenstein (1986) and
Perer (198Gb 1987e))
34 Foreign sector
Most of the open economy models follow the assumptions of balanced trade with import and
export net demands characterized by constant elasticities along the line of Ballardmiddot
-Fullerton-Shaven-Whalley (1965) Such is the case of Salfard (1963) Ballard-Goulder
(19S5) and Goulder-Sllmmers (1987)_ In Feltenstin (1985) the rest of the world is treated as
21
an addjlion~ consumer group
Bovcoberg (1986) develops a model In which two economies lIr considered each following
intertempoml perlect foresight paths These economies meet in the international forum Their
trade relaroMhips are characterized by yearly balanced trade accounts
None of the new generation of dynamic cae tax models has yet IncorPorated the lnternational
capital flows as done in aoulder-Shoven-Whalley (1983) for the earlier Ballarrshy
-Fullertonmiddot Shaven-Whalley (1985) model This is unfortunate as there is an important
Intertmporal aspect to international lending The first al1empts along Ihese lines are due to
Andersson (1987) and Erlich-Glnsburgh-Heyden (1987) Andersson (1987) in his model of the
Swedish economy adopts a closed economy appraoch in which rates of return in the domesJic
economy are largely determined by the international capital markets Fidinterest rate induce
Intlnatlonal capital flows which determine and finance the International lrade imbalance In
lurn In ErlichGinsburghHeydenmiddot (1987) foreign trade Is generaled according to an
intertemporal Irade welfare lunction with constant import and export elasticities In the shortmiddotrun
they allow inlernational trade inbalances which generate capilal ftows to the domestic households In
the long run however trade balance is assumed
35 The need for financial markets
A dynamic economic Wucture not only provides the ideal environment 10 model many features of
economic behavior it also permits one to Incorporate the financial side of the economy If the
government is allowed to run deficits the question 01 deficll financing automatically follows If
investment is optimally dewmined and returns to capital are different aClOSS sectors problems of
investment financing arise If there are seeral final1cfal assets In the economymiddot government
4-
I i
ibull
I I I I
I
j
22
bonds privpte bonds and equity (or simply physical capital installed inmiddot different sectors) the
problem of allocation of saving among assets with potentially different returns arises
The papers surveyed var greatly with respect to l~e extent of their attention to the financial
side of the economy At Clne extreme are the models in ~anardmiddotFuliertonmiddotShoven-Whalley (1985)
Andersson (1987) Ballltld (1983) Ballard-Goulder (1985) Bovenberg (1985 1986) and
Erlich-Gjnsburgh-Heyden C19B which are devoted exclusively to the real side of the economy
In turn Auerbach-KotliKoff (1984 1987) Feltenstein (1984 1986) and Goulde (1985)
allow for government debt In these models saving finances changes in government debt and physical
capital Private and publlc 2ssets are perceived by the households as perfect substitutes The
allocation of saving merely adjusts to the relative demands for funds
Feltenstein (1984 1986) is the only model surveyed which introduces money Government deficits are financed by issuing money and bonds according to an exogenously given rule Money is
demanded by consumers for transaction motives and an exogenously given fraction as a store of value
On the other hand government bonds and physical capital are the vehicles for the intertemporal
transfer of wealth
Summers (1985) and GOlldermiddotSummers (1987) introduce a whole menu of financial assets shy
firm specific equity capital Different assets earn different rates of retum However such rates
are equal up to constant and exogenous sector speCific risk premia Therefore the introduction of
constant exogenous risk premia as helpful as it may be in the context of calibration does not solve
the main issue of the non-optimality of the allocation of saving Also talking about risk premia in a
deterministic context is somewhat unsatisfactory
Pereira (1986b 1987d) also introduces a whole menu of assets private and public bonds and
firm specific equity However all the assets are expected to yield the same rate of return and
therefore perc~jved as perfect substitutes The different asset types allow consideration of
23
exogenOU$~8tUequjty triO dividendlrc~enlion rufes and therefore seve-ral sources of investment
financing bond equity and retained earnings
The nonmiddotcptimalHy of the allocation of saving and the absence or exogenelty of corporate
financial rules is _a reftectlon of the limitations of the determInistic approach Either in a context of
perfect anticipation oi the future prices Of in general within the t$atm of point price
expectations all the papers follow a deterministic approach Under such circumstances consumers
either expect diHerent rales of return (inclusive of risk premium) across assets in which case
they wilt buy only one asset (that with highest rate) Of they expect equal rat~s of relurn in which
case Ihey are indifferent about the asset composition of Iheir portfoliO There is no way of
Iradingmiddotoff rales of return and risks to obtain an optimal interiof solution to the problem of the
allocation of saving
The most advanced contribution in the modelling of saving allocation in a CGE setting is due to
Slemrod (1980 1983) in the context of a static one-period model In his model consumers act
according to a two stage separable decision process They Hrsl decide on how much to save Then
they decide on the allocation of saving according to an Indirect uttlity function dependent on the rates
of ratum and variances offered by the different assets in the eoonomy The sourCe of riskiness in the
ampConory comes from an uncertain marginal product of capital On the other hand aside from
portfolio decisions the res of the economy is insulated from uncertainty
The most complete contribution in terms of the treatment of the corporate financial rules is
FulienonmiddotGordon (1983) In a variant 01 Ihe BailardmiddotFulienon-ShovenmiddotWhelley (1985) model
Fullerton and Gordon have capital intensity and optimal financial decisions Jointly deter~ined
through a two siage process The cosl of financing capital is minimized by trading off the tax
advantages of debt against the epected real bankrupcy coSls inherenl with high debVequity ratios
Given the optimal debtequity ratio the level of inveSlment Is chosen such that at Ihe margin the
24 l
bull
rerurn pn ~uity equas the return on bonds plus an exogenous risk premium
36 Market assumptions equilibrium and expectations
Virtually all of the papers arc characterized by Walrasian market clearing assumptions All
markets are perlect1y competi1ive Atomistic competition among agents 1s assumed even thoughmiddot only
a flnile number of agents is considered Virtually no mark~t disequilibria or price stickiness afe
considered The only exception is ErlichmiddotGinsburghmiddotHeyden (1987) bull In thei model of the Belgium
economy I the _wage rate is fixed in the short-run Therefore in the shOrlwrun desequilibrium in
the labor market will generate endogenous unemployment However in lh$ long-run all prices
Including the wage rata are Ilexible and accordingly ali markets ciear
Given the dynamic nature of behavlor in the economy markelmiddotcfearing prices in each period
depend on expectations of fture prices and on tax variables In the economy There are essentially
two ways at interpreting the economic equilibrium in such a dynamic conte~bull If future prices are
perfectly anticipated (io expectations are self-fulfilling) a perfect foresight equilibrium
prevaHs Then future actions are merely the implementation of current decisions for future
periods However jf ptica expecla110ns are not perfect (Le agents make mistakes w1th re$pect 10
future prices) then a temporary or shortrun equilibrium prevails Markets ciear and clearing
prices depend on fu1ure price expectations Current plans about the future are typically not
precisely implemented They will be revised as more or better information becomes available to the
economic agents See Grandmant (1982) for a comprehensive survey 01 the temporary equilibrium
literature
Wilh the exception of Gould (1985) and Pereira (19S6b 19S7d) all the models surveyed
adopt th~ concept of a perfect foresight equilibrium In turn BaliardmiddotGouldr (1965) considers a
J
2S
flexible am9unt of foresight In terms of the number of years over which price movements are ~
foreseen Pereiras model (1986b 1987b)) is flexible in a somewhat different way in that it can
indude any range of foresight from myopia to perfect foresight The choice between the perfct foresight and lemporary equilibrium is ultimately to be made on
philosophic grounds It can be argued that less than perfect expectations Imply that agents are
irralional in some way (see Auerbach-Kctiikoff (1987 p 10) However the reverse argument can
be made One can question wiether agents are reaHy rationar and perfectly knowledgable about
future prices
Recent evidene of Balird (1987) Ballard-Goulder (19851 Goulder (19851 and Pereira
(1986b 1987d) suggests that the choice in modemng expectations is an important one They show
that the degree of foresight inl0 thc future (ranging from perfect foresight to-myopic expectations)
may have dramatic impacts on the pollcy conclusions of Ihe model Accordingly the best research
strategy may b to design models which are flexible enough 10 allow for different rules regarding the
formation of expectations With the exception of the articles just mentioned sensitivity analysis
bullhave previously not been performed along this dimension
In terms of implementation the two concepts of equilibrium - perfect foresight and temporary
equilibrium have different implications The dimensionality of the equ1ibrium soll1ion algorIthm
is involved Suppose we have a model with ten markets to be run for a period of 50 years Aside
from normalization a perfect foresight model implies computing prices In 500 dimensions while a
-t~mporary equilibrium model requires sorving SO equilibria each in ten dimensions Given that
computational speed often varies with the cube of the number of dimensions the lemporary
equilibrium formulation is potentially strikingly more feasible However Ballard (1987) and
Goulder-Ballard (1985) have devoloped techniques to greatly speed the computation of a perfect
foresight equilibrium
26
j
I
4 Jrnplemenlalion and Issues on policy evaluallon
41 Calibration and equilibrium comparlslon
CGE models at typically parameterized by the use of a calibration procedure Some parameters
are exogenously given However some crucial parameters are determined in such a way that the
model replicates the data for a given base year See MansurmiddotWhalleY 11984) for an eXlensive
discussion of 1~i$ issue
CalibratIon in a dynamic context is generaliy interpreted as requiring two properties First
replication of base year data is required Second the model is parameterized to simulate an
intertemporal balanced growth path when Ihe base policy is maintaineq This Is the approach
followed by Ballard (1983) BallardmiddotGoulder (1985) Goulder (1985) and SummersmiddotGoulde
(1986) It follows the practice of BallardmiddotFuliertonmiddotShovenmiddotWhalley (1985) with their
sequential cquilibrium dynamics
Other authors Auerbach-Kotlikof (1983 1984 1987) Boyenberg (1984 1985 1986)
and Pereira (198Gb 1987d) follow whal we cali a qualitalive calibration The slructural
palametars are exogenously chosen so that lh economy follows a reasonable path into Ihe fulure
ThaI has 10 do wllh the fact thaI given the recursive nature 01 the dynamic economy and aside tom
such structural parameters only iniUa stock values are needed to run these models Given initial
conditions on the stocks of say private wealth capltal and government debt agents will optimize
and Ihereby generate a lirst round of net demands and equilibrium conditions In lurn the
equilibrium prices will determine the evolution of the stock variables into the next perIod
There are several potential problems with calibration in the conlext of dynamic models The
assumption of a steadymiddotstate growth ~ath in the base case can be questioned First while
27 I Ibullbull
steadymiddotstaU is a possibity it certainly is not the only meaningfull solution to dynamic models
Even in the case of a perfect foresight equilibrium the model implies an equilibrium path which i may or may not involve colanced growth The model not the modeller should dictate the nature 01 I themiddotbase case path Second in the context of a temporarY equilibrium path a steady state solutlcn is Inol a likely model QutcOn1tl In fact unlessmiddot expectations are stattc short run behavior consistent
with a steady-state evolution will in general not be generated On the other hand if static I expectations are self~fufjili[1g we have in fact a perfect foresight model Third even the base year
replication requirement may cause problems In Ihe cOntexl of temporary equilibrium Any
calibration parameter would be conditional on expectation rules which is probably an undesirable
feature
The nalur of the two-requirement calibration strategy is very muchin the spirit of the bull
traditional design of the comparision of alternative equilibria comparis_ion between a steady-state
base case on one hand and alternative paths including a transition period and a finat sleadyyenstate on
the other hand Pereira (198Gb 1987d) compare different (not necessarily steadystate)
equilibrium paths The arguments against such a procedure are based on the idea that the impact of
policy changes can be observed most easily since all departures from the steadymiddotstate can be
attributed to the altemalive polioy On the other hand the base case so dEfined as a steadymiddotstate is
consist~nl with previous work in a less dynamtc setting and Iherefore allows a common standard
for comparing model results
42 Computation algorithms
We are still al a stage in which basically each author uses a different computational technique
Th development of dynamic models - in parlicular wilh adjustment costs andior perfect foresight
bull
28
has corresponded with the decline in the use of fixed point algorithms In -fact given the relative
arge dimensions inevitably involved such algorithms lend to be very inefficient at the best and
chen prohibitively stow See Stone (1985) and Preckel (1965) for a comparative assessment of
different computation techniques Among the models surveyed only Ball1rdFulle(ton~
-ShavenmiddotWhalley (1985) and Feltensteln 1985 use Merrills variant of the fixed point
algorithm 1echnique
AuerbachmiddotKoHikoff (1963 1984 1967) follow a three stage procedure They first compute a
base case steady-state then a revised case steady state and finally transition path for the economy
berveen these two steady-states tn all stages a Gauss-Seidel iterative procedure is used
Ballard (1982) Ballard-Goulder (1965) Goulder (HiSS) and Summers-Goulder (1986)
~
use a method developed by Ballard and Goutder which is similar in many awects to the FairmiddotTaylQr bull
(1983) algorl1hm Short-run equilibria are calculated (using Merrills algorithm) parametric on
nrice expeC~lions The model is then iterated to generate self-fulfilling intertemporaf expectations
and the corresponding perfect foresight equilibrium In a relatively similar approach Andersson
1987 uses a simulation program SIMNON developed In the University of Lund This program
can handle two-paint boundary problems In a fashion consistent with the multiple shooting
algcrlthm (see Lipton-Poterba-Sachs-Eummers (1982))
Boyenbergs computational approach (1985 1985) differs from the other models surveyed in
that he relies heavily on analytical techniques Computations are done by using a dynamic version of
Johansons linearization method Being essentially determined by the continuous time nature of the
I I
f
model thiS linearization model has the disadvantage of confining the analysls to infinltesinal changes
around the base case equilibrium (see Bovenberg (1985) p 53) Pereira (1986b 1987d) uses an optimization algorithm NPSOL - developped by
rGillMurray-Saunders-Wrlght (1986)) This algorithm is used 10 compute the sequence of
I
29
Shortmiddotrun temporary equilibrium which makes up the intertemporal equilibrium path The
equilibrium conditions are seen as nonlinear equality constraints in the minimization of an artlcial
objective function The prices are normalized to the unit simplex by an additional linear equality I I 1 bull constraint
II ErlichmiddotGinsburgh-Heyden (1987) follow a unique approach in that they use a variant of the
optimization technique introduced by Negishl (1960) The economic equilibrium can be generated i
as a solution of a mathemallcal program Ihe objective function of which is a weighted sum of the
utility functions of the various agents while tho constraints set consists of the market dearing 1
conditions Ginsburgh-Heyden (1985) have extended Negishis resulllO tho case of downward price
lrigidities I
Finally the paper by Jorgonson-Yun (1984) is also unique in tIlat it is the only I
econometrically estimated model Different blocks for the consumption and production Side of the
-
model are separately estimated to provide the necessary structural parameters i
I The diversity of computation techniques is yet another indicator of the exploratory nature of the
body of literature surveyed in this article
43 Equal yield comparisions
The link between a base case and counteriactual simulations Is usually provided by the concept of
equal yield Shaven-Whaley (1977) discuss the meaning 01 equal yield in a general equilibrium
context when government is not allowed to run defICits Equal yield is interpreted to moan constant
public utility Government base case utility is maintained in the counferfactual experiments With
balanced budgets the concepl of equal yield is unambiguous The new equUibrium prices and the
balanced budge condition will del ermine the minimal expendilure and taxes needed to maintain base
30
case publi~ utility Accordingly in general equal yield is inconsistent with equal nominal tax
revenue Some change in tax revenue Is necessary Different tax replacement schemes ate
considered to assure that enough tax revenue is collected This is the approach essentially followed
by most 01 the papers surveyed Only Feltenstein (1985) Goulder (198) and Jorgenson-Yun
(1984) chose not to follow an equal yield strategy
The question is of how to Interprete lila concept of equal yield when the government is alowed to
fUn deflcLts The optimal leve of expenditure for base ease public trliUty can now be taAinanced
bond financed or financed by a mix of bonds and taxation We have several versions of equal yield In
particular equal yield may now be consistent with equal tax revenUE Furthermore some meaSure
ltgtf financial crowding out effects of government deficns can be inferred from the comparislon of the
several equal yield alternatives These issues are extensively discussed ~Pereira (1987b)
44 Price expectations and the dynamic generalization of compensation
Indicators
The sum of equivalent and compensation variations over households is the most widely used
aggregated measure of efficiency gains or losses In the context of $teady~state eomparisions andor
when complete future markets exist andlor perfect foresighl is assumed there are no difficulties
associated with the use of the standard Hicksian indicators In fact correct future prices are known
in these cases
If expectations are not selfmiddot fulfilling andlor future markels are not available a dynamic
generalization is necessary 8allardmiddotGoulder (1985) provide some steps in that direction by
defining an indicator that accomodates periods far into the future when households do not have
perfect foresight but a steady~state prevails On he other hand this issue is more fundamental in
i
[
I
i I I I I [
I I
bullI
31
the contex~ of a Gner1 lemporary equilibrium ftamawor~ In such Circumstances Pereira
(1986a) dev~lops a cnamlc generalization which is obtained as the present discounted value of a i i
sequence of short-fUn optimal expenditure functions consistent with a base case expected future I stream of IlWities I
32
5 Empirical evidenee from selected policy Issues
Dynamic tax models have generated several important results which escaped Sialic modellers
The following is a seleCpoundd set of issuos whIch stress the marginal benefits of dynam~ modelHng of
economic behavior in innovltlttive areas of analysis The discussion of a consumption tax emphasizes
the benefits of dynamic tusehold behavior and intergenerational aspects The study of the
flliminatlon or the re~intToduction of the investsnemt tax credits is made meaningful by the dynamic
modelling of production betlavor Modelling governmenl in a dynamic conlexts allows the modeller
to study the irrrpact of government deflcits Finally a dynamic framework lets us appreCiate the I
relevance of consumer expectations in terms of the evaluation of policy alternatives
i
i
51 Consumption Tax
t
By Its very nature a consumption tax can not be adequately investigated with a static model bull
Fullerion-Shoven-Whaicy (1983) use Ihe model of the US economy s described in
Ballard-Fullerton-Shoven-Whalley (1935) to evaluate the movement from the currenl US tax Isys1em to a progressive consumption tax Since their modeJ Incorporates a Jaborleisure choice
where leisure is an untaxed commodlty their results reflect the fact that both the consumption tax
and Ihe present system are dslortionary
Concentrating on the rntertemporaf distortions they show that sheltering more saving from lhe
current US income tax could improve econornic efficiency even if marginal tax rate increases are
necessary in order to maintain government fevenue~ At first you have a revenue shortfall
However the economy moves to a higher sustatned growth path and uitimatefy lower tax rates
cangenerate the same revenue path Also wages increase in the long run with this policy The
bull
bullbull ~
33 shyt
i
I
I
present value of welfare gi[lS for a potiey of complete savin9 deduction wilh marginal tate
adjustments (consumption lax) is simulated to be around $500 billion to $600 billion 01 1973
dollars
They also investiga1e th 1ngth of ttme it takes th~ economy to adjust to these policy changes
Roughly they estimate the long run to b thirty years although this figure Is very sensitive to the
specification of the savings elasticity - a crucial parameler in this model
52 Investment T~x Credit
Goulder-$ummers (1 S87) address the impact of eliminating the lnvestmenl Tax Cradit (ITC) on
intersectoral capital formation and on economic growth Their model isYa1ticularly adequate to
addies$ this issu~ in that is postulates a forward looking investment m9we with adjustment costs
They show that the eliminating the ITC causes a reduction in the rale of investment Inveslment is
estimated to fall by about 7 in the short-run and by about 12 in Ihe new long run steady state
In lurn a previous policy anncuncement lowers Ihe overall al1ractiveness of Investment and leads 10
a downward shift of the Investrrent profile On the other hand the combined effect of a revenue
neutral simultaneous efimiraron of the ITC and reduction of the corporate tax rates is a long run
reduction ef the capital Sieck by 35 This pattern suggests that a revenue neutral increament in
both Ihe corporale tax rates and the inveSlment lax credits would be preferable 10 the formula
adopted in Ihe US Tax Reform Act of 1986
Pereira (1987d) addresses the impact of the ITe from Ihe oland point of Ihe curTenllax syslem
under the Tax Reform Act of 1986 Given Ihe concern over the potential depressiVe effects upon
savings and investment of the new tax system in conjunction with deficit reduC1ion mechanisms of
the GrammmiddotRudman type a pertinent question is should Ih lTC be e-introduced In the context
bull
34
-of his mod~1 of the US economy Pereira focuses on the trade-off between the distortion in the
intersectoral allocation of capital induced by the lTC the lower tax revenues it generates and
potential financial crowding-out effects on one hand and the positive effects of lowering the relative
price of new capital goods on the other hand Preliminary results confirm the qualitative results in
Goulder-Summers (1987) suggestting a welfare gain of about 2 of the present value of GNP in
the best scenario of absence of replacement taxes Furthermore preliminary results show an
important time pattern to welfare gains with the average benefits increasing the further you look
into the future This pattern is due to the presence of constraints to intersectoral mobility of capital
and inertia in the adjustment towards the optimal capital levels as reflected by the presence of
adjustment costs
I
I
bullAuerbach-Katlikaff (1gB) in the context of their intergenerational model of the US economy
consider the impact on savings capital formation and interest rates of deficit policies and balanced
budget increases in government consumption They conclude that deficit finance and government
consumption can significa-ntly crowd out capital formation and lower the welfare of future
generations However crowding out from deficit finance is a very slow process because it results
from increased government spending over potentially long horizons Also deficit policies may
substantially influence the longmiddotrun interest rates while leaving the short-run rates essentially
unaffected Finally and in opposition to the central role adjustment costs seem to play in both
53 Financial crowding outeff~cts of government deficits
Goulder-Summers and Pereiras models Auerbach-Kotlikoff suggest that the time path of interest
rates induced by a policy of deficit financing seems to be insensitive to the presence of adjustment
costs
bull
35
54 The role 01 consumer expectatlons
The expectations analysis has uncovered one of the benefits of dynamic modelling
Ballard-Goulder (1985) find that the appeal of adopting a consumption tax depends on the level of
foresight possessed by the consumers Furthermore additional foresight may be welfare worsening
This is a second best type of result This tends 10 occur under policies that lead to capital deepening
and declining rate of return to capital over time To the extent that consumers have more foresight
they will be beller equippod to anticipate the fall in the rental price of capital Consequently if a
saving incentive is enacted people save more with myopia than with perfect foresight Given the
exi~tence of taxes on capita and the discrepancy between the private an~--sccial returns to capital
the greater savings with myopia is beller socially Ballard-Goulder find that the welfare gains
from a consumption tax are reduced by about 10 when we move from myopia to farleet foresight
Therefore the attractivenness of adopting a consumption tax seems fairly [obust across these
foresight specifications
I
bull
36
6 Tbe power and weaknesses 01 dynamic modelling the example 01 corporale
tax Integrallon
The Issue of corporate tax integration allows us to sh~w the stre-ngths and weaknesses or the
dynamic approach Empirical evidence using CGE techniques indicates that depending on the precise
scheme of integration and the 1ax replacement methods integratiOn may have substantial effects In
Ihe work of FuliertonmiddotKingmiddotShoven-Whalley (1980 1981) total integration was found to yield an bull
annual static eHielency gain of $4 to $8 billion in 1973 dollars Simulated dynamic gains may be as
large as $695 billion or abcu 14 of the present value of tulUre consumption and leisure in the
US economy These gains result primarily from interindustry reallocalions of Investment and an
improved inter~emporal allocation of consumption
Mora recent work emphasizes hOw consumers asset portfolio decisions and firms finanCial
decisions affect the efficiency gains from integration Slemrod (1980) focusing on consumers
asset ponfoio deCisions finds static efficiency gains which are about twice -as Iarge as those
repOrted iiy Fulierton-King-ShovenWhalley (1961) FuUrlon-Gurdon (1983) focus on the
firms financial decisions They roporl efficiency gains of 6 of GNP from the elimination of the
tax distor1ions favoring debt However when they eliminate the corporate tax and repiace it with
increaSed persona income taxes addlHonal dls10t1ions are created in the optimal labot~leisure
decisions These distortions tend to dominate the analysis slJ(h that the overall effects of complete
integration are very modest Galpr-LuckmiddotTodr (1986) address simultaneously consumers
asset pOrlfollo decisions and the firms financial decisions They also find very modest efficiency
gains from integralion
The above results are important but may be severely biased There are severa aspects of
economic behavior and modelling crucial for the study of Income tax Integration whiCh have no been
i 37
captured innny of the above papers One aspect is the absence of government deficitsmiddot a balanced
budget is assumed It is true Ihal resource crowding OUI is caplured In these mOdels but financial
crowding out induced by government spending is nOI Second investment is not derived from
optimtztlcn behavior Investment behavior passively accomodates endogenous saving decisions As
a consequence the differential impacts of policies in the incentives to save and to invest are not
captured Aso full capital mobility across secors is assumed with instantaneous capital
adjustmenlS towards optimal levels This assumption rules qut different costs of capital across
seclors and therefore differentiated reactions to tax policies changes Finallyhe modelling of both
gov9rnmerlt deficits and of endogenous real and financial investment decisions necessitates the
~nsideraiion of a dynamilt framework and the introduction of financiaJ assets govemment bonds and
iprivate financial assets A dynamic framework also highlights the efficiency effects of Integration on r
ithe ptimal intertemporal decisions The above models are either sIalic (Slemrods and
GalpermiddotltJckemiddotTode(sj or a dynamic sequ~nce of otherwise Slatic model bullbull I Pereir a (1986b) develops a dynamic applied general equilibrium model bull to study the tbull
efficiency effects of integration on the growth and allocation of investmenl across sectors Special
Iattention is paid to the structural affeelS of resource and financial crowding oul induced by
gove mect spending and deficits and to the real and financial Investment reactions across industries
to both tax integration and flnancial crowding out His mode departs from previous work on income
lax Integraticn and for that matler from most 01 the CGIE literalure for tax policy evaluation in
several direcllons It encompasses an endogenous sequential equilibrium struelure founded on
dynamic behavior with flexible expectations Government deficits are optlmaUy determined
Investment decisions are tward looking and the resull of optimizing behavior Several financial
assetsmiddot public and privatemiddot are considered
Simulation resulls Indicale that the welfare gains from Integration under large deficits are at
32
best lJlodest when measured in terms of the GNP The elimination of the corporate tax and its I replacemenl by increased income lax rates yields long run benefits which are never larger than
2 of the presenl yalue ct telr consumption and leisure onder the previous tax regime and 1
under the current tax regime However the short run w~lare effects of full integration tend 10 be
very low and in in some scenarios even negative This is a flew intertemporar pattern of efficiency Ieffects which reflects an adjJsiment lag in the interindustry investment decisions due to the
existence of costs of adjustment On the of her hand partial integration achieved by excluding I dividends from the corporale lax base syslematically yields negalive efficlencyeffeclS In Ihe long Inm these can be as high as 3 of present value of the future value of consumption and leisure This
is a new second best effect suggesllng Ihat less than complete integration may have perverse I efficiency effects
The dynamic slructure of PerclrBs model snowing for an improved freatment of real investment
declsions and government deficts provides a potential setting for a more accurate measure of the
costs and benefits of inlegration The simulation resullS suggest lower benefits and an intertempora Ibull
pattern of growing benefits
Simulation results also clearly suggest robust negative effects from partial integration How
reliable is this result If dJvidends were deductible from the corporate lax base (as are interest
payments) the preferenlial Ireatmant of debl over equily would be ellminaled Corporallons should
be expecled 10 react by decreasing the optimal deWeqully rallo However corporale financial
gecisions are exogenously set In Pereiras model as in aft of the other models surveyed that go as far
as dealing with the issue
Also withdrawing dividends from Ihe corporate tax base is a way of eliminating Ihe double
taxation of dividends at belh Ihe corporate and personnal Income levels How will the optimal
aliocellon of household saving accomodate that change in Ihe relalive return to corporale equity
39 I
t However hOusehold portfolio decisions are exogenouly set in Pereiras model as in all of the other
models surveyeltj go as far as dealing with the issue
IIt is sate to say that the evaluation 01 the benefits 01 partial Integration as defined above are
sevrey underestima1eO Meanwhile the distortionSoenerated by the increase in the marginal
personal tax rates designed to ralse the revenue foregone by the dividend exclusion are fully
aCCltlunted for A good measure of the costs of integration together With a poor evaluation of the
benefils may very well be the reason why partial integration is Simulated 10 yield neg alive
eHikiency gains
On a different vein as most of Ihe model surveyed here Pereiras is a closed economy model It
is legitimate to wonder how the resullS would change If international capital flows were anowed
IThis almost certainly would affect financial crowding out since a good d~al of the capital inflows
could be used to finance public debt
For these various reasons the results should be treated as preliminary but they strongly I suggosl that the dynamic structure provides valuable new insights into the corporate tax integration SSU2
40
7 Concluding remarks I I I i
What has been acccmplished Ihis far The success of Ihe CGE research in laelding the challenge
of dynamic modelling can not be denied
I iGreat progress in the modelling of household behavior has been achieved Promising
developm~nts in the modeUing of production and government behavior have also been accomplished
Interesting innovations in the concept of equilibrium and computation techniques were discussed
The policy analysis was greatly enriched by considering transitional effects together with longer
run steadymiddotstate equilibrium paths generalized equal yield strategies accomodating different
financial crowding out impacts and generalized dynamic policy evaluation indicators
An important set of issues have been addresed by the models surveyed Traditional issues
focusing on the effects of capital taxation and consumption taxes have been pursued In terms of
capital taxation for example the intertemporal nature of the issue was enhanced by allowing an
optimal evolution of capital stock in the economy and forward looking optimal investment decisions
In Andersson (1987) GouldermiddotSummers (1987) and Pereira (1986b 1987d) this is coupled
with a improved treatment of several tax provisions like investment tax credits and depreciation
allowances In terms of the consumption taxes developments in the specification of Intertemporal
household behavior the introduction of overlapping generations and the modelling of
intergenerational links via bequest motives as well as a closer attention to demographic evolution
provided a much improved economic setting for the understanding of the several aspects of the
problem
Furthermore dynamic modelling has permitted the addressing of a set of new issues Policy
Issues ranging from dynamic tax policy analysis to the evaluation of exogenous changes in
government expenditures to the impact of different methods of financing government expenditure to
4 1
the importance of tinanciai croHrii1g out to 1he relevance of consumorS expectations in tne policy
results have been addressed
Despile all of the progress and in part due to such progress several avenues are wide open for
much neded additional research The following seem to be the most Interesting and promising
areasmiddot First the specincation of liquidity constraints to household behavior may help to obtain a
better understanding of policy chenges that affect directly the interest rates in the economy Second
major attention needs to be paid to the incorporafion of a roraign sector (with open capital markets)
Third some efforts are needed in terms of finding adequate computation techniques now that
dynamic modelling eally makes the curse of dimensionality worse than eVOrbull
The modelling of financial markets is the single most unsatisfactory speet of the dynamic
models surveyed here~ The problems associated with provlding a mo~ adequate treatment of
financial markets and in general endogenous and optimal financial deci$ions bull both at the corporate
and at the personal levelsmiddot have 10 be addressed That necessarily requires Introducing uncertainty
into the models To b adequate this must go well beyond the mere assumption of point expeetalions
with the whole array of modelling implementalion and evaluation problems thai generates
bull I i
42
REFERENCES
Abel A and O Blanchard 1983 An Intertemporal Model of Saving and Interest Econometrica
51 pp 675-92
~ Andersson K 1987 Tar-ation of Capital in Sweden - A General Equilibrium Model Ph D
Dissenation UnIversity of Lund Sweder
Auerbach A L KotlikoH 1983 National Savings Economic Welfare and the Structure of
Taxation in M Feldstein ed Behavioral Simulation Methods in Tax Policy- Analysis Chicago
University of Chicago Press
Auerbach A l Kotlikoff 1984 Social Security and the Economics of Demographic-
Transition in H Aaron and G Burtless eds Retirement and Economic Behavior Washington DC
Brookings Institution
Auerbach A L Kotlikoff 1987 Dynamic Fiscal Policy Cambridge Cambridge University
Press
Auerbach A L Kotlikof and J Skinner 1983 The Efficiency Gains from Dynamic Tax
Reform Internatiooaj Economic Reyiew 24 pp 81middot100
Ballard C 1983 Evaluation of the Consumption Tax with Dynamic General Equilibrium
Models Ph D Dissertation Stanford University
Ballard C 1987 Tax Policy and Consumer Foresight A General Equilibrium Simulation
Study Economic Inquiry 25 pp 267-284
Ballard C and L Goulder 1985 Consumption Taxes ForeSight and Welfare a Computatonal
General Equilibrium Analysis in J Piggott and J Whalley eds New Deyelooments in Apolied
General Equilibrium Analysis Cambridge Cambridge University Press
Ballard C D Fullerton J Shoven and J Whalley 1985 A General Equilibrium Model for
t 43
Tax Eoliey Eyalua~iQO NBER University or Chicago Press
Borges A (1987) bull A Survey of Compula1ional General Equilibnum Models CECD Economic
Slud~ 8
Bovenbig L 19B3 imperfect Capital Mobility in a Perfect Foresight Herberger Model
Mimea UC Br~eley
Bovenberg L 1984 Capitel Accumulation and Capital Immobility Qmiddottheory In a Dynamic General Equilibrium Framework PhD dissertation University of California at Berkeley t
Bovenberg L 19850 Dynamic General Equilibrium Tax Models with Adjustment Costs in A iManne ed bull 1985 EtQnomrc ECUjilbrium Model Formulation and Solution Mathematical
lProgramming Siudy 23 Arnsterdam NorthmiddotHoliand
Bovenberg L t 1986 Capital Income Taxation in Growing Open Eeondmies Journa of Pubtic
Eamp(omlcs 3t pp 347middot376 IBovenberg L ald W Keller 1981 Dynamics and Nonlinearlties in Applied General
Equilibrium Models Internal Report Mlmiddot81middot208 Department for Statistical Methods CBS bull
Vooiburg The Netherlands
Charnley C 1981 The Welfare Cost of C~pltal income Taxation In a Growing Economy
IJournal of Political ECQnQm~ 89 3 pp 468middot496 I
Charnley C 19810 Efficient Tax Reform in a Dynamic Model of General Equilibrium I Mimeo Yale University I Charnley C 1982 On the Optimal Taxation In Stylized Models 01 Inlartemporal General
Equilibrium Mimeo Yale University I Decaluw B and A Manens 1985 Pays en Oeveloppement at Modeles Calculables OEquilibre
General - Une Revue de al litterature empirique Monographie Sene 9 Volume 3 Centre de
Recherche et Developpemont Economlque Unrslt de Montreal
44
bull
I I
I
I
1
Erlich $ V Ginsburgh and L van der Heyden 1987 Wher~ do Real Wage Policies Lead
Belgium bull A General Equilibrium Analysis European Economic Reyiew (forthcoming)
air R and J Taylor 1983 Solution and Maximum Likelihood Estimation of Dynamic
Nonlinear Ratioant Expeclations Models Ecooometrica 51 pp 1169middot1186
Feltenstein A 1984 Money and Bonds In a Disaggregated Ec~nonYin Searl 1-1 J Shoven
Eds Aoolied ~eoeral EqJiliDCiulD tioalYsis Cambridge University Press Cambridge
Feltenstein A 19B6 A Dynamic General Equllibrlu Analysis of Financial Crowding Out
Theory with an Application to Australia Journal of Public Economics 31 No1
Fullerlon D R Gorden 1983 A Reexamination of Tax Distorllons in General Equilibrium
Models in M Feldstein Ed Behayora Simulation Methods in Tax pOlicy Analysis Chicago
University of Chipgo Press
Fvllerlon D J Shaven and J Whalley 1983 Replacing the US Income Tax with a
Progressive Consumption Tax JoulOal of Public ECQnomics 20 pp3middot23
Fullerton D Y Henderson J Shaven 1984 A COmparision of Methodologies in Empirical
General Equilibrium Models of Taxation in Searl H J Shoven Eds ~pQlied Geoaral Jguilibrium
lIoalyss Cambridge University Press Cambridge
Fullerton 0 A King J Shaven and J Whaney 1981 Corporate Tax Integration in the United
Statesmiddot a General Equiiibdufi1 Approach American Economic Review 71 pp~ 677~691
Galper H R Lucke E Tader 1985 Taxation portlor Choice and the Allocation of Capital A
General Equilibrium Approach Mimeo
Gill p W Murray M Saunders M Wright 1986 Users Guide for NPSOL (Version 40) A
Fortran Package for Nonlinear Programming T echnicsl Report SOL 86middot2 Deparlment of
Operations Research Stanford University
Glnsbur~h V and L van der Heyden 1985 General Equilibrium with Wage Rigidities An
45
AppJication to Belgium in A Manne ed t Economic EQumprit1m~ Made Formulation and $oJioo
Mathematical Programming Study 23 Amsterdam NorthmiddotHolland ~
Goulet J 1968 1 Adjustrent CO$1S in the Theory o(the Firm Reyiew of EconomiC Studies 35 bull
pp47middot55
Goulder L 1985 Tax-Financed versus BondmiddotFinancedGhanges in Governent Spending
bull Mimeo Harva(d UniversilYmiddot
Goulder l J Shaven and J Whalley 1983 Domestic Tax Policy and the Foreign Sector The
importance of altemative foreign policy formulations to results from a general equilibrium tax
analysis model In ~n methods in Ibe antico of income from caoia1 ed Martin
Feldstein Chicago University of Chicago riSS
Goulder l L Summers 1987 Tax Policy Asset Prices and 3rowtly ~
A General Equilibrium Analysis NBER Working Pper No 212B
Grandmont J 1982 Temporary General Equilibrium Theory Ch 20 in Handbook 01
Mathemalical Eccnornics Vol II NorthmiddotHoliand
Harberger A 1959 The Corporate Income Tax An empirical Appraisal In lax Revision
Comoendjum Voll House Cornmitee on Ways and Means Washington DC Govemment Printing
Office
Harberger A 1962 The Incidence of the Corporate Income Tax JOULn1 of Political
Economy 70 pp 215middot240
Harberger A 1964 Taxation Resource Aliocalrlon and Welfare fn lb Role of Dir bullbulll and
Indjresl Taxes io tbe Fede[al Reserve s~I~Ill Princeton Princeton Unlveresity Press
James J 1985 New Developments in the Applications of General Equilibrium models to
Economic History in Piggott aand Whalley Eds
Jorgenson D and K Yun 1984 Tax Policy and Capital Allocation Harvard Inslitute of
bull
I
I f
Economic Research WP 1107
Judd K 1965 ShonmiddotRun Analysis of Fiscal Policy in a Simple Perfect Foresight Model
Journal of poJiHcal 5CCJiW1l pp 298middot319 r
Upton D J Potbc J Sachsa nd L Summers 1982 Multiple Shooting in Rational
Expectations Models fuoometrlc 50 1329middot1333
Manne A ed 1985 fconQrnjc Eoullibrium Model FOrIDujaljQo and Solution Mathematical
Pogrammlng Study 23 Amsterdam NorthmiddotHolland
Manne A 1985 Or the Formulation and Solution 01 Economic Equilibrium Models In A
Manne ed ECQOQlli-Jr~ibmiddotum ~tQdel EQlJulailon and Solution Malhematical Programming
Study 23 Amsterdam NorthmiddotHolland
Mansur A and J Whalley 1984 Numerical Specification of Applied General Equilibrium
Models in Scarf H J Shoven Eds Aoplied General Eouillbrium AnaiysectIsect Cambridge Universily
Press Cambridge
Merrill 0 1972 Applications and Extensions of an Algorithm thaI Computes FixedmiddotPoints of
Cenain Upper Semimiddotcominuou Point to Set Mappings PhD Disserlation Unlvel$ity 01 Michigan
Negishi T 1960 Welfar Economics and Exislence I an Equnlbrlum for a Competitive
Economy MelroeCOOQ11ka 12 92middot97
Pereira A 1985 On the Existence and Uniqueness of Optimal Ou1put Path lor CRTS Firms in
Adjustment Costs Technologies Mimeo Stanford University
Pereira A 1985a Consumer Expectations and Ihe COSI 01 UllUty In an Intertemporat
Framework Mlmeo Stanford University
Pereira A 1985b Corporale Tax Integrallon and Government Deficits A Dynamic General
Equilibrium Anaiysls Mlmeo Stanford University
Pereir A 1987 On Ihe Computation of the Users Costs of Stocks Ecooomi (forthcoming)
47
Pereira A bull 1987b Equal Yield Alternatives and Government Oeficlts Mimeof Stanford
University
Pereira A 1987c A Dynamic Applied General ~uHlbrium Model for Tax Policy Evaluation
Comlletamp Docurnent2tion Mimeo University of California San Diego (in progress)
Pereira A 1987d Should the investment Tax Credit be Rei(ltroduced Mlmeo University
of California San Diego (ir progreSs)
Piggott J and J Whalley eds 1985 tJew QeyeQpments in Applied General EQuilibrium
bullenalysjs Cambridge Cambridge University Press
Preckel Pbull 1985 Alternative algorithms for computing economic equilibria In A Manne Ed
Radnor R 19amp3 Equllibrium under Uncertainty In K Arrow M Intrilligaor ods
I ibull
Robinson S 1 g86 Mullisectorai Models of Oeveloping CountrieS a Survey Oivision of
Agricullure and Resources UC Borkeley WP 401
Scarf H 1967 The approximation of fixed points of a continuous mapping ~MM Jouwal of
epplid Mathornallcs 15 pp 1328middot43
Scarf H with T Hansen 1973 00 the Computation of EconomiC Eguilibria New Haven Yale
University Press
Scarf H J Shoven Eds 1984 61l1llied Genera Elujlibrium 6nalyssect Cambridge University
~ Press Cambridge
Shoven J 1976 The Incidence and Efficiency Effects of Taxe on Income from Capital
Jurnal of ramie1 Economy 84 pp 1261-83
Shoven J 1983 Applied General Equilibrium Tax Modemng IMF Slaff PaoerS 30
Shoven J and LB Simon 1986 Share Repurchases and Acquisitions An Analysis of which
Firms Participate NBER Working Paper No 2243
bull
48
IShaven J and J Whalley 1972 ft General Equmbrium Calculation of lhe Effects of
Differential Taxation of Incoma from Capital in the US Jauroe of public Econgmics 1
pp281middot321 (
Shoven J bull and J Whalley 1973 General Equilibrium with Tax$s a computation procedure
and an existence pr00f ReviEhi of Economic Studies 40 pp 475middot490
Shoven J J Whalley 1977 Equal Yield Tax Alternatives A General Equilibrium
Computa~ui0nal Tfchnique J~HJrnal of PUblic EconomIcs 8 pp 211~224 bull
Shaven J J Whalley 1984 Applied GeneralmiddotEquilibrlum MOdels of Taxalion and
International Tr8de An Introduction and Survey Journal of Economic literature 23t pp
1007middot1051
Slemrod J 1980 A General Equilibrium Model of Capital Income Taxation Ph D bullDissertation Harvard Univnrsity
Slemrod J 1983 A General Equilibrium Model of Taxation with Endogenous Financial
Behavior in M Feldstein edbull Behavioral Simulation Methods In Tax Policy Analysis Chicago
UniVerSity of Chicago Press
Stone J 1985 Sequential Optimization and Complementarily Techniques for computing
Economic equilibrium in A Manne Ed
Summers L 1981 Taxation and Corporate Investment a q-Theory Approach BroQkinas
Summers L 1985 Taxation and the Size and ComposHion 01 the Capital Stock An Asset Price
Approach Mlmeo Harvard University
Whalley J 1975 A Geneal Equiloibrium Assessment of the 1973 United Kingdom tax
reform Economica 42 pp 139middot61
1JLTIllOS JIQll~)ltG PAPERS PUBLICADOS bull
I
I I
j
112 70 - LUCENA iJiCIIJo ~ IIEnvironrilent Honitoring and Orqanization Structure T (Setellibro 1987)
nI 71 - L~ENA Diogo A Note on the Representation of Information Stxueture (SeteJObro 1987)
nQ 72 - LUCENA D1CgO Envuorment 111)nitoring and Organization
Structure II (Setelilhro 1987)
ril 73 - AlmmE3 Antonio Pais e CAsPAR Yitor bull Tributacao Inoentivos e InvesUlllentos Analise Qualitativa noYellibro 19S7)
n2 74 - liAlltOSA rose Pedro Opt1lllal lIage R1gidity A SUggestedtlthodology to Teet tb Theory and an Application (Outubro 1987)
nI 75 - CABRAL L 11 B Three lIotes on SYJIIlletric Gemes witt itsYJIIlletric Equilibri (NoveJObro 1987)
n2 76 - LUCENlt D10gOTo Search or Not to Search (Fevereiro1967)
nQ 17 - SA Jorge asconcelos euroI A Theory of SynergyB (Fevereiro1987)
n2 76 - VILARES tanuei Jose Os Be IntenaMios IlIIportados ColIIO Factor de Produao (Julho 1987)
nQ 79 - SA J~rge VasCQncel03 e HoY to Copete and Co~unicate in tature Industrial Products_ (aio 1988)
n2 80 - ROIl Rafael Learning and Capacity Expansion in a New larket Under Uncertainty (Fevereiro 1988)
n2 81 PEREIRA Alfredo llarvilo Survey of Dynamic Computational General EquilibriUlll Hodel for Tax Policy Evaluation (Outubro 1987)
Qualquer in ormacao sabre o 1IIorlltlng Papers ja publlcados sera prestada pelo Secretariado de Apoio aos Docentes podendo os ~e3OSf
ser adquirido5 ne Secyo de Vlndas da Faculdade de EconollLia UNL na Travessa Esteviio Pinto Cltpolide - 1000 LISllOA
I 17
to above gynamlc producion and investment b~havior are middotinduced by the existence of capital
adjustment costs and linked to Tobins q theory Adjuslment costs are designed to capture both the
incomplete mobility of capital aoross industries and installation costs ~e the costs of adjusting
capital towards its cpHmai level)
Adjllstment costs can be conceived as internal to the firm and measured in terms of foregone
outpul along the lines of Lucas (1967) Alternatively adjustment costs can be viewed as actual
external cests incurred together with Ihe purchase costs along the Unes of Gould (1969) With the
exception of Pereira (198Gb 1937d) all of the CGE authors follow the former approach
The firms in the economy maximize their market value as the present discountad value of the
future stream of dividends In Andersson (1981) and Pereira (198Gb 1987d) firms are seen as
maximizing the present discounted value of net cash flow Maximization is constrained by Ihe
adjustmenl cosl lechnology and ~n equation of motion describing Ihe evolution of the capital stock
It should be noted that undereogenous divldendrelention rules the two problems are equivalent II
should also be nOled Ihat a salisfectory economic rational for th bullbullistoo of dMdonds with the
present tax code is missing in the profession (Shoven (1986))
The models with static formulation of producers behavior are characterized by passive
investment behavior Investment merely accomodates to saving in the eccnomy With a dynamic
formulation induced b adjustment costs real investmen decisions are forward looking Investmenl
Is endogenously and optimally determined by the firms A fundamental difference between the
shonmiddotrun in which capital stock is given and longmiddotrun in which Ihe level of capital is allowed to be
optimally determined is emphasized
This extra richness of produc1ion dynamics is not without costs A careful look at the summary
tables will clearly show an inverse relation between the adoptlon of dynamic features in production
and the level of disagreggation of the production side of the economy In fact with production
bull
j
1 6
dynamics $e dimension of the problems is immensly increased Let IS be more specific
In a static framework with constant returns to scale production technology the output level ismiddot
indeterminate The optimal aliocation of inputs can be obtained by cost minimization with any
feasible output level generatillg zero profits Such zero profit conditions are used to solve for the
output prices in terms of the factor prices and hencemiddot to reduce the dlmensionality of the problem
from the -number of commodities and factors to the number of factors Thus even if the model deals
with 30 production sectors the computation of an economic equilibrium can take place using only
the dimensionality of the primary inputs in the economy
Now under certain regularity conditions on the production and adjustment costs technologies
leading to enough concavity of the optimality objective the intertemporal output path for the firm is
endogenously optimally and uniquely determined even with constant relums to scale technologies
(see Pereira (1985 1987a) Accordin~y with adjustment costs in the model optimal profits
will in general be non-zero This result has crucial implications for the computatiQn of equilibrium
in the model~ with production dynamics With adjustment costs no reduction of dimensionality is
possible The curse of dimensionality returns as a binding constraint
The introduction of adjustment costs adds another significant complication to the model With
capital being less that perfectly mobile in the economy different rates of return on capital will
exist in different sectors This is a difficult problem to tackle conceptually in the absence of
uncertainty Also in the real world producer maximization choices include also its choice of
financial ratios (debVequity) and payout rates (dividendretained earnings) These financial
subjects are difficult and much Clf this behavior is still taken as exogenous by the modellers We
will come back to these issues below
middot i
19
bull 33 government behavior
middot
Two issues dominate the modelling of government behavIor First govemment behavior has
Ibeen typically seen in Ihe CGE liIerature for lax poliCY evaluation as constrained by yearly balanced
I budsets (see for example 8allard-Fullrton-S~ovn-Whafley 098Sraquo The analysis of
govemment defidlS and public debl in a CelE context requires a dynamic selling Second the level of
government expenditures is either exegenouty given as in Auerbach-KoUikofi (1984 1987)
6ovenbrg (1984 1985) Feltenstein (1984 1986) and Jorgenson-Yun (1984) or
endogenously (but no optimally) determined by the balanced budget conditions as in
6allard-Fullerton-Shoven-Whalley (1985) Andersson (1987) Erllch-Ginshurgh-Heyden
~
(1987) Gouder (19aS) and Gouder-Summers (1987) In the second case the composition of
bull public expenditures is often optimally determined However the leve~ of government expendi1ures
can only 00 endogenously and optimally determined if the government is seen as an optimizing agent
and is allowed to run deficits
The first attempts to deal wiU the government defiCits in CGE tax models are due to
Auerbach-Kollilltoff (1984) and Feensain (1984) In Auerbach-Kotlikof (1984) government
expenditures afe exogenous They grow at the rate of growth of population However given 1he tax
structure and tax revenues yearly deficits and surpluses are atTowed subject to an in1ertempond
constraint that the present value of future tax revenues equals the present value of future
expenditures
In Fellenstein (1984 1986) government expenditures are also exogenously given He also
allows the government to run deficits to finance expenditures in excess to 1ax revenues However
surpluses are returned to CQnsumars in the form cf transfers Accordingly government ts not
subject to any constraint regarding the future repayment of public debt
In GouJder (19851 the 9Vernmenl maximizes a static social welfare ~nctio subject to a-
bull
20 I i
I
I I I
I bull
balanced budget conSirtlL This follows the optimal alocatkm of government expenditures alpng
the lines of Baliard-FHeiton-Shoven-Whalley (1985) Ttle model Is generallzea to allow for
exogenous changes in HIe time path of government expenditures Two financing alternatives are
consjdeft~d and contr8s1ed EldditionaJ tax revenues and bond issuance
Pereira (198Gb 1137d) auempts to address both the incorporation of deficits and the
determination of govement expenditures The path of government expenditures and the path of
delicilSlsurpluses (ano therefore the path lor debt) are endogenously and optimally determined The
government is seen as mltJximizing n intertemporal social welfare function given the 13x structure
Optimization is subject to a sequence of recursive equations of motion reflecting the evolution of the
ptlbllc debt allowing fDr government budget imbalances It should be stressathat this specification
01 the government cOMtcalnt is equivalent to the specificaUon in Auerbaeh-Kotlikoff (1983 1987)
when no liquidity constraints affect -gove~nment behavior
The extra richness in the treatment of government behavior allows tfte examination of
government debt poices in a truly dynamic setting Also financial crowding out effects induced by
government deficits can be analyzed (see Auerbach-Kottlikoff (1987) Feltenstein (1986) and
Perer (198Gb 1987e))
34 Foreign sector
Most of the open economy models follow the assumptions of balanced trade with import and
export net demands characterized by constant elasticities along the line of Ballardmiddot
-Fullerton-Shaven-Whalley (1965) Such is the case of Salfard (1963) Ballard-Goulder
(19S5) and Goulder-Sllmmers (1987)_ In Feltenstin (1985) the rest of the world is treated as
21
an addjlion~ consumer group
Bovcoberg (1986) develops a model In which two economies lIr considered each following
intertempoml perlect foresight paths These economies meet in the international forum Their
trade relaroMhips are characterized by yearly balanced trade accounts
None of the new generation of dynamic cae tax models has yet IncorPorated the lnternational
capital flows as done in aoulder-Shoven-Whalley (1983) for the earlier Ballarrshy
-Fullertonmiddot Shaven-Whalley (1985) model This is unfortunate as there is an important
Intertmporal aspect to international lending The first al1empts along Ihese lines are due to
Andersson (1987) and Erlich-Glnsburgh-Heyden (1987) Andersson (1987) in his model of the
Swedish economy adopts a closed economy appraoch in which rates of return in the domesJic
economy are largely determined by the international capital markets Fidinterest rate induce
Intlnatlonal capital flows which determine and finance the International lrade imbalance In
lurn In ErlichGinsburghHeydenmiddot (1987) foreign trade Is generaled according to an
intertemporal Irade welfare lunction with constant import and export elasticities In the shortmiddotrun
they allow inlernational trade inbalances which generate capilal ftows to the domestic households In
the long run however trade balance is assumed
35 The need for financial markets
A dynamic economic Wucture not only provides the ideal environment 10 model many features of
economic behavior it also permits one to Incorporate the financial side of the economy If the
government is allowed to run deficits the question 01 deficll financing automatically follows If
investment is optimally dewmined and returns to capital are different aClOSS sectors problems of
investment financing arise If there are seeral final1cfal assets In the economymiddot government
4-
I i
ibull
I I I I
I
j
22
bonds privpte bonds and equity (or simply physical capital installed inmiddot different sectors) the
problem of allocation of saving among assets with potentially different returns arises
The papers surveyed var greatly with respect to l~e extent of their attention to the financial
side of the economy At Clne extreme are the models in ~anardmiddotFuliertonmiddotShoven-Whalley (1985)
Andersson (1987) Ballltld (1983) Ballard-Goulder (1985) Bovenberg (1985 1986) and
Erlich-Gjnsburgh-Heyden C19B which are devoted exclusively to the real side of the economy
In turn Auerbach-KotliKoff (1984 1987) Feltenstein (1984 1986) and Goulde (1985)
allow for government debt In these models saving finances changes in government debt and physical
capital Private and publlc 2ssets are perceived by the households as perfect substitutes The
allocation of saving merely adjusts to the relative demands for funds
Feltenstein (1984 1986) is the only model surveyed which introduces money Government deficits are financed by issuing money and bonds according to an exogenously given rule Money is
demanded by consumers for transaction motives and an exogenously given fraction as a store of value
On the other hand government bonds and physical capital are the vehicles for the intertemporal
transfer of wealth
Summers (1985) and GOlldermiddotSummers (1987) introduce a whole menu of financial assets shy
firm specific equity capital Different assets earn different rates of retum However such rates
are equal up to constant and exogenous sector speCific risk premia Therefore the introduction of
constant exogenous risk premia as helpful as it may be in the context of calibration does not solve
the main issue of the non-optimality of the allocation of saving Also talking about risk premia in a
deterministic context is somewhat unsatisfactory
Pereira (1986b 1987d) also introduces a whole menu of assets private and public bonds and
firm specific equity However all the assets are expected to yield the same rate of return and
therefore perc~jved as perfect substitutes The different asset types allow consideration of
23
exogenOU$~8tUequjty triO dividendlrc~enlion rufes and therefore seve-ral sources of investment
financing bond equity and retained earnings
The nonmiddotcptimalHy of the allocation of saving and the absence or exogenelty of corporate
financial rules is _a reftectlon of the limitations of the determInistic approach Either in a context of
perfect anticipation oi the future prices Of in general within the t$atm of point price
expectations all the papers follow a deterministic approach Under such circumstances consumers
either expect diHerent rales of return (inclusive of risk premium) across assets in which case
they wilt buy only one asset (that with highest rate) Of they expect equal rat~s of relurn in which
case Ihey are indifferent about the asset composition of Iheir portfoliO There is no way of
Iradingmiddotoff rales of return and risks to obtain an optimal interiof solution to the problem of the
allocation of saving
The most advanced contribution in the modelling of saving allocation in a CGE setting is due to
Slemrod (1980 1983) in the context of a static one-period model In his model consumers act
according to a two stage separable decision process They Hrsl decide on how much to save Then
they decide on the allocation of saving according to an Indirect uttlity function dependent on the rates
of ratum and variances offered by the different assets in the eoonomy The sourCe of riskiness in the
ampConory comes from an uncertain marginal product of capital On the other hand aside from
portfolio decisions the res of the economy is insulated from uncertainty
The most complete contribution in terms of the treatment of the corporate financial rules is
FulienonmiddotGordon (1983) In a variant 01 Ihe BailardmiddotFulienon-ShovenmiddotWhelley (1985) model
Fullerton and Gordon have capital intensity and optimal financial decisions Jointly deter~ined
through a two siage process The cosl of financing capital is minimized by trading off the tax
advantages of debt against the epected real bankrupcy coSls inherenl with high debVequity ratios
Given the optimal debtequity ratio the level of inveSlment Is chosen such that at Ihe margin the
24 l
bull
rerurn pn ~uity equas the return on bonds plus an exogenous risk premium
36 Market assumptions equilibrium and expectations
Virtually all of the papers arc characterized by Walrasian market clearing assumptions All
markets are perlect1y competi1ive Atomistic competition among agents 1s assumed even thoughmiddot only
a flnile number of agents is considered Virtually no mark~t disequilibria or price stickiness afe
considered The only exception is ErlichmiddotGinsburghmiddotHeyden (1987) bull In thei model of the Belgium
economy I the _wage rate is fixed in the short-run Therefore in the shOrlwrun desequilibrium in
the labor market will generate endogenous unemployment However in lh$ long-run all prices
Including the wage rata are Ilexible and accordingly ali markets ciear
Given the dynamic nature of behavlor in the economy markelmiddotcfearing prices in each period
depend on expectations of fture prices and on tax variables In the economy There are essentially
two ways at interpreting the economic equilibrium in such a dynamic conte~bull If future prices are
perfectly anticipated (io expectations are self-fulfilling) a perfect foresight equilibrium
prevaHs Then future actions are merely the implementation of current decisions for future
periods However jf ptica expecla110ns are not perfect (Le agents make mistakes w1th re$pect 10
future prices) then a temporary or shortrun equilibrium prevails Markets ciear and clearing
prices depend on fu1ure price expectations Current plans about the future are typically not
precisely implemented They will be revised as more or better information becomes available to the
economic agents See Grandmant (1982) for a comprehensive survey 01 the temporary equilibrium
literature
Wilh the exception of Gould (1985) and Pereira (19S6b 19S7d) all the models surveyed
adopt th~ concept of a perfect foresight equilibrium In turn BaliardmiddotGouldr (1965) considers a
J
2S
flexible am9unt of foresight In terms of the number of years over which price movements are ~
foreseen Pereiras model (1986b 1987b)) is flexible in a somewhat different way in that it can
indude any range of foresight from myopia to perfect foresight The choice between the perfct foresight and lemporary equilibrium is ultimately to be made on
philosophic grounds It can be argued that less than perfect expectations Imply that agents are
irralional in some way (see Auerbach-Kctiikoff (1987 p 10) However the reverse argument can
be made One can question wiether agents are reaHy rationar and perfectly knowledgable about
future prices
Recent evidene of Balird (1987) Ballard-Goulder (19851 Goulder (19851 and Pereira
(1986b 1987d) suggests that the choice in modemng expectations is an important one They show
that the degree of foresight inl0 thc future (ranging from perfect foresight to-myopic expectations)
may have dramatic impacts on the pollcy conclusions of Ihe model Accordingly the best research
strategy may b to design models which are flexible enough 10 allow for different rules regarding the
formation of expectations With the exception of the articles just mentioned sensitivity analysis
bullhave previously not been performed along this dimension
In terms of implementation the two concepts of equilibrium - perfect foresight and temporary
equilibrium have different implications The dimensionality of the equ1ibrium soll1ion algorIthm
is involved Suppose we have a model with ten markets to be run for a period of 50 years Aside
from normalization a perfect foresight model implies computing prices In 500 dimensions while a
-t~mporary equilibrium model requires sorving SO equilibria each in ten dimensions Given that
computational speed often varies with the cube of the number of dimensions the lemporary
equilibrium formulation is potentially strikingly more feasible However Ballard (1987) and
Goulder-Ballard (1985) have devoloped techniques to greatly speed the computation of a perfect
foresight equilibrium
26
j
I
4 Jrnplemenlalion and Issues on policy evaluallon
41 Calibration and equilibrium comparlslon
CGE models at typically parameterized by the use of a calibration procedure Some parameters
are exogenously given However some crucial parameters are determined in such a way that the
model replicates the data for a given base year See MansurmiddotWhalleY 11984) for an eXlensive
discussion of 1~i$ issue
CalibratIon in a dynamic context is generaliy interpreted as requiring two properties First
replication of base year data is required Second the model is parameterized to simulate an
intertemporal balanced growth path when Ihe base policy is maintaineq This Is the approach
followed by Ballard (1983) BallardmiddotGoulder (1985) Goulder (1985) and SummersmiddotGoulde
(1986) It follows the practice of BallardmiddotFuliertonmiddotShovenmiddotWhalley (1985) with their
sequential cquilibrium dynamics
Other authors Auerbach-Kotlikof (1983 1984 1987) Boyenberg (1984 1985 1986)
and Pereira (198Gb 1987d) follow whal we cali a qualitalive calibration The slructural
palametars are exogenously chosen so that lh economy follows a reasonable path into Ihe fulure
ThaI has 10 do wllh the fact thaI given the recursive nature 01 the dynamic economy and aside tom
such structural parameters only iniUa stock values are needed to run these models Given initial
conditions on the stocks of say private wealth capltal and government debt agents will optimize
and Ihereby generate a lirst round of net demands and equilibrium conditions In lurn the
equilibrium prices will determine the evolution of the stock variables into the next perIod
There are several potential problems with calibration in the conlext of dynamic models The
assumption of a steadymiddotstate growth ~ath in the base case can be questioned First while
27 I Ibullbull
steadymiddotstaU is a possibity it certainly is not the only meaningfull solution to dynamic models
Even in the case of a perfect foresight equilibrium the model implies an equilibrium path which i may or may not involve colanced growth The model not the modeller should dictate the nature 01 I themiddotbase case path Second in the context of a temporarY equilibrium path a steady state solutlcn is Inol a likely model QutcOn1tl In fact unlessmiddot expectations are stattc short run behavior consistent
with a steady-state evolution will in general not be generated On the other hand if static I expectations are self~fufjili[1g we have in fact a perfect foresight model Third even the base year
replication requirement may cause problems In Ihe cOntexl of temporary equilibrium Any
calibration parameter would be conditional on expectation rules which is probably an undesirable
feature
The nalur of the two-requirement calibration strategy is very muchin the spirit of the bull
traditional design of the comparision of alternative equilibria comparis_ion between a steady-state
base case on one hand and alternative paths including a transition period and a finat sleadyyenstate on
the other hand Pereira (198Gb 1987d) compare different (not necessarily steadystate)
equilibrium paths The arguments against such a procedure are based on the idea that the impact of
policy changes can be observed most easily since all departures from the steadymiddotstate can be
attributed to the altemalive polioy On the other hand the base case so dEfined as a steadymiddotstate is
consist~nl with previous work in a less dynamtc setting and Iherefore allows a common standard
for comparing model results
42 Computation algorithms
We are still al a stage in which basically each author uses a different computational technique
Th development of dynamic models - in parlicular wilh adjustment costs andior perfect foresight
bull
28
has corresponded with the decline in the use of fixed point algorithms In -fact given the relative
arge dimensions inevitably involved such algorithms lend to be very inefficient at the best and
chen prohibitively stow See Stone (1985) and Preckel (1965) for a comparative assessment of
different computation techniques Among the models surveyed only Ball1rdFulle(ton~
-ShavenmiddotWhalley (1985) and Feltensteln 1985 use Merrills variant of the fixed point
algorithm 1echnique
AuerbachmiddotKoHikoff (1963 1984 1967) follow a three stage procedure They first compute a
base case steady-state then a revised case steady state and finally transition path for the economy
berveen these two steady-states tn all stages a Gauss-Seidel iterative procedure is used
Ballard (1982) Ballard-Goulder (1965) Goulder (HiSS) and Summers-Goulder (1986)
~
use a method developed by Ballard and Goutder which is similar in many awects to the FairmiddotTaylQr bull
(1983) algorl1hm Short-run equilibria are calculated (using Merrills algorithm) parametric on
nrice expeC~lions The model is then iterated to generate self-fulfilling intertemporaf expectations
and the corresponding perfect foresight equilibrium In a relatively similar approach Andersson
1987 uses a simulation program SIMNON developed In the University of Lund This program
can handle two-paint boundary problems In a fashion consistent with the multiple shooting
algcrlthm (see Lipton-Poterba-Sachs-Eummers (1982))
Boyenbergs computational approach (1985 1985) differs from the other models surveyed in
that he relies heavily on analytical techniques Computations are done by using a dynamic version of
Johansons linearization method Being essentially determined by the continuous time nature of the
I I
f
model thiS linearization model has the disadvantage of confining the analysls to infinltesinal changes
around the base case equilibrium (see Bovenberg (1985) p 53) Pereira (1986b 1987d) uses an optimization algorithm NPSOL - developped by
rGillMurray-Saunders-Wrlght (1986)) This algorithm is used 10 compute the sequence of
I
29
Shortmiddotrun temporary equilibrium which makes up the intertemporal equilibrium path The
equilibrium conditions are seen as nonlinear equality constraints in the minimization of an artlcial
objective function The prices are normalized to the unit simplex by an additional linear equality I I 1 bull constraint
II ErlichmiddotGinsburgh-Heyden (1987) follow a unique approach in that they use a variant of the
optimization technique introduced by Negishl (1960) The economic equilibrium can be generated i
as a solution of a mathemallcal program Ihe objective function of which is a weighted sum of the
utility functions of the various agents while tho constraints set consists of the market dearing 1
conditions Ginsburgh-Heyden (1985) have extended Negishis resulllO tho case of downward price
lrigidities I
Finally the paper by Jorgonson-Yun (1984) is also unique in tIlat it is the only I
econometrically estimated model Different blocks for the consumption and production Side of the
-
model are separately estimated to provide the necessary structural parameters i
I The diversity of computation techniques is yet another indicator of the exploratory nature of the
body of literature surveyed in this article
43 Equal yield comparisions
The link between a base case and counteriactual simulations Is usually provided by the concept of
equal yield Shaven-Whaley (1977) discuss the meaning 01 equal yield in a general equilibrium
context when government is not allowed to run defICits Equal yield is interpreted to moan constant
public utility Government base case utility is maintained in the counferfactual experiments With
balanced budgets the concepl of equal yield is unambiguous The new equUibrium prices and the
balanced budge condition will del ermine the minimal expendilure and taxes needed to maintain base
30
case publi~ utility Accordingly in general equal yield is inconsistent with equal nominal tax
revenue Some change in tax revenue Is necessary Different tax replacement schemes ate
considered to assure that enough tax revenue is collected This is the approach essentially followed
by most 01 the papers surveyed Only Feltenstein (1985) Goulder (198) and Jorgenson-Yun
(1984) chose not to follow an equal yield strategy
The question is of how to Interprete lila concept of equal yield when the government is alowed to
fUn deflcLts The optimal leve of expenditure for base ease public trliUty can now be taAinanced
bond financed or financed by a mix of bonds and taxation We have several versions of equal yield In
particular equal yield may now be consistent with equal tax revenUE Furthermore some meaSure
ltgtf financial crowding out effects of government deficns can be inferred from the comparislon of the
several equal yield alternatives These issues are extensively discussed ~Pereira (1987b)
44 Price expectations and the dynamic generalization of compensation
Indicators
The sum of equivalent and compensation variations over households is the most widely used
aggregated measure of efficiency gains or losses In the context of $teady~state eomparisions andor
when complete future markets exist andlor perfect foresighl is assumed there are no difficulties
associated with the use of the standard Hicksian indicators In fact correct future prices are known
in these cases
If expectations are not selfmiddot fulfilling andlor future markels are not available a dynamic
generalization is necessary 8allardmiddotGoulder (1985) provide some steps in that direction by
defining an indicator that accomodates periods far into the future when households do not have
perfect foresight but a steady~state prevails On he other hand this issue is more fundamental in
i
[
I
i I I I I [
I I
bullI
31
the contex~ of a Gner1 lemporary equilibrium ftamawor~ In such Circumstances Pereira
(1986a) dev~lops a cnamlc generalization which is obtained as the present discounted value of a i i
sequence of short-fUn optimal expenditure functions consistent with a base case expected future I stream of IlWities I
32
5 Empirical evidenee from selected policy Issues
Dynamic tax models have generated several important results which escaped Sialic modellers
The following is a seleCpoundd set of issuos whIch stress the marginal benefits of dynam~ modelHng of
economic behavior in innovltlttive areas of analysis The discussion of a consumption tax emphasizes
the benefits of dynamic tusehold behavior and intergenerational aspects The study of the
flliminatlon or the re~intToduction of the investsnemt tax credits is made meaningful by the dynamic
modelling of production betlavor Modelling governmenl in a dynamic conlexts allows the modeller
to study the irrrpact of government deflcits Finally a dynamic framework lets us appreCiate the I
relevance of consumer expectations in terms of the evaluation of policy alternatives
i
i
51 Consumption Tax
t
By Its very nature a consumption tax can not be adequately investigated with a static model bull
Fullerion-Shoven-Whaicy (1983) use Ihe model of the US economy s described in
Ballard-Fullerton-Shoven-Whalley (1935) to evaluate the movement from the currenl US tax Isys1em to a progressive consumption tax Since their modeJ Incorporates a Jaborleisure choice
where leisure is an untaxed commodlty their results reflect the fact that both the consumption tax
and Ihe present system are dslortionary
Concentrating on the rntertemporaf distortions they show that sheltering more saving from lhe
current US income tax could improve econornic efficiency even if marginal tax rate increases are
necessary in order to maintain government fevenue~ At first you have a revenue shortfall
However the economy moves to a higher sustatned growth path and uitimatefy lower tax rates
cangenerate the same revenue path Also wages increase in the long run with this policy The
bull
bullbull ~
33 shyt
i
I
I
present value of welfare gi[lS for a potiey of complete savin9 deduction wilh marginal tate
adjustments (consumption lax) is simulated to be around $500 billion to $600 billion 01 1973
dollars
They also investiga1e th 1ngth of ttme it takes th~ economy to adjust to these policy changes
Roughly they estimate the long run to b thirty years although this figure Is very sensitive to the
specification of the savings elasticity - a crucial parameler in this model
52 Investment T~x Credit
Goulder-$ummers (1 S87) address the impact of eliminating the lnvestmenl Tax Cradit (ITC) on
intersectoral capital formation and on economic growth Their model isYa1ticularly adequate to
addies$ this issu~ in that is postulates a forward looking investment m9we with adjustment costs
They show that the eliminating the ITC causes a reduction in the rale of investment Inveslment is
estimated to fall by about 7 in the short-run and by about 12 in Ihe new long run steady state
In lurn a previous policy anncuncement lowers Ihe overall al1ractiveness of Investment and leads 10
a downward shift of the Investrrent profile On the other hand the combined effect of a revenue
neutral simultaneous efimiraron of the ITC and reduction of the corporate tax rates is a long run
reduction ef the capital Sieck by 35 This pattern suggests that a revenue neutral increament in
both Ihe corporale tax rates and the inveSlment lax credits would be preferable 10 the formula
adopted in Ihe US Tax Reform Act of 1986
Pereira (1987d) addresses the impact of the ITe from Ihe oland point of Ihe curTenllax syslem
under the Tax Reform Act of 1986 Given Ihe concern over the potential depressiVe effects upon
savings and investment of the new tax system in conjunction with deficit reduC1ion mechanisms of
the GrammmiddotRudman type a pertinent question is should Ih lTC be e-introduced In the context
bull
34
-of his mod~1 of the US economy Pereira focuses on the trade-off between the distortion in the
intersectoral allocation of capital induced by the lTC the lower tax revenues it generates and
potential financial crowding-out effects on one hand and the positive effects of lowering the relative
price of new capital goods on the other hand Preliminary results confirm the qualitative results in
Goulder-Summers (1987) suggestting a welfare gain of about 2 of the present value of GNP in
the best scenario of absence of replacement taxes Furthermore preliminary results show an
important time pattern to welfare gains with the average benefits increasing the further you look
into the future This pattern is due to the presence of constraints to intersectoral mobility of capital
and inertia in the adjustment towards the optimal capital levels as reflected by the presence of
adjustment costs
I
I
bullAuerbach-Katlikaff (1gB) in the context of their intergenerational model of the US economy
consider the impact on savings capital formation and interest rates of deficit policies and balanced
budget increases in government consumption They conclude that deficit finance and government
consumption can significa-ntly crowd out capital formation and lower the welfare of future
generations However crowding out from deficit finance is a very slow process because it results
from increased government spending over potentially long horizons Also deficit policies may
substantially influence the longmiddotrun interest rates while leaving the short-run rates essentially
unaffected Finally and in opposition to the central role adjustment costs seem to play in both
53 Financial crowding outeff~cts of government deficits
Goulder-Summers and Pereiras models Auerbach-Kotlikoff suggest that the time path of interest
rates induced by a policy of deficit financing seems to be insensitive to the presence of adjustment
costs
bull
35
54 The role 01 consumer expectatlons
The expectations analysis has uncovered one of the benefits of dynamic modelling
Ballard-Goulder (1985) find that the appeal of adopting a consumption tax depends on the level of
foresight possessed by the consumers Furthermore additional foresight may be welfare worsening
This is a second best type of result This tends 10 occur under policies that lead to capital deepening
and declining rate of return to capital over time To the extent that consumers have more foresight
they will be beller equippod to anticipate the fall in the rental price of capital Consequently if a
saving incentive is enacted people save more with myopia than with perfect foresight Given the
exi~tence of taxes on capita and the discrepancy between the private an~--sccial returns to capital
the greater savings with myopia is beller socially Ballard-Goulder find that the welfare gains
from a consumption tax are reduced by about 10 when we move from myopia to farleet foresight
Therefore the attractivenness of adopting a consumption tax seems fairly [obust across these
foresight specifications
I
bull
36
6 Tbe power and weaknesses 01 dynamic modelling the example 01 corporale
tax Integrallon
The Issue of corporate tax integration allows us to sh~w the stre-ngths and weaknesses or the
dynamic approach Empirical evidence using CGE techniques indicates that depending on the precise
scheme of integration and the 1ax replacement methods integratiOn may have substantial effects In
Ihe work of FuliertonmiddotKingmiddotShoven-Whalley (1980 1981) total integration was found to yield an bull
annual static eHielency gain of $4 to $8 billion in 1973 dollars Simulated dynamic gains may be as
large as $695 billion or abcu 14 of the present value of tulUre consumption and leisure in the
US economy These gains result primarily from interindustry reallocalions of Investment and an
improved inter~emporal allocation of consumption
Mora recent work emphasizes hOw consumers asset portfolio decisions and firms finanCial
decisions affect the efficiency gains from integration Slemrod (1980) focusing on consumers
asset ponfoio deCisions finds static efficiency gains which are about twice -as Iarge as those
repOrted iiy Fulierton-King-ShovenWhalley (1961) FuUrlon-Gurdon (1983) focus on the
firms financial decisions They roporl efficiency gains of 6 of GNP from the elimination of the
tax distor1ions favoring debt However when they eliminate the corporate tax and repiace it with
increaSed persona income taxes addlHonal dls10t1ions are created in the optimal labot~leisure
decisions These distortions tend to dominate the analysis slJ(h that the overall effects of complete
integration are very modest Galpr-LuckmiddotTodr (1986) address simultaneously consumers
asset pOrlfollo decisions and the firms financial decisions They also find very modest efficiency
gains from integralion
The above results are important but may be severely biased There are severa aspects of
economic behavior and modelling crucial for the study of Income tax Integration whiCh have no been
i 37
captured innny of the above papers One aspect is the absence of government deficitsmiddot a balanced
budget is assumed It is true Ihal resource crowding OUI is caplured In these mOdels but financial
crowding out induced by government spending is nOI Second investment is not derived from
optimtztlcn behavior Investment behavior passively accomodates endogenous saving decisions As
a consequence the differential impacts of policies in the incentives to save and to invest are not
captured Aso full capital mobility across secors is assumed with instantaneous capital
adjustmenlS towards optimal levels This assumption rules qut different costs of capital across
seclors and therefore differentiated reactions to tax policies changes Finallyhe modelling of both
gov9rnmerlt deficits and of endogenous real and financial investment decisions necessitates the
~nsideraiion of a dynamilt framework and the introduction of financiaJ assets govemment bonds and
iprivate financial assets A dynamic framework also highlights the efficiency effects of Integration on r
ithe ptimal intertemporal decisions The above models are either sIalic (Slemrods and
GalpermiddotltJckemiddotTode(sj or a dynamic sequ~nce of otherwise Slatic model bullbull I Pereir a (1986b) develops a dynamic applied general equilibrium model bull to study the tbull
efficiency effects of integration on the growth and allocation of investmenl across sectors Special
Iattention is paid to the structural affeelS of resource and financial crowding oul induced by
gove mect spending and deficits and to the real and financial Investment reactions across industries
to both tax integration and flnancial crowding out His mode departs from previous work on income
lax Integraticn and for that matler from most 01 the CGIE literalure for tax policy evaluation in
several direcllons It encompasses an endogenous sequential equilibrium struelure founded on
dynamic behavior with flexible expectations Government deficits are optlmaUy determined
Investment decisions are tward looking and the resull of optimizing behavior Several financial
assetsmiddot public and privatemiddot are considered
Simulation resulls Indicale that the welfare gains from Integration under large deficits are at
32
best lJlodest when measured in terms of the GNP The elimination of the corporate tax and its I replacemenl by increased income lax rates yields long run benefits which are never larger than
2 of the presenl yalue ct telr consumption and leisure onder the previous tax regime and 1
under the current tax regime However the short run w~lare effects of full integration tend 10 be
very low and in in some scenarios even negative This is a flew intertemporar pattern of efficiency Ieffects which reflects an adjJsiment lag in the interindustry investment decisions due to the
existence of costs of adjustment On the of her hand partial integration achieved by excluding I dividends from the corporale lax base syslematically yields negalive efficlencyeffeclS In Ihe long Inm these can be as high as 3 of present value of the future value of consumption and leisure This
is a new second best effect suggesllng Ihat less than complete integration may have perverse I efficiency effects
The dynamic slructure of PerclrBs model snowing for an improved freatment of real investment
declsions and government deficts provides a potential setting for a more accurate measure of the
costs and benefits of inlegration The simulation resullS suggest lower benefits and an intertempora Ibull
pattern of growing benefits
Simulation results also clearly suggest robust negative effects from partial integration How
reliable is this result If dJvidends were deductible from the corporate lax base (as are interest
payments) the preferenlial Ireatmant of debl over equily would be ellminaled Corporallons should
be expecled 10 react by decreasing the optimal deWeqully rallo However corporale financial
gecisions are exogenously set In Pereiras model as in aft of the other models surveyed that go as far
as dealing with the issue
Also withdrawing dividends from Ihe corporate tax base is a way of eliminating Ihe double
taxation of dividends at belh Ihe corporate and personnal Income levels How will the optimal
aliocellon of household saving accomodate that change in Ihe relalive return to corporale equity
39 I
t However hOusehold portfolio decisions are exogenouly set in Pereiras model as in all of the other
models surveyeltj go as far as dealing with the issue
IIt is sate to say that the evaluation 01 the benefits 01 partial Integration as defined above are
sevrey underestima1eO Meanwhile the distortionSoenerated by the increase in the marginal
personal tax rates designed to ralse the revenue foregone by the dividend exclusion are fully
aCCltlunted for A good measure of the costs of integration together With a poor evaluation of the
benefils may very well be the reason why partial integration is Simulated 10 yield neg alive
eHikiency gains
On a different vein as most of Ihe model surveyed here Pereiras is a closed economy model It
is legitimate to wonder how the resullS would change If international capital flows were anowed
IThis almost certainly would affect financial crowding out since a good d~al of the capital inflows
could be used to finance public debt
For these various reasons the results should be treated as preliminary but they strongly I suggosl that the dynamic structure provides valuable new insights into the corporate tax integration SSU2
40
7 Concluding remarks I I I i
What has been acccmplished Ihis far The success of Ihe CGE research in laelding the challenge
of dynamic modelling can not be denied
I iGreat progress in the modelling of household behavior has been achieved Promising
developm~nts in the modeUing of production and government behavior have also been accomplished
Interesting innovations in the concept of equilibrium and computation techniques were discussed
The policy analysis was greatly enriched by considering transitional effects together with longer
run steadymiddotstate equilibrium paths generalized equal yield strategies accomodating different
financial crowding out impacts and generalized dynamic policy evaluation indicators
An important set of issues have been addresed by the models surveyed Traditional issues
focusing on the effects of capital taxation and consumption taxes have been pursued In terms of
capital taxation for example the intertemporal nature of the issue was enhanced by allowing an
optimal evolution of capital stock in the economy and forward looking optimal investment decisions
In Andersson (1987) GouldermiddotSummers (1987) and Pereira (1986b 1987d) this is coupled
with a improved treatment of several tax provisions like investment tax credits and depreciation
allowances In terms of the consumption taxes developments in the specification of Intertemporal
household behavior the introduction of overlapping generations and the modelling of
intergenerational links via bequest motives as well as a closer attention to demographic evolution
provided a much improved economic setting for the understanding of the several aspects of the
problem
Furthermore dynamic modelling has permitted the addressing of a set of new issues Policy
Issues ranging from dynamic tax policy analysis to the evaluation of exogenous changes in
government expenditures to the impact of different methods of financing government expenditure to
4 1
the importance of tinanciai croHrii1g out to 1he relevance of consumorS expectations in tne policy
results have been addressed
Despile all of the progress and in part due to such progress several avenues are wide open for
much neded additional research The following seem to be the most Interesting and promising
areasmiddot First the specincation of liquidity constraints to household behavior may help to obtain a
better understanding of policy chenges that affect directly the interest rates in the economy Second
major attention needs to be paid to the incorporafion of a roraign sector (with open capital markets)
Third some efforts are needed in terms of finding adequate computation techniques now that
dynamic modelling eally makes the curse of dimensionality worse than eVOrbull
The modelling of financial markets is the single most unsatisfactory speet of the dynamic
models surveyed here~ The problems associated with provlding a mo~ adequate treatment of
financial markets and in general endogenous and optimal financial deci$ions bull both at the corporate
and at the personal levelsmiddot have 10 be addressed That necessarily requires Introducing uncertainty
into the models To b adequate this must go well beyond the mere assumption of point expeetalions
with the whole array of modelling implementalion and evaluation problems thai generates
bull I i
42
REFERENCES
Abel A and O Blanchard 1983 An Intertemporal Model of Saving and Interest Econometrica
51 pp 675-92
~ Andersson K 1987 Tar-ation of Capital in Sweden - A General Equilibrium Model Ph D
Dissenation UnIversity of Lund Sweder
Auerbach A L KotlikoH 1983 National Savings Economic Welfare and the Structure of
Taxation in M Feldstein ed Behavioral Simulation Methods in Tax Policy- Analysis Chicago
University of Chicago Press
Auerbach A l Kotlikoff 1984 Social Security and the Economics of Demographic-
Transition in H Aaron and G Burtless eds Retirement and Economic Behavior Washington DC
Brookings Institution
Auerbach A L Kotlikoff 1987 Dynamic Fiscal Policy Cambridge Cambridge University
Press
Auerbach A L Kotlikof and J Skinner 1983 The Efficiency Gains from Dynamic Tax
Reform Internatiooaj Economic Reyiew 24 pp 81middot100
Ballard C 1983 Evaluation of the Consumption Tax with Dynamic General Equilibrium
Models Ph D Dissertation Stanford University
Ballard C 1987 Tax Policy and Consumer Foresight A General Equilibrium Simulation
Study Economic Inquiry 25 pp 267-284
Ballard C and L Goulder 1985 Consumption Taxes ForeSight and Welfare a Computatonal
General Equilibrium Analysis in J Piggott and J Whalley eds New Deyelooments in Apolied
General Equilibrium Analysis Cambridge Cambridge University Press
Ballard C D Fullerton J Shoven and J Whalley 1985 A General Equilibrium Model for
t 43
Tax Eoliey Eyalua~iQO NBER University or Chicago Press
Borges A (1987) bull A Survey of Compula1ional General Equilibnum Models CECD Economic
Slud~ 8
Bovenbig L 19B3 imperfect Capital Mobility in a Perfect Foresight Herberger Model
Mimea UC Br~eley
Bovenberg L 1984 Capitel Accumulation and Capital Immobility Qmiddottheory In a Dynamic General Equilibrium Framework PhD dissertation University of California at Berkeley t
Bovenberg L 19850 Dynamic General Equilibrium Tax Models with Adjustment Costs in A iManne ed bull 1985 EtQnomrc ECUjilbrium Model Formulation and Solution Mathematical
lProgramming Siudy 23 Arnsterdam NorthmiddotHoliand
Bovenberg L t 1986 Capital Income Taxation in Growing Open Eeondmies Journa of Pubtic
Eamp(omlcs 3t pp 347middot376 IBovenberg L ald W Keller 1981 Dynamics and Nonlinearlties in Applied General
Equilibrium Models Internal Report Mlmiddot81middot208 Department for Statistical Methods CBS bull
Vooiburg The Netherlands
Charnley C 1981 The Welfare Cost of C~pltal income Taxation In a Growing Economy
IJournal of Political ECQnQm~ 89 3 pp 468middot496 I
Charnley C 19810 Efficient Tax Reform in a Dynamic Model of General Equilibrium I Mimeo Yale University I Charnley C 1982 On the Optimal Taxation In Stylized Models 01 Inlartemporal General
Equilibrium Mimeo Yale University I Decaluw B and A Manens 1985 Pays en Oeveloppement at Modeles Calculables OEquilibre
General - Une Revue de al litterature empirique Monographie Sene 9 Volume 3 Centre de
Recherche et Developpemont Economlque Unrslt de Montreal
44
bull
I I
I
I
1
Erlich $ V Ginsburgh and L van der Heyden 1987 Wher~ do Real Wage Policies Lead
Belgium bull A General Equilibrium Analysis European Economic Reyiew (forthcoming)
air R and J Taylor 1983 Solution and Maximum Likelihood Estimation of Dynamic
Nonlinear Ratioant Expeclations Models Ecooometrica 51 pp 1169middot1186
Feltenstein A 1984 Money and Bonds In a Disaggregated Ec~nonYin Searl 1-1 J Shoven
Eds Aoolied ~eoeral EqJiliDCiulD tioalYsis Cambridge University Press Cambridge
Feltenstein A 19B6 A Dynamic General Equllibrlu Analysis of Financial Crowding Out
Theory with an Application to Australia Journal of Public Economics 31 No1
Fullerlon D R Gorden 1983 A Reexamination of Tax Distorllons in General Equilibrium
Models in M Feldstein Ed Behayora Simulation Methods in Tax pOlicy Analysis Chicago
University of Chipgo Press
Fvllerlon D J Shaven and J Whalley 1983 Replacing the US Income Tax with a
Progressive Consumption Tax JoulOal of Public ECQnomics 20 pp3middot23
Fullerton D Y Henderson J Shaven 1984 A COmparision of Methodologies in Empirical
General Equilibrium Models of Taxation in Searl H J Shoven Eds ~pQlied Geoaral Jguilibrium
lIoalyss Cambridge University Press Cambridge
Fullerton 0 A King J Shaven and J Whaney 1981 Corporate Tax Integration in the United
Statesmiddot a General Equiiibdufi1 Approach American Economic Review 71 pp~ 677~691
Galper H R Lucke E Tader 1985 Taxation portlor Choice and the Allocation of Capital A
General Equilibrium Approach Mimeo
Gill p W Murray M Saunders M Wright 1986 Users Guide for NPSOL (Version 40) A
Fortran Package for Nonlinear Programming T echnicsl Report SOL 86middot2 Deparlment of
Operations Research Stanford University
Glnsbur~h V and L van der Heyden 1985 General Equilibrium with Wage Rigidities An
45
AppJication to Belgium in A Manne ed t Economic EQumprit1m~ Made Formulation and $oJioo
Mathematical Programming Study 23 Amsterdam NorthmiddotHolland ~
Goulet J 1968 1 Adjustrent CO$1S in the Theory o(the Firm Reyiew of EconomiC Studies 35 bull
pp47middot55
Goulder L 1985 Tax-Financed versus BondmiddotFinancedGhanges in Governent Spending
bull Mimeo Harva(d UniversilYmiddot
Goulder l J Shaven and J Whalley 1983 Domestic Tax Policy and the Foreign Sector The
importance of altemative foreign policy formulations to results from a general equilibrium tax
analysis model In ~n methods in Ibe antico of income from caoia1 ed Martin
Feldstein Chicago University of Chicago riSS
Goulder l L Summers 1987 Tax Policy Asset Prices and 3rowtly ~
A General Equilibrium Analysis NBER Working Pper No 212B
Grandmont J 1982 Temporary General Equilibrium Theory Ch 20 in Handbook 01
Mathemalical Eccnornics Vol II NorthmiddotHoliand
Harberger A 1959 The Corporate Income Tax An empirical Appraisal In lax Revision
Comoendjum Voll House Cornmitee on Ways and Means Washington DC Govemment Printing
Office
Harberger A 1962 The Incidence of the Corporate Income Tax JOULn1 of Political
Economy 70 pp 215middot240
Harberger A 1964 Taxation Resource Aliocalrlon and Welfare fn lb Role of Dir bullbulll and
Indjresl Taxes io tbe Fede[al Reserve s~I~Ill Princeton Princeton Unlveresity Press
James J 1985 New Developments in the Applications of General Equilibrium models to
Economic History in Piggott aand Whalley Eds
Jorgenson D and K Yun 1984 Tax Policy and Capital Allocation Harvard Inslitute of
bull
I
I f
Economic Research WP 1107
Judd K 1965 ShonmiddotRun Analysis of Fiscal Policy in a Simple Perfect Foresight Model
Journal of poJiHcal 5CCJiW1l pp 298middot319 r
Upton D J Potbc J Sachsa nd L Summers 1982 Multiple Shooting in Rational
Expectations Models fuoometrlc 50 1329middot1333
Manne A ed 1985 fconQrnjc Eoullibrium Model FOrIDujaljQo and Solution Mathematical
Pogrammlng Study 23 Amsterdam NorthmiddotHolland
Manne A 1985 Or the Formulation and Solution 01 Economic Equilibrium Models In A
Manne ed ECQOQlli-Jr~ibmiddotum ~tQdel EQlJulailon and Solution Malhematical Programming
Study 23 Amsterdam NorthmiddotHolland
Mansur A and J Whalley 1984 Numerical Specification of Applied General Equilibrium
Models in Scarf H J Shoven Eds Aoplied General Eouillbrium AnaiysectIsect Cambridge Universily
Press Cambridge
Merrill 0 1972 Applications and Extensions of an Algorithm thaI Computes FixedmiddotPoints of
Cenain Upper Semimiddotcominuou Point to Set Mappings PhD Disserlation Unlvel$ity 01 Michigan
Negishi T 1960 Welfar Economics and Exislence I an Equnlbrlum for a Competitive
Economy MelroeCOOQ11ka 12 92middot97
Pereira A 1985 On the Existence and Uniqueness of Optimal Ou1put Path lor CRTS Firms in
Adjustment Costs Technologies Mimeo Stanford University
Pereira A 1985a Consumer Expectations and Ihe COSI 01 UllUty In an Intertemporat
Framework Mlmeo Stanford University
Pereira A 1985b Corporale Tax Integrallon and Government Deficits A Dynamic General
Equilibrium Anaiysls Mlmeo Stanford University
Pereir A 1987 On Ihe Computation of the Users Costs of Stocks Ecooomi (forthcoming)
47
Pereira A bull 1987b Equal Yield Alternatives and Government Oeficlts Mimeof Stanford
University
Pereira A 1987c A Dynamic Applied General ~uHlbrium Model for Tax Policy Evaluation
Comlletamp Docurnent2tion Mimeo University of California San Diego (in progress)
Pereira A 1987d Should the investment Tax Credit be Rei(ltroduced Mlmeo University
of California San Diego (ir progreSs)
Piggott J and J Whalley eds 1985 tJew QeyeQpments in Applied General EQuilibrium
bullenalysjs Cambridge Cambridge University Press
Preckel Pbull 1985 Alternative algorithms for computing economic equilibria In A Manne Ed
Radnor R 19amp3 Equllibrium under Uncertainty In K Arrow M Intrilligaor ods
I ibull
Robinson S 1 g86 Mullisectorai Models of Oeveloping CountrieS a Survey Oivision of
Agricullure and Resources UC Borkeley WP 401
Scarf H 1967 The approximation of fixed points of a continuous mapping ~MM Jouwal of
epplid Mathornallcs 15 pp 1328middot43
Scarf H with T Hansen 1973 00 the Computation of EconomiC Eguilibria New Haven Yale
University Press
Scarf H J Shoven Eds 1984 61l1llied Genera Elujlibrium 6nalyssect Cambridge University
~ Press Cambridge
Shoven J 1976 The Incidence and Efficiency Effects of Taxe on Income from Capital
Jurnal of ramie1 Economy 84 pp 1261-83
Shoven J 1983 Applied General Equilibrium Tax Modemng IMF Slaff PaoerS 30
Shoven J and LB Simon 1986 Share Repurchases and Acquisitions An Analysis of which
Firms Participate NBER Working Paper No 2243
bull
48
IShaven J and J Whalley 1972 ft General Equmbrium Calculation of lhe Effects of
Differential Taxation of Incoma from Capital in the US Jauroe of public Econgmics 1
pp281middot321 (
Shoven J bull and J Whalley 1973 General Equilibrium with Tax$s a computation procedure
and an existence pr00f ReviEhi of Economic Studies 40 pp 475middot490
Shoven J J Whalley 1977 Equal Yield Tax Alternatives A General Equilibrium
Computa~ui0nal Tfchnique J~HJrnal of PUblic EconomIcs 8 pp 211~224 bull
Shaven J J Whalley 1984 Applied GeneralmiddotEquilibrlum MOdels of Taxalion and
International Tr8de An Introduction and Survey Journal of Economic literature 23t pp
1007middot1051
Slemrod J 1980 A General Equilibrium Model of Capital Income Taxation Ph D bullDissertation Harvard Univnrsity
Slemrod J 1983 A General Equilibrium Model of Taxation with Endogenous Financial
Behavior in M Feldstein edbull Behavioral Simulation Methods In Tax Policy Analysis Chicago
UniVerSity of Chicago Press
Stone J 1985 Sequential Optimization and Complementarily Techniques for computing
Economic equilibrium in A Manne Ed
Summers L 1981 Taxation and Corporate Investment a q-Theory Approach BroQkinas
Summers L 1985 Taxation and the Size and ComposHion 01 the Capital Stock An Asset Price
Approach Mlmeo Harvard University
Whalley J 1975 A Geneal Equiloibrium Assessment of the 1973 United Kingdom tax
reform Economica 42 pp 139middot61
1JLTIllOS JIQll~)ltG PAPERS PUBLICADOS bull
I
I I
j
112 70 - LUCENA iJiCIIJo ~ IIEnvironrilent Honitoring and Orqanization Structure T (Setellibro 1987)
nI 71 - L~ENA Diogo A Note on the Representation of Information Stxueture (SeteJObro 1987)
nQ 72 - LUCENA D1CgO Envuorment 111)nitoring and Organization
Structure II (Setelilhro 1987)
ril 73 - AlmmE3 Antonio Pais e CAsPAR Yitor bull Tributacao Inoentivos e InvesUlllentos Analise Qualitativa noYellibro 19S7)
n2 74 - liAlltOSA rose Pedro Opt1lllal lIage R1gidity A SUggestedtlthodology to Teet tb Theory and an Application (Outubro 1987)
nI 75 - CABRAL L 11 B Three lIotes on SYJIIlletric Gemes witt itsYJIIlletric Equilibri (NoveJObro 1987)
n2 76 - LUCENlt D10gOTo Search or Not to Search (Fevereiro1967)
nQ 17 - SA Jorge asconcelos euroI A Theory of SynergyB (Fevereiro1987)
n2 76 - VILARES tanuei Jose Os Be IntenaMios IlIIportados ColIIO Factor de Produao (Julho 1987)
nQ 79 - SA J~rge VasCQncel03 e HoY to Copete and Co~unicate in tature Industrial Products_ (aio 1988)
n2 80 - ROIl Rafael Learning and Capacity Expansion in a New larket Under Uncertainty (Fevereiro 1988)
n2 81 PEREIRA Alfredo llarvilo Survey of Dynamic Computational General EquilibriUlll Hodel for Tax Policy Evaluation (Outubro 1987)
Qualquer in ormacao sabre o 1IIorlltlng Papers ja publlcados sera prestada pelo Secretariado de Apoio aos Docentes podendo os ~e3OSf
ser adquirido5 ne Secyo de Vlndas da Faculdade de EconollLia UNL na Travessa Esteviio Pinto Cltpolide - 1000 LISllOA
j
1 6
dynamics $e dimension of the problems is immensly increased Let IS be more specific
In a static framework with constant returns to scale production technology the output level ismiddot
indeterminate The optimal aliocation of inputs can be obtained by cost minimization with any
feasible output level generatillg zero profits Such zero profit conditions are used to solve for the
output prices in terms of the factor prices and hencemiddot to reduce the dlmensionality of the problem
from the -number of commodities and factors to the number of factors Thus even if the model deals
with 30 production sectors the computation of an economic equilibrium can take place using only
the dimensionality of the primary inputs in the economy
Now under certain regularity conditions on the production and adjustment costs technologies
leading to enough concavity of the optimality objective the intertemporal output path for the firm is
endogenously optimally and uniquely determined even with constant relums to scale technologies
(see Pereira (1985 1987a) Accordin~y with adjustment costs in the model optimal profits
will in general be non-zero This result has crucial implications for the computatiQn of equilibrium
in the model~ with production dynamics With adjustment costs no reduction of dimensionality is
possible The curse of dimensionality returns as a binding constraint
The introduction of adjustment costs adds another significant complication to the model With
capital being less that perfectly mobile in the economy different rates of return on capital will
exist in different sectors This is a difficult problem to tackle conceptually in the absence of
uncertainty Also in the real world producer maximization choices include also its choice of
financial ratios (debVequity) and payout rates (dividendretained earnings) These financial
subjects are difficult and much Clf this behavior is still taken as exogenous by the modellers We
will come back to these issues below
middot i
19
bull 33 government behavior
middot
Two issues dominate the modelling of government behavIor First govemment behavior has
Ibeen typically seen in Ihe CGE liIerature for lax poliCY evaluation as constrained by yearly balanced
I budsets (see for example 8allard-Fullrton-S~ovn-Whafley 098Sraquo The analysis of
govemment defidlS and public debl in a CelE context requires a dynamic selling Second the level of
government expenditures is either exegenouty given as in Auerbach-KoUikofi (1984 1987)
6ovenbrg (1984 1985) Feltenstein (1984 1986) and Jorgenson-Yun (1984) or
endogenously (but no optimally) determined by the balanced budget conditions as in
6allard-Fullerton-Shoven-Whalley (1985) Andersson (1987) Erllch-Ginshurgh-Heyden
~
(1987) Gouder (19aS) and Gouder-Summers (1987) In the second case the composition of
bull public expenditures is often optimally determined However the leve~ of government expendi1ures
can only 00 endogenously and optimally determined if the government is seen as an optimizing agent
and is allowed to run deficits
The first attempts to deal wiU the government defiCits in CGE tax models are due to
Auerbach-Kollilltoff (1984) and Feensain (1984) In Auerbach-Kotlikof (1984) government
expenditures afe exogenous They grow at the rate of growth of population However given 1he tax
structure and tax revenues yearly deficits and surpluses are atTowed subject to an in1ertempond
constraint that the present value of future tax revenues equals the present value of future
expenditures
In Fellenstein (1984 1986) government expenditures are also exogenously given He also
allows the government to run deficits to finance expenditures in excess to 1ax revenues However
surpluses are returned to CQnsumars in the form cf transfers Accordingly government ts not
subject to any constraint regarding the future repayment of public debt
In GouJder (19851 the 9Vernmenl maximizes a static social welfare ~nctio subject to a-
bull
20 I i
I
I I I
I bull
balanced budget conSirtlL This follows the optimal alocatkm of government expenditures alpng
the lines of Baliard-FHeiton-Shoven-Whalley (1985) Ttle model Is generallzea to allow for
exogenous changes in HIe time path of government expenditures Two financing alternatives are
consjdeft~d and contr8s1ed EldditionaJ tax revenues and bond issuance
Pereira (198Gb 1137d) auempts to address both the incorporation of deficits and the
determination of govement expenditures The path of government expenditures and the path of
delicilSlsurpluses (ano therefore the path lor debt) are endogenously and optimally determined The
government is seen as mltJximizing n intertemporal social welfare function given the 13x structure
Optimization is subject to a sequence of recursive equations of motion reflecting the evolution of the
ptlbllc debt allowing fDr government budget imbalances It should be stressathat this specification
01 the government cOMtcalnt is equivalent to the specificaUon in Auerbaeh-Kotlikoff (1983 1987)
when no liquidity constraints affect -gove~nment behavior
The extra richness in the treatment of government behavior allows tfte examination of
government debt poices in a truly dynamic setting Also financial crowding out effects induced by
government deficits can be analyzed (see Auerbach-Kottlikoff (1987) Feltenstein (1986) and
Perer (198Gb 1987e))
34 Foreign sector
Most of the open economy models follow the assumptions of balanced trade with import and
export net demands characterized by constant elasticities along the line of Ballardmiddot
-Fullerton-Shaven-Whalley (1965) Such is the case of Salfard (1963) Ballard-Goulder
(19S5) and Goulder-Sllmmers (1987)_ In Feltenstin (1985) the rest of the world is treated as
21
an addjlion~ consumer group
Bovcoberg (1986) develops a model In which two economies lIr considered each following
intertempoml perlect foresight paths These economies meet in the international forum Their
trade relaroMhips are characterized by yearly balanced trade accounts
None of the new generation of dynamic cae tax models has yet IncorPorated the lnternational
capital flows as done in aoulder-Shoven-Whalley (1983) for the earlier Ballarrshy
-Fullertonmiddot Shaven-Whalley (1985) model This is unfortunate as there is an important
Intertmporal aspect to international lending The first al1empts along Ihese lines are due to
Andersson (1987) and Erlich-Glnsburgh-Heyden (1987) Andersson (1987) in his model of the
Swedish economy adopts a closed economy appraoch in which rates of return in the domesJic
economy are largely determined by the international capital markets Fidinterest rate induce
Intlnatlonal capital flows which determine and finance the International lrade imbalance In
lurn In ErlichGinsburghHeydenmiddot (1987) foreign trade Is generaled according to an
intertemporal Irade welfare lunction with constant import and export elasticities In the shortmiddotrun
they allow inlernational trade inbalances which generate capilal ftows to the domestic households In
the long run however trade balance is assumed
35 The need for financial markets
A dynamic economic Wucture not only provides the ideal environment 10 model many features of
economic behavior it also permits one to Incorporate the financial side of the economy If the
government is allowed to run deficits the question 01 deficll financing automatically follows If
investment is optimally dewmined and returns to capital are different aClOSS sectors problems of
investment financing arise If there are seeral final1cfal assets In the economymiddot government
4-
I i
ibull
I I I I
I
j
22
bonds privpte bonds and equity (or simply physical capital installed inmiddot different sectors) the
problem of allocation of saving among assets with potentially different returns arises
The papers surveyed var greatly with respect to l~e extent of their attention to the financial
side of the economy At Clne extreme are the models in ~anardmiddotFuliertonmiddotShoven-Whalley (1985)
Andersson (1987) Ballltld (1983) Ballard-Goulder (1985) Bovenberg (1985 1986) and
Erlich-Gjnsburgh-Heyden C19B which are devoted exclusively to the real side of the economy
In turn Auerbach-KotliKoff (1984 1987) Feltenstein (1984 1986) and Goulde (1985)
allow for government debt In these models saving finances changes in government debt and physical
capital Private and publlc 2ssets are perceived by the households as perfect substitutes The
allocation of saving merely adjusts to the relative demands for funds
Feltenstein (1984 1986) is the only model surveyed which introduces money Government deficits are financed by issuing money and bonds according to an exogenously given rule Money is
demanded by consumers for transaction motives and an exogenously given fraction as a store of value
On the other hand government bonds and physical capital are the vehicles for the intertemporal
transfer of wealth
Summers (1985) and GOlldermiddotSummers (1987) introduce a whole menu of financial assets shy
firm specific equity capital Different assets earn different rates of retum However such rates
are equal up to constant and exogenous sector speCific risk premia Therefore the introduction of
constant exogenous risk premia as helpful as it may be in the context of calibration does not solve
the main issue of the non-optimality of the allocation of saving Also talking about risk premia in a
deterministic context is somewhat unsatisfactory
Pereira (1986b 1987d) also introduces a whole menu of assets private and public bonds and
firm specific equity However all the assets are expected to yield the same rate of return and
therefore perc~jved as perfect substitutes The different asset types allow consideration of
23
exogenOU$~8tUequjty triO dividendlrc~enlion rufes and therefore seve-ral sources of investment
financing bond equity and retained earnings
The nonmiddotcptimalHy of the allocation of saving and the absence or exogenelty of corporate
financial rules is _a reftectlon of the limitations of the determInistic approach Either in a context of
perfect anticipation oi the future prices Of in general within the t$atm of point price
expectations all the papers follow a deterministic approach Under such circumstances consumers
either expect diHerent rales of return (inclusive of risk premium) across assets in which case
they wilt buy only one asset (that with highest rate) Of they expect equal rat~s of relurn in which
case Ihey are indifferent about the asset composition of Iheir portfoliO There is no way of
Iradingmiddotoff rales of return and risks to obtain an optimal interiof solution to the problem of the
allocation of saving
The most advanced contribution in the modelling of saving allocation in a CGE setting is due to
Slemrod (1980 1983) in the context of a static one-period model In his model consumers act
according to a two stage separable decision process They Hrsl decide on how much to save Then
they decide on the allocation of saving according to an Indirect uttlity function dependent on the rates
of ratum and variances offered by the different assets in the eoonomy The sourCe of riskiness in the
ampConory comes from an uncertain marginal product of capital On the other hand aside from
portfolio decisions the res of the economy is insulated from uncertainty
The most complete contribution in terms of the treatment of the corporate financial rules is
FulienonmiddotGordon (1983) In a variant 01 Ihe BailardmiddotFulienon-ShovenmiddotWhelley (1985) model
Fullerton and Gordon have capital intensity and optimal financial decisions Jointly deter~ined
through a two siage process The cosl of financing capital is minimized by trading off the tax
advantages of debt against the epected real bankrupcy coSls inherenl with high debVequity ratios
Given the optimal debtequity ratio the level of inveSlment Is chosen such that at Ihe margin the
24 l
bull
rerurn pn ~uity equas the return on bonds plus an exogenous risk premium
36 Market assumptions equilibrium and expectations
Virtually all of the papers arc characterized by Walrasian market clearing assumptions All
markets are perlect1y competi1ive Atomistic competition among agents 1s assumed even thoughmiddot only
a flnile number of agents is considered Virtually no mark~t disequilibria or price stickiness afe
considered The only exception is ErlichmiddotGinsburghmiddotHeyden (1987) bull In thei model of the Belgium
economy I the _wage rate is fixed in the short-run Therefore in the shOrlwrun desequilibrium in
the labor market will generate endogenous unemployment However in lh$ long-run all prices
Including the wage rata are Ilexible and accordingly ali markets ciear
Given the dynamic nature of behavlor in the economy markelmiddotcfearing prices in each period
depend on expectations of fture prices and on tax variables In the economy There are essentially
two ways at interpreting the economic equilibrium in such a dynamic conte~bull If future prices are
perfectly anticipated (io expectations are self-fulfilling) a perfect foresight equilibrium
prevaHs Then future actions are merely the implementation of current decisions for future
periods However jf ptica expecla110ns are not perfect (Le agents make mistakes w1th re$pect 10
future prices) then a temporary or shortrun equilibrium prevails Markets ciear and clearing
prices depend on fu1ure price expectations Current plans about the future are typically not
precisely implemented They will be revised as more or better information becomes available to the
economic agents See Grandmant (1982) for a comprehensive survey 01 the temporary equilibrium
literature
Wilh the exception of Gould (1985) and Pereira (19S6b 19S7d) all the models surveyed
adopt th~ concept of a perfect foresight equilibrium In turn BaliardmiddotGouldr (1965) considers a
J
2S
flexible am9unt of foresight In terms of the number of years over which price movements are ~
foreseen Pereiras model (1986b 1987b)) is flexible in a somewhat different way in that it can
indude any range of foresight from myopia to perfect foresight The choice between the perfct foresight and lemporary equilibrium is ultimately to be made on
philosophic grounds It can be argued that less than perfect expectations Imply that agents are
irralional in some way (see Auerbach-Kctiikoff (1987 p 10) However the reverse argument can
be made One can question wiether agents are reaHy rationar and perfectly knowledgable about
future prices
Recent evidene of Balird (1987) Ballard-Goulder (19851 Goulder (19851 and Pereira
(1986b 1987d) suggests that the choice in modemng expectations is an important one They show
that the degree of foresight inl0 thc future (ranging from perfect foresight to-myopic expectations)
may have dramatic impacts on the pollcy conclusions of Ihe model Accordingly the best research
strategy may b to design models which are flexible enough 10 allow for different rules regarding the
formation of expectations With the exception of the articles just mentioned sensitivity analysis
bullhave previously not been performed along this dimension
In terms of implementation the two concepts of equilibrium - perfect foresight and temporary
equilibrium have different implications The dimensionality of the equ1ibrium soll1ion algorIthm
is involved Suppose we have a model with ten markets to be run for a period of 50 years Aside
from normalization a perfect foresight model implies computing prices In 500 dimensions while a
-t~mporary equilibrium model requires sorving SO equilibria each in ten dimensions Given that
computational speed often varies with the cube of the number of dimensions the lemporary
equilibrium formulation is potentially strikingly more feasible However Ballard (1987) and
Goulder-Ballard (1985) have devoloped techniques to greatly speed the computation of a perfect
foresight equilibrium
26
j
I
4 Jrnplemenlalion and Issues on policy evaluallon
41 Calibration and equilibrium comparlslon
CGE models at typically parameterized by the use of a calibration procedure Some parameters
are exogenously given However some crucial parameters are determined in such a way that the
model replicates the data for a given base year See MansurmiddotWhalleY 11984) for an eXlensive
discussion of 1~i$ issue
CalibratIon in a dynamic context is generaliy interpreted as requiring two properties First
replication of base year data is required Second the model is parameterized to simulate an
intertemporal balanced growth path when Ihe base policy is maintaineq This Is the approach
followed by Ballard (1983) BallardmiddotGoulder (1985) Goulder (1985) and SummersmiddotGoulde
(1986) It follows the practice of BallardmiddotFuliertonmiddotShovenmiddotWhalley (1985) with their
sequential cquilibrium dynamics
Other authors Auerbach-Kotlikof (1983 1984 1987) Boyenberg (1984 1985 1986)
and Pereira (198Gb 1987d) follow whal we cali a qualitalive calibration The slructural
palametars are exogenously chosen so that lh economy follows a reasonable path into Ihe fulure
ThaI has 10 do wllh the fact thaI given the recursive nature 01 the dynamic economy and aside tom
such structural parameters only iniUa stock values are needed to run these models Given initial
conditions on the stocks of say private wealth capltal and government debt agents will optimize
and Ihereby generate a lirst round of net demands and equilibrium conditions In lurn the
equilibrium prices will determine the evolution of the stock variables into the next perIod
There are several potential problems with calibration in the conlext of dynamic models The
assumption of a steadymiddotstate growth ~ath in the base case can be questioned First while
27 I Ibullbull
steadymiddotstaU is a possibity it certainly is not the only meaningfull solution to dynamic models
Even in the case of a perfect foresight equilibrium the model implies an equilibrium path which i may or may not involve colanced growth The model not the modeller should dictate the nature 01 I themiddotbase case path Second in the context of a temporarY equilibrium path a steady state solutlcn is Inol a likely model QutcOn1tl In fact unlessmiddot expectations are stattc short run behavior consistent
with a steady-state evolution will in general not be generated On the other hand if static I expectations are self~fufjili[1g we have in fact a perfect foresight model Third even the base year
replication requirement may cause problems In Ihe cOntexl of temporary equilibrium Any
calibration parameter would be conditional on expectation rules which is probably an undesirable
feature
The nalur of the two-requirement calibration strategy is very muchin the spirit of the bull
traditional design of the comparision of alternative equilibria comparis_ion between a steady-state
base case on one hand and alternative paths including a transition period and a finat sleadyyenstate on
the other hand Pereira (198Gb 1987d) compare different (not necessarily steadystate)
equilibrium paths The arguments against such a procedure are based on the idea that the impact of
policy changes can be observed most easily since all departures from the steadymiddotstate can be
attributed to the altemalive polioy On the other hand the base case so dEfined as a steadymiddotstate is
consist~nl with previous work in a less dynamtc setting and Iherefore allows a common standard
for comparing model results
42 Computation algorithms
We are still al a stage in which basically each author uses a different computational technique
Th development of dynamic models - in parlicular wilh adjustment costs andior perfect foresight
bull
28
has corresponded with the decline in the use of fixed point algorithms In -fact given the relative
arge dimensions inevitably involved such algorithms lend to be very inefficient at the best and
chen prohibitively stow See Stone (1985) and Preckel (1965) for a comparative assessment of
different computation techniques Among the models surveyed only Ball1rdFulle(ton~
-ShavenmiddotWhalley (1985) and Feltensteln 1985 use Merrills variant of the fixed point
algorithm 1echnique
AuerbachmiddotKoHikoff (1963 1984 1967) follow a three stage procedure They first compute a
base case steady-state then a revised case steady state and finally transition path for the economy
berveen these two steady-states tn all stages a Gauss-Seidel iterative procedure is used
Ballard (1982) Ballard-Goulder (1965) Goulder (HiSS) and Summers-Goulder (1986)
~
use a method developed by Ballard and Goutder which is similar in many awects to the FairmiddotTaylQr bull
(1983) algorl1hm Short-run equilibria are calculated (using Merrills algorithm) parametric on
nrice expeC~lions The model is then iterated to generate self-fulfilling intertemporaf expectations
and the corresponding perfect foresight equilibrium In a relatively similar approach Andersson
1987 uses a simulation program SIMNON developed In the University of Lund This program
can handle two-paint boundary problems In a fashion consistent with the multiple shooting
algcrlthm (see Lipton-Poterba-Sachs-Eummers (1982))
Boyenbergs computational approach (1985 1985) differs from the other models surveyed in
that he relies heavily on analytical techniques Computations are done by using a dynamic version of
Johansons linearization method Being essentially determined by the continuous time nature of the
I I
f
model thiS linearization model has the disadvantage of confining the analysls to infinltesinal changes
around the base case equilibrium (see Bovenberg (1985) p 53) Pereira (1986b 1987d) uses an optimization algorithm NPSOL - developped by
rGillMurray-Saunders-Wrlght (1986)) This algorithm is used 10 compute the sequence of
I
29
Shortmiddotrun temporary equilibrium which makes up the intertemporal equilibrium path The
equilibrium conditions are seen as nonlinear equality constraints in the minimization of an artlcial
objective function The prices are normalized to the unit simplex by an additional linear equality I I 1 bull constraint
II ErlichmiddotGinsburgh-Heyden (1987) follow a unique approach in that they use a variant of the
optimization technique introduced by Negishl (1960) The economic equilibrium can be generated i
as a solution of a mathemallcal program Ihe objective function of which is a weighted sum of the
utility functions of the various agents while tho constraints set consists of the market dearing 1
conditions Ginsburgh-Heyden (1985) have extended Negishis resulllO tho case of downward price
lrigidities I
Finally the paper by Jorgonson-Yun (1984) is also unique in tIlat it is the only I
econometrically estimated model Different blocks for the consumption and production Side of the
-
model are separately estimated to provide the necessary structural parameters i
I The diversity of computation techniques is yet another indicator of the exploratory nature of the
body of literature surveyed in this article
43 Equal yield comparisions
The link between a base case and counteriactual simulations Is usually provided by the concept of
equal yield Shaven-Whaley (1977) discuss the meaning 01 equal yield in a general equilibrium
context when government is not allowed to run defICits Equal yield is interpreted to moan constant
public utility Government base case utility is maintained in the counferfactual experiments With
balanced budgets the concepl of equal yield is unambiguous The new equUibrium prices and the
balanced budge condition will del ermine the minimal expendilure and taxes needed to maintain base
30
case publi~ utility Accordingly in general equal yield is inconsistent with equal nominal tax
revenue Some change in tax revenue Is necessary Different tax replacement schemes ate
considered to assure that enough tax revenue is collected This is the approach essentially followed
by most 01 the papers surveyed Only Feltenstein (1985) Goulder (198) and Jorgenson-Yun
(1984) chose not to follow an equal yield strategy
The question is of how to Interprete lila concept of equal yield when the government is alowed to
fUn deflcLts The optimal leve of expenditure for base ease public trliUty can now be taAinanced
bond financed or financed by a mix of bonds and taxation We have several versions of equal yield In
particular equal yield may now be consistent with equal tax revenUE Furthermore some meaSure
ltgtf financial crowding out effects of government deficns can be inferred from the comparislon of the
several equal yield alternatives These issues are extensively discussed ~Pereira (1987b)
44 Price expectations and the dynamic generalization of compensation
Indicators
The sum of equivalent and compensation variations over households is the most widely used
aggregated measure of efficiency gains or losses In the context of $teady~state eomparisions andor
when complete future markets exist andlor perfect foresighl is assumed there are no difficulties
associated with the use of the standard Hicksian indicators In fact correct future prices are known
in these cases
If expectations are not selfmiddot fulfilling andlor future markels are not available a dynamic
generalization is necessary 8allardmiddotGoulder (1985) provide some steps in that direction by
defining an indicator that accomodates periods far into the future when households do not have
perfect foresight but a steady~state prevails On he other hand this issue is more fundamental in
i
[
I
i I I I I [
I I
bullI
31
the contex~ of a Gner1 lemporary equilibrium ftamawor~ In such Circumstances Pereira
(1986a) dev~lops a cnamlc generalization which is obtained as the present discounted value of a i i
sequence of short-fUn optimal expenditure functions consistent with a base case expected future I stream of IlWities I
32
5 Empirical evidenee from selected policy Issues
Dynamic tax models have generated several important results which escaped Sialic modellers
The following is a seleCpoundd set of issuos whIch stress the marginal benefits of dynam~ modelHng of
economic behavior in innovltlttive areas of analysis The discussion of a consumption tax emphasizes
the benefits of dynamic tusehold behavior and intergenerational aspects The study of the
flliminatlon or the re~intToduction of the investsnemt tax credits is made meaningful by the dynamic
modelling of production betlavor Modelling governmenl in a dynamic conlexts allows the modeller
to study the irrrpact of government deflcits Finally a dynamic framework lets us appreCiate the I
relevance of consumer expectations in terms of the evaluation of policy alternatives
i
i
51 Consumption Tax
t
By Its very nature a consumption tax can not be adequately investigated with a static model bull
Fullerion-Shoven-Whaicy (1983) use Ihe model of the US economy s described in
Ballard-Fullerton-Shoven-Whalley (1935) to evaluate the movement from the currenl US tax Isys1em to a progressive consumption tax Since their modeJ Incorporates a Jaborleisure choice
where leisure is an untaxed commodlty their results reflect the fact that both the consumption tax
and Ihe present system are dslortionary
Concentrating on the rntertemporaf distortions they show that sheltering more saving from lhe
current US income tax could improve econornic efficiency even if marginal tax rate increases are
necessary in order to maintain government fevenue~ At first you have a revenue shortfall
However the economy moves to a higher sustatned growth path and uitimatefy lower tax rates
cangenerate the same revenue path Also wages increase in the long run with this policy The
bull
bullbull ~
33 shyt
i
I
I
present value of welfare gi[lS for a potiey of complete savin9 deduction wilh marginal tate
adjustments (consumption lax) is simulated to be around $500 billion to $600 billion 01 1973
dollars
They also investiga1e th 1ngth of ttme it takes th~ economy to adjust to these policy changes
Roughly they estimate the long run to b thirty years although this figure Is very sensitive to the
specification of the savings elasticity - a crucial parameler in this model
52 Investment T~x Credit
Goulder-$ummers (1 S87) address the impact of eliminating the lnvestmenl Tax Cradit (ITC) on
intersectoral capital formation and on economic growth Their model isYa1ticularly adequate to
addies$ this issu~ in that is postulates a forward looking investment m9we with adjustment costs
They show that the eliminating the ITC causes a reduction in the rale of investment Inveslment is
estimated to fall by about 7 in the short-run and by about 12 in Ihe new long run steady state
In lurn a previous policy anncuncement lowers Ihe overall al1ractiveness of Investment and leads 10
a downward shift of the Investrrent profile On the other hand the combined effect of a revenue
neutral simultaneous efimiraron of the ITC and reduction of the corporate tax rates is a long run
reduction ef the capital Sieck by 35 This pattern suggests that a revenue neutral increament in
both Ihe corporale tax rates and the inveSlment lax credits would be preferable 10 the formula
adopted in Ihe US Tax Reform Act of 1986
Pereira (1987d) addresses the impact of the ITe from Ihe oland point of Ihe curTenllax syslem
under the Tax Reform Act of 1986 Given Ihe concern over the potential depressiVe effects upon
savings and investment of the new tax system in conjunction with deficit reduC1ion mechanisms of
the GrammmiddotRudman type a pertinent question is should Ih lTC be e-introduced In the context
bull
34
-of his mod~1 of the US economy Pereira focuses on the trade-off between the distortion in the
intersectoral allocation of capital induced by the lTC the lower tax revenues it generates and
potential financial crowding-out effects on one hand and the positive effects of lowering the relative
price of new capital goods on the other hand Preliminary results confirm the qualitative results in
Goulder-Summers (1987) suggestting a welfare gain of about 2 of the present value of GNP in
the best scenario of absence of replacement taxes Furthermore preliminary results show an
important time pattern to welfare gains with the average benefits increasing the further you look
into the future This pattern is due to the presence of constraints to intersectoral mobility of capital
and inertia in the adjustment towards the optimal capital levels as reflected by the presence of
adjustment costs
I
I
bullAuerbach-Katlikaff (1gB) in the context of their intergenerational model of the US economy
consider the impact on savings capital formation and interest rates of deficit policies and balanced
budget increases in government consumption They conclude that deficit finance and government
consumption can significa-ntly crowd out capital formation and lower the welfare of future
generations However crowding out from deficit finance is a very slow process because it results
from increased government spending over potentially long horizons Also deficit policies may
substantially influence the longmiddotrun interest rates while leaving the short-run rates essentially
unaffected Finally and in opposition to the central role adjustment costs seem to play in both
53 Financial crowding outeff~cts of government deficits
Goulder-Summers and Pereiras models Auerbach-Kotlikoff suggest that the time path of interest
rates induced by a policy of deficit financing seems to be insensitive to the presence of adjustment
costs
bull
35
54 The role 01 consumer expectatlons
The expectations analysis has uncovered one of the benefits of dynamic modelling
Ballard-Goulder (1985) find that the appeal of adopting a consumption tax depends on the level of
foresight possessed by the consumers Furthermore additional foresight may be welfare worsening
This is a second best type of result This tends 10 occur under policies that lead to capital deepening
and declining rate of return to capital over time To the extent that consumers have more foresight
they will be beller equippod to anticipate the fall in the rental price of capital Consequently if a
saving incentive is enacted people save more with myopia than with perfect foresight Given the
exi~tence of taxes on capita and the discrepancy between the private an~--sccial returns to capital
the greater savings with myopia is beller socially Ballard-Goulder find that the welfare gains
from a consumption tax are reduced by about 10 when we move from myopia to farleet foresight
Therefore the attractivenness of adopting a consumption tax seems fairly [obust across these
foresight specifications
I
bull
36
6 Tbe power and weaknesses 01 dynamic modelling the example 01 corporale
tax Integrallon
The Issue of corporate tax integration allows us to sh~w the stre-ngths and weaknesses or the
dynamic approach Empirical evidence using CGE techniques indicates that depending on the precise
scheme of integration and the 1ax replacement methods integratiOn may have substantial effects In
Ihe work of FuliertonmiddotKingmiddotShoven-Whalley (1980 1981) total integration was found to yield an bull
annual static eHielency gain of $4 to $8 billion in 1973 dollars Simulated dynamic gains may be as
large as $695 billion or abcu 14 of the present value of tulUre consumption and leisure in the
US economy These gains result primarily from interindustry reallocalions of Investment and an
improved inter~emporal allocation of consumption
Mora recent work emphasizes hOw consumers asset portfolio decisions and firms finanCial
decisions affect the efficiency gains from integration Slemrod (1980) focusing on consumers
asset ponfoio deCisions finds static efficiency gains which are about twice -as Iarge as those
repOrted iiy Fulierton-King-ShovenWhalley (1961) FuUrlon-Gurdon (1983) focus on the
firms financial decisions They roporl efficiency gains of 6 of GNP from the elimination of the
tax distor1ions favoring debt However when they eliminate the corporate tax and repiace it with
increaSed persona income taxes addlHonal dls10t1ions are created in the optimal labot~leisure
decisions These distortions tend to dominate the analysis slJ(h that the overall effects of complete
integration are very modest Galpr-LuckmiddotTodr (1986) address simultaneously consumers
asset pOrlfollo decisions and the firms financial decisions They also find very modest efficiency
gains from integralion
The above results are important but may be severely biased There are severa aspects of
economic behavior and modelling crucial for the study of Income tax Integration whiCh have no been
i 37
captured innny of the above papers One aspect is the absence of government deficitsmiddot a balanced
budget is assumed It is true Ihal resource crowding OUI is caplured In these mOdels but financial
crowding out induced by government spending is nOI Second investment is not derived from
optimtztlcn behavior Investment behavior passively accomodates endogenous saving decisions As
a consequence the differential impacts of policies in the incentives to save and to invest are not
captured Aso full capital mobility across secors is assumed with instantaneous capital
adjustmenlS towards optimal levels This assumption rules qut different costs of capital across
seclors and therefore differentiated reactions to tax policies changes Finallyhe modelling of both
gov9rnmerlt deficits and of endogenous real and financial investment decisions necessitates the
~nsideraiion of a dynamilt framework and the introduction of financiaJ assets govemment bonds and
iprivate financial assets A dynamic framework also highlights the efficiency effects of Integration on r
ithe ptimal intertemporal decisions The above models are either sIalic (Slemrods and
GalpermiddotltJckemiddotTode(sj or a dynamic sequ~nce of otherwise Slatic model bullbull I Pereir a (1986b) develops a dynamic applied general equilibrium model bull to study the tbull
efficiency effects of integration on the growth and allocation of investmenl across sectors Special
Iattention is paid to the structural affeelS of resource and financial crowding oul induced by
gove mect spending and deficits and to the real and financial Investment reactions across industries
to both tax integration and flnancial crowding out His mode departs from previous work on income
lax Integraticn and for that matler from most 01 the CGIE literalure for tax policy evaluation in
several direcllons It encompasses an endogenous sequential equilibrium struelure founded on
dynamic behavior with flexible expectations Government deficits are optlmaUy determined
Investment decisions are tward looking and the resull of optimizing behavior Several financial
assetsmiddot public and privatemiddot are considered
Simulation resulls Indicale that the welfare gains from Integration under large deficits are at
32
best lJlodest when measured in terms of the GNP The elimination of the corporate tax and its I replacemenl by increased income lax rates yields long run benefits which are never larger than
2 of the presenl yalue ct telr consumption and leisure onder the previous tax regime and 1
under the current tax regime However the short run w~lare effects of full integration tend 10 be
very low and in in some scenarios even negative This is a flew intertemporar pattern of efficiency Ieffects which reflects an adjJsiment lag in the interindustry investment decisions due to the
existence of costs of adjustment On the of her hand partial integration achieved by excluding I dividends from the corporale lax base syslematically yields negalive efficlencyeffeclS In Ihe long Inm these can be as high as 3 of present value of the future value of consumption and leisure This
is a new second best effect suggesllng Ihat less than complete integration may have perverse I efficiency effects
The dynamic slructure of PerclrBs model snowing for an improved freatment of real investment
declsions and government deficts provides a potential setting for a more accurate measure of the
costs and benefits of inlegration The simulation resullS suggest lower benefits and an intertempora Ibull
pattern of growing benefits
Simulation results also clearly suggest robust negative effects from partial integration How
reliable is this result If dJvidends were deductible from the corporate lax base (as are interest
payments) the preferenlial Ireatmant of debl over equily would be ellminaled Corporallons should
be expecled 10 react by decreasing the optimal deWeqully rallo However corporale financial
gecisions are exogenously set In Pereiras model as in aft of the other models surveyed that go as far
as dealing with the issue
Also withdrawing dividends from Ihe corporate tax base is a way of eliminating Ihe double
taxation of dividends at belh Ihe corporate and personnal Income levels How will the optimal
aliocellon of household saving accomodate that change in Ihe relalive return to corporale equity
39 I
t However hOusehold portfolio decisions are exogenouly set in Pereiras model as in all of the other
models surveyeltj go as far as dealing with the issue
IIt is sate to say that the evaluation 01 the benefits 01 partial Integration as defined above are
sevrey underestima1eO Meanwhile the distortionSoenerated by the increase in the marginal
personal tax rates designed to ralse the revenue foregone by the dividend exclusion are fully
aCCltlunted for A good measure of the costs of integration together With a poor evaluation of the
benefils may very well be the reason why partial integration is Simulated 10 yield neg alive
eHikiency gains
On a different vein as most of Ihe model surveyed here Pereiras is a closed economy model It
is legitimate to wonder how the resullS would change If international capital flows were anowed
IThis almost certainly would affect financial crowding out since a good d~al of the capital inflows
could be used to finance public debt
For these various reasons the results should be treated as preliminary but they strongly I suggosl that the dynamic structure provides valuable new insights into the corporate tax integration SSU2
40
7 Concluding remarks I I I i
What has been acccmplished Ihis far The success of Ihe CGE research in laelding the challenge
of dynamic modelling can not be denied
I iGreat progress in the modelling of household behavior has been achieved Promising
developm~nts in the modeUing of production and government behavior have also been accomplished
Interesting innovations in the concept of equilibrium and computation techniques were discussed
The policy analysis was greatly enriched by considering transitional effects together with longer
run steadymiddotstate equilibrium paths generalized equal yield strategies accomodating different
financial crowding out impacts and generalized dynamic policy evaluation indicators
An important set of issues have been addresed by the models surveyed Traditional issues
focusing on the effects of capital taxation and consumption taxes have been pursued In terms of
capital taxation for example the intertemporal nature of the issue was enhanced by allowing an
optimal evolution of capital stock in the economy and forward looking optimal investment decisions
In Andersson (1987) GouldermiddotSummers (1987) and Pereira (1986b 1987d) this is coupled
with a improved treatment of several tax provisions like investment tax credits and depreciation
allowances In terms of the consumption taxes developments in the specification of Intertemporal
household behavior the introduction of overlapping generations and the modelling of
intergenerational links via bequest motives as well as a closer attention to demographic evolution
provided a much improved economic setting for the understanding of the several aspects of the
problem
Furthermore dynamic modelling has permitted the addressing of a set of new issues Policy
Issues ranging from dynamic tax policy analysis to the evaluation of exogenous changes in
government expenditures to the impact of different methods of financing government expenditure to
4 1
the importance of tinanciai croHrii1g out to 1he relevance of consumorS expectations in tne policy
results have been addressed
Despile all of the progress and in part due to such progress several avenues are wide open for
much neded additional research The following seem to be the most Interesting and promising
areasmiddot First the specincation of liquidity constraints to household behavior may help to obtain a
better understanding of policy chenges that affect directly the interest rates in the economy Second
major attention needs to be paid to the incorporafion of a roraign sector (with open capital markets)
Third some efforts are needed in terms of finding adequate computation techniques now that
dynamic modelling eally makes the curse of dimensionality worse than eVOrbull
The modelling of financial markets is the single most unsatisfactory speet of the dynamic
models surveyed here~ The problems associated with provlding a mo~ adequate treatment of
financial markets and in general endogenous and optimal financial deci$ions bull both at the corporate
and at the personal levelsmiddot have 10 be addressed That necessarily requires Introducing uncertainty
into the models To b adequate this must go well beyond the mere assumption of point expeetalions
with the whole array of modelling implementalion and evaluation problems thai generates
bull I i
42
REFERENCES
Abel A and O Blanchard 1983 An Intertemporal Model of Saving and Interest Econometrica
51 pp 675-92
~ Andersson K 1987 Tar-ation of Capital in Sweden - A General Equilibrium Model Ph D
Dissenation UnIversity of Lund Sweder
Auerbach A L KotlikoH 1983 National Savings Economic Welfare and the Structure of
Taxation in M Feldstein ed Behavioral Simulation Methods in Tax Policy- Analysis Chicago
University of Chicago Press
Auerbach A l Kotlikoff 1984 Social Security and the Economics of Demographic-
Transition in H Aaron and G Burtless eds Retirement and Economic Behavior Washington DC
Brookings Institution
Auerbach A L Kotlikoff 1987 Dynamic Fiscal Policy Cambridge Cambridge University
Press
Auerbach A L Kotlikof and J Skinner 1983 The Efficiency Gains from Dynamic Tax
Reform Internatiooaj Economic Reyiew 24 pp 81middot100
Ballard C 1983 Evaluation of the Consumption Tax with Dynamic General Equilibrium
Models Ph D Dissertation Stanford University
Ballard C 1987 Tax Policy and Consumer Foresight A General Equilibrium Simulation
Study Economic Inquiry 25 pp 267-284
Ballard C and L Goulder 1985 Consumption Taxes ForeSight and Welfare a Computatonal
General Equilibrium Analysis in J Piggott and J Whalley eds New Deyelooments in Apolied
General Equilibrium Analysis Cambridge Cambridge University Press
Ballard C D Fullerton J Shoven and J Whalley 1985 A General Equilibrium Model for
t 43
Tax Eoliey Eyalua~iQO NBER University or Chicago Press
Borges A (1987) bull A Survey of Compula1ional General Equilibnum Models CECD Economic
Slud~ 8
Bovenbig L 19B3 imperfect Capital Mobility in a Perfect Foresight Herberger Model
Mimea UC Br~eley
Bovenberg L 1984 Capitel Accumulation and Capital Immobility Qmiddottheory In a Dynamic General Equilibrium Framework PhD dissertation University of California at Berkeley t
Bovenberg L 19850 Dynamic General Equilibrium Tax Models with Adjustment Costs in A iManne ed bull 1985 EtQnomrc ECUjilbrium Model Formulation and Solution Mathematical
lProgramming Siudy 23 Arnsterdam NorthmiddotHoliand
Bovenberg L t 1986 Capital Income Taxation in Growing Open Eeondmies Journa of Pubtic
Eamp(omlcs 3t pp 347middot376 IBovenberg L ald W Keller 1981 Dynamics and Nonlinearlties in Applied General
Equilibrium Models Internal Report Mlmiddot81middot208 Department for Statistical Methods CBS bull
Vooiburg The Netherlands
Charnley C 1981 The Welfare Cost of C~pltal income Taxation In a Growing Economy
IJournal of Political ECQnQm~ 89 3 pp 468middot496 I
Charnley C 19810 Efficient Tax Reform in a Dynamic Model of General Equilibrium I Mimeo Yale University I Charnley C 1982 On the Optimal Taxation In Stylized Models 01 Inlartemporal General
Equilibrium Mimeo Yale University I Decaluw B and A Manens 1985 Pays en Oeveloppement at Modeles Calculables OEquilibre
General - Une Revue de al litterature empirique Monographie Sene 9 Volume 3 Centre de
Recherche et Developpemont Economlque Unrslt de Montreal
44
bull
I I
I
I
1
Erlich $ V Ginsburgh and L van der Heyden 1987 Wher~ do Real Wage Policies Lead
Belgium bull A General Equilibrium Analysis European Economic Reyiew (forthcoming)
air R and J Taylor 1983 Solution and Maximum Likelihood Estimation of Dynamic
Nonlinear Ratioant Expeclations Models Ecooometrica 51 pp 1169middot1186
Feltenstein A 1984 Money and Bonds In a Disaggregated Ec~nonYin Searl 1-1 J Shoven
Eds Aoolied ~eoeral EqJiliDCiulD tioalYsis Cambridge University Press Cambridge
Feltenstein A 19B6 A Dynamic General Equllibrlu Analysis of Financial Crowding Out
Theory with an Application to Australia Journal of Public Economics 31 No1
Fullerlon D R Gorden 1983 A Reexamination of Tax Distorllons in General Equilibrium
Models in M Feldstein Ed Behayora Simulation Methods in Tax pOlicy Analysis Chicago
University of Chipgo Press
Fvllerlon D J Shaven and J Whalley 1983 Replacing the US Income Tax with a
Progressive Consumption Tax JoulOal of Public ECQnomics 20 pp3middot23
Fullerton D Y Henderson J Shaven 1984 A COmparision of Methodologies in Empirical
General Equilibrium Models of Taxation in Searl H J Shoven Eds ~pQlied Geoaral Jguilibrium
lIoalyss Cambridge University Press Cambridge
Fullerton 0 A King J Shaven and J Whaney 1981 Corporate Tax Integration in the United
Statesmiddot a General Equiiibdufi1 Approach American Economic Review 71 pp~ 677~691
Galper H R Lucke E Tader 1985 Taxation portlor Choice and the Allocation of Capital A
General Equilibrium Approach Mimeo
Gill p W Murray M Saunders M Wright 1986 Users Guide for NPSOL (Version 40) A
Fortran Package for Nonlinear Programming T echnicsl Report SOL 86middot2 Deparlment of
Operations Research Stanford University
Glnsbur~h V and L van der Heyden 1985 General Equilibrium with Wage Rigidities An
45
AppJication to Belgium in A Manne ed t Economic EQumprit1m~ Made Formulation and $oJioo
Mathematical Programming Study 23 Amsterdam NorthmiddotHolland ~
Goulet J 1968 1 Adjustrent CO$1S in the Theory o(the Firm Reyiew of EconomiC Studies 35 bull
pp47middot55
Goulder L 1985 Tax-Financed versus BondmiddotFinancedGhanges in Governent Spending
bull Mimeo Harva(d UniversilYmiddot
Goulder l J Shaven and J Whalley 1983 Domestic Tax Policy and the Foreign Sector The
importance of altemative foreign policy formulations to results from a general equilibrium tax
analysis model In ~n methods in Ibe antico of income from caoia1 ed Martin
Feldstein Chicago University of Chicago riSS
Goulder l L Summers 1987 Tax Policy Asset Prices and 3rowtly ~
A General Equilibrium Analysis NBER Working Pper No 212B
Grandmont J 1982 Temporary General Equilibrium Theory Ch 20 in Handbook 01
Mathemalical Eccnornics Vol II NorthmiddotHoliand
Harberger A 1959 The Corporate Income Tax An empirical Appraisal In lax Revision
Comoendjum Voll House Cornmitee on Ways and Means Washington DC Govemment Printing
Office
Harberger A 1962 The Incidence of the Corporate Income Tax JOULn1 of Political
Economy 70 pp 215middot240
Harberger A 1964 Taxation Resource Aliocalrlon and Welfare fn lb Role of Dir bullbulll and
Indjresl Taxes io tbe Fede[al Reserve s~I~Ill Princeton Princeton Unlveresity Press
James J 1985 New Developments in the Applications of General Equilibrium models to
Economic History in Piggott aand Whalley Eds
Jorgenson D and K Yun 1984 Tax Policy and Capital Allocation Harvard Inslitute of
bull
I
I f
Economic Research WP 1107
Judd K 1965 ShonmiddotRun Analysis of Fiscal Policy in a Simple Perfect Foresight Model
Journal of poJiHcal 5CCJiW1l pp 298middot319 r
Upton D J Potbc J Sachsa nd L Summers 1982 Multiple Shooting in Rational
Expectations Models fuoometrlc 50 1329middot1333
Manne A ed 1985 fconQrnjc Eoullibrium Model FOrIDujaljQo and Solution Mathematical
Pogrammlng Study 23 Amsterdam NorthmiddotHolland
Manne A 1985 Or the Formulation and Solution 01 Economic Equilibrium Models In A
Manne ed ECQOQlli-Jr~ibmiddotum ~tQdel EQlJulailon and Solution Malhematical Programming
Study 23 Amsterdam NorthmiddotHolland
Mansur A and J Whalley 1984 Numerical Specification of Applied General Equilibrium
Models in Scarf H J Shoven Eds Aoplied General Eouillbrium AnaiysectIsect Cambridge Universily
Press Cambridge
Merrill 0 1972 Applications and Extensions of an Algorithm thaI Computes FixedmiddotPoints of
Cenain Upper Semimiddotcominuou Point to Set Mappings PhD Disserlation Unlvel$ity 01 Michigan
Negishi T 1960 Welfar Economics and Exislence I an Equnlbrlum for a Competitive
Economy MelroeCOOQ11ka 12 92middot97
Pereira A 1985 On the Existence and Uniqueness of Optimal Ou1put Path lor CRTS Firms in
Adjustment Costs Technologies Mimeo Stanford University
Pereira A 1985a Consumer Expectations and Ihe COSI 01 UllUty In an Intertemporat
Framework Mlmeo Stanford University
Pereira A 1985b Corporale Tax Integrallon and Government Deficits A Dynamic General
Equilibrium Anaiysls Mlmeo Stanford University
Pereir A 1987 On Ihe Computation of the Users Costs of Stocks Ecooomi (forthcoming)
47
Pereira A bull 1987b Equal Yield Alternatives and Government Oeficlts Mimeof Stanford
University
Pereira A 1987c A Dynamic Applied General ~uHlbrium Model for Tax Policy Evaluation
Comlletamp Docurnent2tion Mimeo University of California San Diego (in progress)
Pereira A 1987d Should the investment Tax Credit be Rei(ltroduced Mlmeo University
of California San Diego (ir progreSs)
Piggott J and J Whalley eds 1985 tJew QeyeQpments in Applied General EQuilibrium
bullenalysjs Cambridge Cambridge University Press
Preckel Pbull 1985 Alternative algorithms for computing economic equilibria In A Manne Ed
Radnor R 19amp3 Equllibrium under Uncertainty In K Arrow M Intrilligaor ods
I ibull
Robinson S 1 g86 Mullisectorai Models of Oeveloping CountrieS a Survey Oivision of
Agricullure and Resources UC Borkeley WP 401
Scarf H 1967 The approximation of fixed points of a continuous mapping ~MM Jouwal of
epplid Mathornallcs 15 pp 1328middot43
Scarf H with T Hansen 1973 00 the Computation of EconomiC Eguilibria New Haven Yale
University Press
Scarf H J Shoven Eds 1984 61l1llied Genera Elujlibrium 6nalyssect Cambridge University
~ Press Cambridge
Shoven J 1976 The Incidence and Efficiency Effects of Taxe on Income from Capital
Jurnal of ramie1 Economy 84 pp 1261-83
Shoven J 1983 Applied General Equilibrium Tax Modemng IMF Slaff PaoerS 30
Shoven J and LB Simon 1986 Share Repurchases and Acquisitions An Analysis of which
Firms Participate NBER Working Paper No 2243
bull
48
IShaven J and J Whalley 1972 ft General Equmbrium Calculation of lhe Effects of
Differential Taxation of Incoma from Capital in the US Jauroe of public Econgmics 1
pp281middot321 (
Shoven J bull and J Whalley 1973 General Equilibrium with Tax$s a computation procedure
and an existence pr00f ReviEhi of Economic Studies 40 pp 475middot490
Shoven J J Whalley 1977 Equal Yield Tax Alternatives A General Equilibrium
Computa~ui0nal Tfchnique J~HJrnal of PUblic EconomIcs 8 pp 211~224 bull
Shaven J J Whalley 1984 Applied GeneralmiddotEquilibrlum MOdels of Taxalion and
International Tr8de An Introduction and Survey Journal of Economic literature 23t pp
1007middot1051
Slemrod J 1980 A General Equilibrium Model of Capital Income Taxation Ph D bullDissertation Harvard Univnrsity
Slemrod J 1983 A General Equilibrium Model of Taxation with Endogenous Financial
Behavior in M Feldstein edbull Behavioral Simulation Methods In Tax Policy Analysis Chicago
UniVerSity of Chicago Press
Stone J 1985 Sequential Optimization and Complementarily Techniques for computing
Economic equilibrium in A Manne Ed
Summers L 1981 Taxation and Corporate Investment a q-Theory Approach BroQkinas
Summers L 1985 Taxation and the Size and ComposHion 01 the Capital Stock An Asset Price
Approach Mlmeo Harvard University
Whalley J 1975 A Geneal Equiloibrium Assessment of the 1973 United Kingdom tax
reform Economica 42 pp 139middot61
1JLTIllOS JIQll~)ltG PAPERS PUBLICADOS bull
I
I I
j
112 70 - LUCENA iJiCIIJo ~ IIEnvironrilent Honitoring and Orqanization Structure T (Setellibro 1987)
nI 71 - L~ENA Diogo A Note on the Representation of Information Stxueture (SeteJObro 1987)
nQ 72 - LUCENA D1CgO Envuorment 111)nitoring and Organization
Structure II (Setelilhro 1987)
ril 73 - AlmmE3 Antonio Pais e CAsPAR Yitor bull Tributacao Inoentivos e InvesUlllentos Analise Qualitativa noYellibro 19S7)
n2 74 - liAlltOSA rose Pedro Opt1lllal lIage R1gidity A SUggestedtlthodology to Teet tb Theory and an Application (Outubro 1987)
nI 75 - CABRAL L 11 B Three lIotes on SYJIIlletric Gemes witt itsYJIIlletric Equilibri (NoveJObro 1987)
n2 76 - LUCENlt D10gOTo Search or Not to Search (Fevereiro1967)
nQ 17 - SA Jorge asconcelos euroI A Theory of SynergyB (Fevereiro1987)
n2 76 - VILARES tanuei Jose Os Be IntenaMios IlIIportados ColIIO Factor de Produao (Julho 1987)
nQ 79 - SA J~rge VasCQncel03 e HoY to Copete and Co~unicate in tature Industrial Products_ (aio 1988)
n2 80 - ROIl Rafael Learning and Capacity Expansion in a New larket Under Uncertainty (Fevereiro 1988)
n2 81 PEREIRA Alfredo llarvilo Survey of Dynamic Computational General EquilibriUlll Hodel for Tax Policy Evaluation (Outubro 1987)
Qualquer in ormacao sabre o 1IIorlltlng Papers ja publlcados sera prestada pelo Secretariado de Apoio aos Docentes podendo os ~e3OSf
ser adquirido5 ne Secyo de Vlndas da Faculdade de EconollLia UNL na Travessa Esteviio Pinto Cltpolide - 1000 LISllOA
middot i
19
bull 33 government behavior
middot
Two issues dominate the modelling of government behavIor First govemment behavior has
Ibeen typically seen in Ihe CGE liIerature for lax poliCY evaluation as constrained by yearly balanced
I budsets (see for example 8allard-Fullrton-S~ovn-Whafley 098Sraquo The analysis of
govemment defidlS and public debl in a CelE context requires a dynamic selling Second the level of
government expenditures is either exegenouty given as in Auerbach-KoUikofi (1984 1987)
6ovenbrg (1984 1985) Feltenstein (1984 1986) and Jorgenson-Yun (1984) or
endogenously (but no optimally) determined by the balanced budget conditions as in
6allard-Fullerton-Shoven-Whalley (1985) Andersson (1987) Erllch-Ginshurgh-Heyden
~
(1987) Gouder (19aS) and Gouder-Summers (1987) In the second case the composition of
bull public expenditures is often optimally determined However the leve~ of government expendi1ures
can only 00 endogenously and optimally determined if the government is seen as an optimizing agent
and is allowed to run deficits
The first attempts to deal wiU the government defiCits in CGE tax models are due to
Auerbach-Kollilltoff (1984) and Feensain (1984) In Auerbach-Kotlikof (1984) government
expenditures afe exogenous They grow at the rate of growth of population However given 1he tax
structure and tax revenues yearly deficits and surpluses are atTowed subject to an in1ertempond
constraint that the present value of future tax revenues equals the present value of future
expenditures
In Fellenstein (1984 1986) government expenditures are also exogenously given He also
allows the government to run deficits to finance expenditures in excess to 1ax revenues However
surpluses are returned to CQnsumars in the form cf transfers Accordingly government ts not
subject to any constraint regarding the future repayment of public debt
In GouJder (19851 the 9Vernmenl maximizes a static social welfare ~nctio subject to a-
bull
20 I i
I
I I I
I bull
balanced budget conSirtlL This follows the optimal alocatkm of government expenditures alpng
the lines of Baliard-FHeiton-Shoven-Whalley (1985) Ttle model Is generallzea to allow for
exogenous changes in HIe time path of government expenditures Two financing alternatives are
consjdeft~d and contr8s1ed EldditionaJ tax revenues and bond issuance
Pereira (198Gb 1137d) auempts to address both the incorporation of deficits and the
determination of govement expenditures The path of government expenditures and the path of
delicilSlsurpluses (ano therefore the path lor debt) are endogenously and optimally determined The
government is seen as mltJximizing n intertemporal social welfare function given the 13x structure
Optimization is subject to a sequence of recursive equations of motion reflecting the evolution of the
ptlbllc debt allowing fDr government budget imbalances It should be stressathat this specification
01 the government cOMtcalnt is equivalent to the specificaUon in Auerbaeh-Kotlikoff (1983 1987)
when no liquidity constraints affect -gove~nment behavior
The extra richness in the treatment of government behavior allows tfte examination of
government debt poices in a truly dynamic setting Also financial crowding out effects induced by
government deficits can be analyzed (see Auerbach-Kottlikoff (1987) Feltenstein (1986) and
Perer (198Gb 1987e))
34 Foreign sector
Most of the open economy models follow the assumptions of balanced trade with import and
export net demands characterized by constant elasticities along the line of Ballardmiddot
-Fullerton-Shaven-Whalley (1965) Such is the case of Salfard (1963) Ballard-Goulder
(19S5) and Goulder-Sllmmers (1987)_ In Feltenstin (1985) the rest of the world is treated as
21
an addjlion~ consumer group
Bovcoberg (1986) develops a model In which two economies lIr considered each following
intertempoml perlect foresight paths These economies meet in the international forum Their
trade relaroMhips are characterized by yearly balanced trade accounts
None of the new generation of dynamic cae tax models has yet IncorPorated the lnternational
capital flows as done in aoulder-Shoven-Whalley (1983) for the earlier Ballarrshy
-Fullertonmiddot Shaven-Whalley (1985) model This is unfortunate as there is an important
Intertmporal aspect to international lending The first al1empts along Ihese lines are due to
Andersson (1987) and Erlich-Glnsburgh-Heyden (1987) Andersson (1987) in his model of the
Swedish economy adopts a closed economy appraoch in which rates of return in the domesJic
economy are largely determined by the international capital markets Fidinterest rate induce
Intlnatlonal capital flows which determine and finance the International lrade imbalance In
lurn In ErlichGinsburghHeydenmiddot (1987) foreign trade Is generaled according to an
intertemporal Irade welfare lunction with constant import and export elasticities In the shortmiddotrun
they allow inlernational trade inbalances which generate capilal ftows to the domestic households In
the long run however trade balance is assumed
35 The need for financial markets
A dynamic economic Wucture not only provides the ideal environment 10 model many features of
economic behavior it also permits one to Incorporate the financial side of the economy If the
government is allowed to run deficits the question 01 deficll financing automatically follows If
investment is optimally dewmined and returns to capital are different aClOSS sectors problems of
investment financing arise If there are seeral final1cfal assets In the economymiddot government
4-
I i
ibull
I I I I
I
j
22
bonds privpte bonds and equity (or simply physical capital installed inmiddot different sectors) the
problem of allocation of saving among assets with potentially different returns arises
The papers surveyed var greatly with respect to l~e extent of their attention to the financial
side of the economy At Clne extreme are the models in ~anardmiddotFuliertonmiddotShoven-Whalley (1985)
Andersson (1987) Ballltld (1983) Ballard-Goulder (1985) Bovenberg (1985 1986) and
Erlich-Gjnsburgh-Heyden C19B which are devoted exclusively to the real side of the economy
In turn Auerbach-KotliKoff (1984 1987) Feltenstein (1984 1986) and Goulde (1985)
allow for government debt In these models saving finances changes in government debt and physical
capital Private and publlc 2ssets are perceived by the households as perfect substitutes The
allocation of saving merely adjusts to the relative demands for funds
Feltenstein (1984 1986) is the only model surveyed which introduces money Government deficits are financed by issuing money and bonds according to an exogenously given rule Money is
demanded by consumers for transaction motives and an exogenously given fraction as a store of value
On the other hand government bonds and physical capital are the vehicles for the intertemporal
transfer of wealth
Summers (1985) and GOlldermiddotSummers (1987) introduce a whole menu of financial assets shy
firm specific equity capital Different assets earn different rates of retum However such rates
are equal up to constant and exogenous sector speCific risk premia Therefore the introduction of
constant exogenous risk premia as helpful as it may be in the context of calibration does not solve
the main issue of the non-optimality of the allocation of saving Also talking about risk premia in a
deterministic context is somewhat unsatisfactory
Pereira (1986b 1987d) also introduces a whole menu of assets private and public bonds and
firm specific equity However all the assets are expected to yield the same rate of return and
therefore perc~jved as perfect substitutes The different asset types allow consideration of
23
exogenOU$~8tUequjty triO dividendlrc~enlion rufes and therefore seve-ral sources of investment
financing bond equity and retained earnings
The nonmiddotcptimalHy of the allocation of saving and the absence or exogenelty of corporate
financial rules is _a reftectlon of the limitations of the determInistic approach Either in a context of
perfect anticipation oi the future prices Of in general within the t$atm of point price
expectations all the papers follow a deterministic approach Under such circumstances consumers
either expect diHerent rales of return (inclusive of risk premium) across assets in which case
they wilt buy only one asset (that with highest rate) Of they expect equal rat~s of relurn in which
case Ihey are indifferent about the asset composition of Iheir portfoliO There is no way of
Iradingmiddotoff rales of return and risks to obtain an optimal interiof solution to the problem of the
allocation of saving
The most advanced contribution in the modelling of saving allocation in a CGE setting is due to
Slemrod (1980 1983) in the context of a static one-period model In his model consumers act
according to a two stage separable decision process They Hrsl decide on how much to save Then
they decide on the allocation of saving according to an Indirect uttlity function dependent on the rates
of ratum and variances offered by the different assets in the eoonomy The sourCe of riskiness in the
ampConory comes from an uncertain marginal product of capital On the other hand aside from
portfolio decisions the res of the economy is insulated from uncertainty
The most complete contribution in terms of the treatment of the corporate financial rules is
FulienonmiddotGordon (1983) In a variant 01 Ihe BailardmiddotFulienon-ShovenmiddotWhelley (1985) model
Fullerton and Gordon have capital intensity and optimal financial decisions Jointly deter~ined
through a two siage process The cosl of financing capital is minimized by trading off the tax
advantages of debt against the epected real bankrupcy coSls inherenl with high debVequity ratios
Given the optimal debtequity ratio the level of inveSlment Is chosen such that at Ihe margin the
24 l
bull
rerurn pn ~uity equas the return on bonds plus an exogenous risk premium
36 Market assumptions equilibrium and expectations
Virtually all of the papers arc characterized by Walrasian market clearing assumptions All
markets are perlect1y competi1ive Atomistic competition among agents 1s assumed even thoughmiddot only
a flnile number of agents is considered Virtually no mark~t disequilibria or price stickiness afe
considered The only exception is ErlichmiddotGinsburghmiddotHeyden (1987) bull In thei model of the Belgium
economy I the _wage rate is fixed in the short-run Therefore in the shOrlwrun desequilibrium in
the labor market will generate endogenous unemployment However in lh$ long-run all prices
Including the wage rata are Ilexible and accordingly ali markets ciear
Given the dynamic nature of behavlor in the economy markelmiddotcfearing prices in each period
depend on expectations of fture prices and on tax variables In the economy There are essentially
two ways at interpreting the economic equilibrium in such a dynamic conte~bull If future prices are
perfectly anticipated (io expectations are self-fulfilling) a perfect foresight equilibrium
prevaHs Then future actions are merely the implementation of current decisions for future
periods However jf ptica expecla110ns are not perfect (Le agents make mistakes w1th re$pect 10
future prices) then a temporary or shortrun equilibrium prevails Markets ciear and clearing
prices depend on fu1ure price expectations Current plans about the future are typically not
precisely implemented They will be revised as more or better information becomes available to the
economic agents See Grandmant (1982) for a comprehensive survey 01 the temporary equilibrium
literature
Wilh the exception of Gould (1985) and Pereira (19S6b 19S7d) all the models surveyed
adopt th~ concept of a perfect foresight equilibrium In turn BaliardmiddotGouldr (1965) considers a
J
2S
flexible am9unt of foresight In terms of the number of years over which price movements are ~
foreseen Pereiras model (1986b 1987b)) is flexible in a somewhat different way in that it can
indude any range of foresight from myopia to perfect foresight The choice between the perfct foresight and lemporary equilibrium is ultimately to be made on
philosophic grounds It can be argued that less than perfect expectations Imply that agents are
irralional in some way (see Auerbach-Kctiikoff (1987 p 10) However the reverse argument can
be made One can question wiether agents are reaHy rationar and perfectly knowledgable about
future prices
Recent evidene of Balird (1987) Ballard-Goulder (19851 Goulder (19851 and Pereira
(1986b 1987d) suggests that the choice in modemng expectations is an important one They show
that the degree of foresight inl0 thc future (ranging from perfect foresight to-myopic expectations)
may have dramatic impacts on the pollcy conclusions of Ihe model Accordingly the best research
strategy may b to design models which are flexible enough 10 allow for different rules regarding the
formation of expectations With the exception of the articles just mentioned sensitivity analysis
bullhave previously not been performed along this dimension
In terms of implementation the two concepts of equilibrium - perfect foresight and temporary
equilibrium have different implications The dimensionality of the equ1ibrium soll1ion algorIthm
is involved Suppose we have a model with ten markets to be run for a period of 50 years Aside
from normalization a perfect foresight model implies computing prices In 500 dimensions while a
-t~mporary equilibrium model requires sorving SO equilibria each in ten dimensions Given that
computational speed often varies with the cube of the number of dimensions the lemporary
equilibrium formulation is potentially strikingly more feasible However Ballard (1987) and
Goulder-Ballard (1985) have devoloped techniques to greatly speed the computation of a perfect
foresight equilibrium
26
j
I
4 Jrnplemenlalion and Issues on policy evaluallon
41 Calibration and equilibrium comparlslon
CGE models at typically parameterized by the use of a calibration procedure Some parameters
are exogenously given However some crucial parameters are determined in such a way that the
model replicates the data for a given base year See MansurmiddotWhalleY 11984) for an eXlensive
discussion of 1~i$ issue
CalibratIon in a dynamic context is generaliy interpreted as requiring two properties First
replication of base year data is required Second the model is parameterized to simulate an
intertemporal balanced growth path when Ihe base policy is maintaineq This Is the approach
followed by Ballard (1983) BallardmiddotGoulder (1985) Goulder (1985) and SummersmiddotGoulde
(1986) It follows the practice of BallardmiddotFuliertonmiddotShovenmiddotWhalley (1985) with their
sequential cquilibrium dynamics
Other authors Auerbach-Kotlikof (1983 1984 1987) Boyenberg (1984 1985 1986)
and Pereira (198Gb 1987d) follow whal we cali a qualitalive calibration The slructural
palametars are exogenously chosen so that lh economy follows a reasonable path into Ihe fulure
ThaI has 10 do wllh the fact thaI given the recursive nature 01 the dynamic economy and aside tom
such structural parameters only iniUa stock values are needed to run these models Given initial
conditions on the stocks of say private wealth capltal and government debt agents will optimize
and Ihereby generate a lirst round of net demands and equilibrium conditions In lurn the
equilibrium prices will determine the evolution of the stock variables into the next perIod
There are several potential problems with calibration in the conlext of dynamic models The
assumption of a steadymiddotstate growth ~ath in the base case can be questioned First while
27 I Ibullbull
steadymiddotstaU is a possibity it certainly is not the only meaningfull solution to dynamic models
Even in the case of a perfect foresight equilibrium the model implies an equilibrium path which i may or may not involve colanced growth The model not the modeller should dictate the nature 01 I themiddotbase case path Second in the context of a temporarY equilibrium path a steady state solutlcn is Inol a likely model QutcOn1tl In fact unlessmiddot expectations are stattc short run behavior consistent
with a steady-state evolution will in general not be generated On the other hand if static I expectations are self~fufjili[1g we have in fact a perfect foresight model Third even the base year
replication requirement may cause problems In Ihe cOntexl of temporary equilibrium Any
calibration parameter would be conditional on expectation rules which is probably an undesirable
feature
The nalur of the two-requirement calibration strategy is very muchin the spirit of the bull
traditional design of the comparision of alternative equilibria comparis_ion between a steady-state
base case on one hand and alternative paths including a transition period and a finat sleadyyenstate on
the other hand Pereira (198Gb 1987d) compare different (not necessarily steadystate)
equilibrium paths The arguments against such a procedure are based on the idea that the impact of
policy changes can be observed most easily since all departures from the steadymiddotstate can be
attributed to the altemalive polioy On the other hand the base case so dEfined as a steadymiddotstate is
consist~nl with previous work in a less dynamtc setting and Iherefore allows a common standard
for comparing model results
42 Computation algorithms
We are still al a stage in which basically each author uses a different computational technique
Th development of dynamic models - in parlicular wilh adjustment costs andior perfect foresight
bull
28
has corresponded with the decline in the use of fixed point algorithms In -fact given the relative
arge dimensions inevitably involved such algorithms lend to be very inefficient at the best and
chen prohibitively stow See Stone (1985) and Preckel (1965) for a comparative assessment of
different computation techniques Among the models surveyed only Ball1rdFulle(ton~
-ShavenmiddotWhalley (1985) and Feltensteln 1985 use Merrills variant of the fixed point
algorithm 1echnique
AuerbachmiddotKoHikoff (1963 1984 1967) follow a three stage procedure They first compute a
base case steady-state then a revised case steady state and finally transition path for the economy
berveen these two steady-states tn all stages a Gauss-Seidel iterative procedure is used
Ballard (1982) Ballard-Goulder (1965) Goulder (HiSS) and Summers-Goulder (1986)
~
use a method developed by Ballard and Goutder which is similar in many awects to the FairmiddotTaylQr bull
(1983) algorl1hm Short-run equilibria are calculated (using Merrills algorithm) parametric on
nrice expeC~lions The model is then iterated to generate self-fulfilling intertemporaf expectations
and the corresponding perfect foresight equilibrium In a relatively similar approach Andersson
1987 uses a simulation program SIMNON developed In the University of Lund This program
can handle two-paint boundary problems In a fashion consistent with the multiple shooting
algcrlthm (see Lipton-Poterba-Sachs-Eummers (1982))
Boyenbergs computational approach (1985 1985) differs from the other models surveyed in
that he relies heavily on analytical techniques Computations are done by using a dynamic version of
Johansons linearization method Being essentially determined by the continuous time nature of the
I I
f
model thiS linearization model has the disadvantage of confining the analysls to infinltesinal changes
around the base case equilibrium (see Bovenberg (1985) p 53) Pereira (1986b 1987d) uses an optimization algorithm NPSOL - developped by
rGillMurray-Saunders-Wrlght (1986)) This algorithm is used 10 compute the sequence of
I
29
Shortmiddotrun temporary equilibrium which makes up the intertemporal equilibrium path The
equilibrium conditions are seen as nonlinear equality constraints in the minimization of an artlcial
objective function The prices are normalized to the unit simplex by an additional linear equality I I 1 bull constraint
II ErlichmiddotGinsburgh-Heyden (1987) follow a unique approach in that they use a variant of the
optimization technique introduced by Negishl (1960) The economic equilibrium can be generated i
as a solution of a mathemallcal program Ihe objective function of which is a weighted sum of the
utility functions of the various agents while tho constraints set consists of the market dearing 1
conditions Ginsburgh-Heyden (1985) have extended Negishis resulllO tho case of downward price
lrigidities I
Finally the paper by Jorgonson-Yun (1984) is also unique in tIlat it is the only I
econometrically estimated model Different blocks for the consumption and production Side of the
-
model are separately estimated to provide the necessary structural parameters i
I The diversity of computation techniques is yet another indicator of the exploratory nature of the
body of literature surveyed in this article
43 Equal yield comparisions
The link between a base case and counteriactual simulations Is usually provided by the concept of
equal yield Shaven-Whaley (1977) discuss the meaning 01 equal yield in a general equilibrium
context when government is not allowed to run defICits Equal yield is interpreted to moan constant
public utility Government base case utility is maintained in the counferfactual experiments With
balanced budgets the concepl of equal yield is unambiguous The new equUibrium prices and the
balanced budge condition will del ermine the minimal expendilure and taxes needed to maintain base
30
case publi~ utility Accordingly in general equal yield is inconsistent with equal nominal tax
revenue Some change in tax revenue Is necessary Different tax replacement schemes ate
considered to assure that enough tax revenue is collected This is the approach essentially followed
by most 01 the papers surveyed Only Feltenstein (1985) Goulder (198) and Jorgenson-Yun
(1984) chose not to follow an equal yield strategy
The question is of how to Interprete lila concept of equal yield when the government is alowed to
fUn deflcLts The optimal leve of expenditure for base ease public trliUty can now be taAinanced
bond financed or financed by a mix of bonds and taxation We have several versions of equal yield In
particular equal yield may now be consistent with equal tax revenUE Furthermore some meaSure
ltgtf financial crowding out effects of government deficns can be inferred from the comparislon of the
several equal yield alternatives These issues are extensively discussed ~Pereira (1987b)
44 Price expectations and the dynamic generalization of compensation
Indicators
The sum of equivalent and compensation variations over households is the most widely used
aggregated measure of efficiency gains or losses In the context of $teady~state eomparisions andor
when complete future markets exist andlor perfect foresighl is assumed there are no difficulties
associated with the use of the standard Hicksian indicators In fact correct future prices are known
in these cases
If expectations are not selfmiddot fulfilling andlor future markels are not available a dynamic
generalization is necessary 8allardmiddotGoulder (1985) provide some steps in that direction by
defining an indicator that accomodates periods far into the future when households do not have
perfect foresight but a steady~state prevails On he other hand this issue is more fundamental in
i
[
I
i I I I I [
I I
bullI
31
the contex~ of a Gner1 lemporary equilibrium ftamawor~ In such Circumstances Pereira
(1986a) dev~lops a cnamlc generalization which is obtained as the present discounted value of a i i
sequence of short-fUn optimal expenditure functions consistent with a base case expected future I stream of IlWities I
32
5 Empirical evidenee from selected policy Issues
Dynamic tax models have generated several important results which escaped Sialic modellers
The following is a seleCpoundd set of issuos whIch stress the marginal benefits of dynam~ modelHng of
economic behavior in innovltlttive areas of analysis The discussion of a consumption tax emphasizes
the benefits of dynamic tusehold behavior and intergenerational aspects The study of the
flliminatlon or the re~intToduction of the investsnemt tax credits is made meaningful by the dynamic
modelling of production betlavor Modelling governmenl in a dynamic conlexts allows the modeller
to study the irrrpact of government deflcits Finally a dynamic framework lets us appreCiate the I
relevance of consumer expectations in terms of the evaluation of policy alternatives
i
i
51 Consumption Tax
t
By Its very nature a consumption tax can not be adequately investigated with a static model bull
Fullerion-Shoven-Whaicy (1983) use Ihe model of the US economy s described in
Ballard-Fullerton-Shoven-Whalley (1935) to evaluate the movement from the currenl US tax Isys1em to a progressive consumption tax Since their modeJ Incorporates a Jaborleisure choice
where leisure is an untaxed commodlty their results reflect the fact that both the consumption tax
and Ihe present system are dslortionary
Concentrating on the rntertemporaf distortions they show that sheltering more saving from lhe
current US income tax could improve econornic efficiency even if marginal tax rate increases are
necessary in order to maintain government fevenue~ At first you have a revenue shortfall
However the economy moves to a higher sustatned growth path and uitimatefy lower tax rates
cangenerate the same revenue path Also wages increase in the long run with this policy The
bull
bullbull ~
33 shyt
i
I
I
present value of welfare gi[lS for a potiey of complete savin9 deduction wilh marginal tate
adjustments (consumption lax) is simulated to be around $500 billion to $600 billion 01 1973
dollars
They also investiga1e th 1ngth of ttme it takes th~ economy to adjust to these policy changes
Roughly they estimate the long run to b thirty years although this figure Is very sensitive to the
specification of the savings elasticity - a crucial parameler in this model
52 Investment T~x Credit
Goulder-$ummers (1 S87) address the impact of eliminating the lnvestmenl Tax Cradit (ITC) on
intersectoral capital formation and on economic growth Their model isYa1ticularly adequate to
addies$ this issu~ in that is postulates a forward looking investment m9we with adjustment costs
They show that the eliminating the ITC causes a reduction in the rale of investment Inveslment is
estimated to fall by about 7 in the short-run and by about 12 in Ihe new long run steady state
In lurn a previous policy anncuncement lowers Ihe overall al1ractiveness of Investment and leads 10
a downward shift of the Investrrent profile On the other hand the combined effect of a revenue
neutral simultaneous efimiraron of the ITC and reduction of the corporate tax rates is a long run
reduction ef the capital Sieck by 35 This pattern suggests that a revenue neutral increament in
both Ihe corporale tax rates and the inveSlment lax credits would be preferable 10 the formula
adopted in Ihe US Tax Reform Act of 1986
Pereira (1987d) addresses the impact of the ITe from Ihe oland point of Ihe curTenllax syslem
under the Tax Reform Act of 1986 Given Ihe concern over the potential depressiVe effects upon
savings and investment of the new tax system in conjunction with deficit reduC1ion mechanisms of
the GrammmiddotRudman type a pertinent question is should Ih lTC be e-introduced In the context
bull
34
-of his mod~1 of the US economy Pereira focuses on the trade-off between the distortion in the
intersectoral allocation of capital induced by the lTC the lower tax revenues it generates and
potential financial crowding-out effects on one hand and the positive effects of lowering the relative
price of new capital goods on the other hand Preliminary results confirm the qualitative results in
Goulder-Summers (1987) suggestting a welfare gain of about 2 of the present value of GNP in
the best scenario of absence of replacement taxes Furthermore preliminary results show an
important time pattern to welfare gains with the average benefits increasing the further you look
into the future This pattern is due to the presence of constraints to intersectoral mobility of capital
and inertia in the adjustment towards the optimal capital levels as reflected by the presence of
adjustment costs
I
I
bullAuerbach-Katlikaff (1gB) in the context of their intergenerational model of the US economy
consider the impact on savings capital formation and interest rates of deficit policies and balanced
budget increases in government consumption They conclude that deficit finance and government
consumption can significa-ntly crowd out capital formation and lower the welfare of future
generations However crowding out from deficit finance is a very slow process because it results
from increased government spending over potentially long horizons Also deficit policies may
substantially influence the longmiddotrun interest rates while leaving the short-run rates essentially
unaffected Finally and in opposition to the central role adjustment costs seem to play in both
53 Financial crowding outeff~cts of government deficits
Goulder-Summers and Pereiras models Auerbach-Kotlikoff suggest that the time path of interest
rates induced by a policy of deficit financing seems to be insensitive to the presence of adjustment
costs
bull
35
54 The role 01 consumer expectatlons
The expectations analysis has uncovered one of the benefits of dynamic modelling
Ballard-Goulder (1985) find that the appeal of adopting a consumption tax depends on the level of
foresight possessed by the consumers Furthermore additional foresight may be welfare worsening
This is a second best type of result This tends 10 occur under policies that lead to capital deepening
and declining rate of return to capital over time To the extent that consumers have more foresight
they will be beller equippod to anticipate the fall in the rental price of capital Consequently if a
saving incentive is enacted people save more with myopia than with perfect foresight Given the
exi~tence of taxes on capita and the discrepancy between the private an~--sccial returns to capital
the greater savings with myopia is beller socially Ballard-Goulder find that the welfare gains
from a consumption tax are reduced by about 10 when we move from myopia to farleet foresight
Therefore the attractivenness of adopting a consumption tax seems fairly [obust across these
foresight specifications
I
bull
36
6 Tbe power and weaknesses 01 dynamic modelling the example 01 corporale
tax Integrallon
The Issue of corporate tax integration allows us to sh~w the stre-ngths and weaknesses or the
dynamic approach Empirical evidence using CGE techniques indicates that depending on the precise
scheme of integration and the 1ax replacement methods integratiOn may have substantial effects In
Ihe work of FuliertonmiddotKingmiddotShoven-Whalley (1980 1981) total integration was found to yield an bull
annual static eHielency gain of $4 to $8 billion in 1973 dollars Simulated dynamic gains may be as
large as $695 billion or abcu 14 of the present value of tulUre consumption and leisure in the
US economy These gains result primarily from interindustry reallocalions of Investment and an
improved inter~emporal allocation of consumption
Mora recent work emphasizes hOw consumers asset portfolio decisions and firms finanCial
decisions affect the efficiency gains from integration Slemrod (1980) focusing on consumers
asset ponfoio deCisions finds static efficiency gains which are about twice -as Iarge as those
repOrted iiy Fulierton-King-ShovenWhalley (1961) FuUrlon-Gurdon (1983) focus on the
firms financial decisions They roporl efficiency gains of 6 of GNP from the elimination of the
tax distor1ions favoring debt However when they eliminate the corporate tax and repiace it with
increaSed persona income taxes addlHonal dls10t1ions are created in the optimal labot~leisure
decisions These distortions tend to dominate the analysis slJ(h that the overall effects of complete
integration are very modest Galpr-LuckmiddotTodr (1986) address simultaneously consumers
asset pOrlfollo decisions and the firms financial decisions They also find very modest efficiency
gains from integralion
The above results are important but may be severely biased There are severa aspects of
economic behavior and modelling crucial for the study of Income tax Integration whiCh have no been
i 37
captured innny of the above papers One aspect is the absence of government deficitsmiddot a balanced
budget is assumed It is true Ihal resource crowding OUI is caplured In these mOdels but financial
crowding out induced by government spending is nOI Second investment is not derived from
optimtztlcn behavior Investment behavior passively accomodates endogenous saving decisions As
a consequence the differential impacts of policies in the incentives to save and to invest are not
captured Aso full capital mobility across secors is assumed with instantaneous capital
adjustmenlS towards optimal levels This assumption rules qut different costs of capital across
seclors and therefore differentiated reactions to tax policies changes Finallyhe modelling of both
gov9rnmerlt deficits and of endogenous real and financial investment decisions necessitates the
~nsideraiion of a dynamilt framework and the introduction of financiaJ assets govemment bonds and
iprivate financial assets A dynamic framework also highlights the efficiency effects of Integration on r
ithe ptimal intertemporal decisions The above models are either sIalic (Slemrods and
GalpermiddotltJckemiddotTode(sj or a dynamic sequ~nce of otherwise Slatic model bullbull I Pereir a (1986b) develops a dynamic applied general equilibrium model bull to study the tbull
efficiency effects of integration on the growth and allocation of investmenl across sectors Special
Iattention is paid to the structural affeelS of resource and financial crowding oul induced by
gove mect spending and deficits and to the real and financial Investment reactions across industries
to both tax integration and flnancial crowding out His mode departs from previous work on income
lax Integraticn and for that matler from most 01 the CGIE literalure for tax policy evaluation in
several direcllons It encompasses an endogenous sequential equilibrium struelure founded on
dynamic behavior with flexible expectations Government deficits are optlmaUy determined
Investment decisions are tward looking and the resull of optimizing behavior Several financial
assetsmiddot public and privatemiddot are considered
Simulation resulls Indicale that the welfare gains from Integration under large deficits are at
32
best lJlodest when measured in terms of the GNP The elimination of the corporate tax and its I replacemenl by increased income lax rates yields long run benefits which are never larger than
2 of the presenl yalue ct telr consumption and leisure onder the previous tax regime and 1
under the current tax regime However the short run w~lare effects of full integration tend 10 be
very low and in in some scenarios even negative This is a flew intertemporar pattern of efficiency Ieffects which reflects an adjJsiment lag in the interindustry investment decisions due to the
existence of costs of adjustment On the of her hand partial integration achieved by excluding I dividends from the corporale lax base syslematically yields negalive efficlencyeffeclS In Ihe long Inm these can be as high as 3 of present value of the future value of consumption and leisure This
is a new second best effect suggesllng Ihat less than complete integration may have perverse I efficiency effects
The dynamic slructure of PerclrBs model snowing for an improved freatment of real investment
declsions and government deficts provides a potential setting for a more accurate measure of the
costs and benefits of inlegration The simulation resullS suggest lower benefits and an intertempora Ibull
pattern of growing benefits
Simulation results also clearly suggest robust negative effects from partial integration How
reliable is this result If dJvidends were deductible from the corporate lax base (as are interest
payments) the preferenlial Ireatmant of debl over equily would be ellminaled Corporallons should
be expecled 10 react by decreasing the optimal deWeqully rallo However corporale financial
gecisions are exogenously set In Pereiras model as in aft of the other models surveyed that go as far
as dealing with the issue
Also withdrawing dividends from Ihe corporate tax base is a way of eliminating Ihe double
taxation of dividends at belh Ihe corporate and personnal Income levels How will the optimal
aliocellon of household saving accomodate that change in Ihe relalive return to corporale equity
39 I
t However hOusehold portfolio decisions are exogenouly set in Pereiras model as in all of the other
models surveyeltj go as far as dealing with the issue
IIt is sate to say that the evaluation 01 the benefits 01 partial Integration as defined above are
sevrey underestima1eO Meanwhile the distortionSoenerated by the increase in the marginal
personal tax rates designed to ralse the revenue foregone by the dividend exclusion are fully
aCCltlunted for A good measure of the costs of integration together With a poor evaluation of the
benefils may very well be the reason why partial integration is Simulated 10 yield neg alive
eHikiency gains
On a different vein as most of Ihe model surveyed here Pereiras is a closed economy model It
is legitimate to wonder how the resullS would change If international capital flows were anowed
IThis almost certainly would affect financial crowding out since a good d~al of the capital inflows
could be used to finance public debt
For these various reasons the results should be treated as preliminary but they strongly I suggosl that the dynamic structure provides valuable new insights into the corporate tax integration SSU2
40
7 Concluding remarks I I I i
What has been acccmplished Ihis far The success of Ihe CGE research in laelding the challenge
of dynamic modelling can not be denied
I iGreat progress in the modelling of household behavior has been achieved Promising
developm~nts in the modeUing of production and government behavior have also been accomplished
Interesting innovations in the concept of equilibrium and computation techniques were discussed
The policy analysis was greatly enriched by considering transitional effects together with longer
run steadymiddotstate equilibrium paths generalized equal yield strategies accomodating different
financial crowding out impacts and generalized dynamic policy evaluation indicators
An important set of issues have been addresed by the models surveyed Traditional issues
focusing on the effects of capital taxation and consumption taxes have been pursued In terms of
capital taxation for example the intertemporal nature of the issue was enhanced by allowing an
optimal evolution of capital stock in the economy and forward looking optimal investment decisions
In Andersson (1987) GouldermiddotSummers (1987) and Pereira (1986b 1987d) this is coupled
with a improved treatment of several tax provisions like investment tax credits and depreciation
allowances In terms of the consumption taxes developments in the specification of Intertemporal
household behavior the introduction of overlapping generations and the modelling of
intergenerational links via bequest motives as well as a closer attention to demographic evolution
provided a much improved economic setting for the understanding of the several aspects of the
problem
Furthermore dynamic modelling has permitted the addressing of a set of new issues Policy
Issues ranging from dynamic tax policy analysis to the evaluation of exogenous changes in
government expenditures to the impact of different methods of financing government expenditure to
4 1
the importance of tinanciai croHrii1g out to 1he relevance of consumorS expectations in tne policy
results have been addressed
Despile all of the progress and in part due to such progress several avenues are wide open for
much neded additional research The following seem to be the most Interesting and promising
areasmiddot First the specincation of liquidity constraints to household behavior may help to obtain a
better understanding of policy chenges that affect directly the interest rates in the economy Second
major attention needs to be paid to the incorporafion of a roraign sector (with open capital markets)
Third some efforts are needed in terms of finding adequate computation techniques now that
dynamic modelling eally makes the curse of dimensionality worse than eVOrbull
The modelling of financial markets is the single most unsatisfactory speet of the dynamic
models surveyed here~ The problems associated with provlding a mo~ adequate treatment of
financial markets and in general endogenous and optimal financial deci$ions bull both at the corporate
and at the personal levelsmiddot have 10 be addressed That necessarily requires Introducing uncertainty
into the models To b adequate this must go well beyond the mere assumption of point expeetalions
with the whole array of modelling implementalion and evaluation problems thai generates
bull I i
42
REFERENCES
Abel A and O Blanchard 1983 An Intertemporal Model of Saving and Interest Econometrica
51 pp 675-92
~ Andersson K 1987 Tar-ation of Capital in Sweden - A General Equilibrium Model Ph D
Dissenation UnIversity of Lund Sweder
Auerbach A L KotlikoH 1983 National Savings Economic Welfare and the Structure of
Taxation in M Feldstein ed Behavioral Simulation Methods in Tax Policy- Analysis Chicago
University of Chicago Press
Auerbach A l Kotlikoff 1984 Social Security and the Economics of Demographic-
Transition in H Aaron and G Burtless eds Retirement and Economic Behavior Washington DC
Brookings Institution
Auerbach A L Kotlikoff 1987 Dynamic Fiscal Policy Cambridge Cambridge University
Press
Auerbach A L Kotlikof and J Skinner 1983 The Efficiency Gains from Dynamic Tax
Reform Internatiooaj Economic Reyiew 24 pp 81middot100
Ballard C 1983 Evaluation of the Consumption Tax with Dynamic General Equilibrium
Models Ph D Dissertation Stanford University
Ballard C 1987 Tax Policy and Consumer Foresight A General Equilibrium Simulation
Study Economic Inquiry 25 pp 267-284
Ballard C and L Goulder 1985 Consumption Taxes ForeSight and Welfare a Computatonal
General Equilibrium Analysis in J Piggott and J Whalley eds New Deyelooments in Apolied
General Equilibrium Analysis Cambridge Cambridge University Press
Ballard C D Fullerton J Shoven and J Whalley 1985 A General Equilibrium Model for
t 43
Tax Eoliey Eyalua~iQO NBER University or Chicago Press
Borges A (1987) bull A Survey of Compula1ional General Equilibnum Models CECD Economic
Slud~ 8
Bovenbig L 19B3 imperfect Capital Mobility in a Perfect Foresight Herberger Model
Mimea UC Br~eley
Bovenberg L 1984 Capitel Accumulation and Capital Immobility Qmiddottheory In a Dynamic General Equilibrium Framework PhD dissertation University of California at Berkeley t
Bovenberg L 19850 Dynamic General Equilibrium Tax Models with Adjustment Costs in A iManne ed bull 1985 EtQnomrc ECUjilbrium Model Formulation and Solution Mathematical
lProgramming Siudy 23 Arnsterdam NorthmiddotHoliand
Bovenberg L t 1986 Capital Income Taxation in Growing Open Eeondmies Journa of Pubtic
Eamp(omlcs 3t pp 347middot376 IBovenberg L ald W Keller 1981 Dynamics and Nonlinearlties in Applied General
Equilibrium Models Internal Report Mlmiddot81middot208 Department for Statistical Methods CBS bull
Vooiburg The Netherlands
Charnley C 1981 The Welfare Cost of C~pltal income Taxation In a Growing Economy
IJournal of Political ECQnQm~ 89 3 pp 468middot496 I
Charnley C 19810 Efficient Tax Reform in a Dynamic Model of General Equilibrium I Mimeo Yale University I Charnley C 1982 On the Optimal Taxation In Stylized Models 01 Inlartemporal General
Equilibrium Mimeo Yale University I Decaluw B and A Manens 1985 Pays en Oeveloppement at Modeles Calculables OEquilibre
General - Une Revue de al litterature empirique Monographie Sene 9 Volume 3 Centre de
Recherche et Developpemont Economlque Unrslt de Montreal
44
bull
I I
I
I
1
Erlich $ V Ginsburgh and L van der Heyden 1987 Wher~ do Real Wage Policies Lead
Belgium bull A General Equilibrium Analysis European Economic Reyiew (forthcoming)
air R and J Taylor 1983 Solution and Maximum Likelihood Estimation of Dynamic
Nonlinear Ratioant Expeclations Models Ecooometrica 51 pp 1169middot1186
Feltenstein A 1984 Money and Bonds In a Disaggregated Ec~nonYin Searl 1-1 J Shoven
Eds Aoolied ~eoeral EqJiliDCiulD tioalYsis Cambridge University Press Cambridge
Feltenstein A 19B6 A Dynamic General Equllibrlu Analysis of Financial Crowding Out
Theory with an Application to Australia Journal of Public Economics 31 No1
Fullerlon D R Gorden 1983 A Reexamination of Tax Distorllons in General Equilibrium
Models in M Feldstein Ed Behayora Simulation Methods in Tax pOlicy Analysis Chicago
University of Chipgo Press
Fvllerlon D J Shaven and J Whalley 1983 Replacing the US Income Tax with a
Progressive Consumption Tax JoulOal of Public ECQnomics 20 pp3middot23
Fullerton D Y Henderson J Shaven 1984 A COmparision of Methodologies in Empirical
General Equilibrium Models of Taxation in Searl H J Shoven Eds ~pQlied Geoaral Jguilibrium
lIoalyss Cambridge University Press Cambridge
Fullerton 0 A King J Shaven and J Whaney 1981 Corporate Tax Integration in the United
Statesmiddot a General Equiiibdufi1 Approach American Economic Review 71 pp~ 677~691
Galper H R Lucke E Tader 1985 Taxation portlor Choice and the Allocation of Capital A
General Equilibrium Approach Mimeo
Gill p W Murray M Saunders M Wright 1986 Users Guide for NPSOL (Version 40) A
Fortran Package for Nonlinear Programming T echnicsl Report SOL 86middot2 Deparlment of
Operations Research Stanford University
Glnsbur~h V and L van der Heyden 1985 General Equilibrium with Wage Rigidities An
45
AppJication to Belgium in A Manne ed t Economic EQumprit1m~ Made Formulation and $oJioo
Mathematical Programming Study 23 Amsterdam NorthmiddotHolland ~
Goulet J 1968 1 Adjustrent CO$1S in the Theory o(the Firm Reyiew of EconomiC Studies 35 bull
pp47middot55
Goulder L 1985 Tax-Financed versus BondmiddotFinancedGhanges in Governent Spending
bull Mimeo Harva(d UniversilYmiddot
Goulder l J Shaven and J Whalley 1983 Domestic Tax Policy and the Foreign Sector The
importance of altemative foreign policy formulations to results from a general equilibrium tax
analysis model In ~n methods in Ibe antico of income from caoia1 ed Martin
Feldstein Chicago University of Chicago riSS
Goulder l L Summers 1987 Tax Policy Asset Prices and 3rowtly ~
A General Equilibrium Analysis NBER Working Pper No 212B
Grandmont J 1982 Temporary General Equilibrium Theory Ch 20 in Handbook 01
Mathemalical Eccnornics Vol II NorthmiddotHoliand
Harberger A 1959 The Corporate Income Tax An empirical Appraisal In lax Revision
Comoendjum Voll House Cornmitee on Ways and Means Washington DC Govemment Printing
Office
Harberger A 1962 The Incidence of the Corporate Income Tax JOULn1 of Political
Economy 70 pp 215middot240
Harberger A 1964 Taxation Resource Aliocalrlon and Welfare fn lb Role of Dir bullbulll and
Indjresl Taxes io tbe Fede[al Reserve s~I~Ill Princeton Princeton Unlveresity Press
James J 1985 New Developments in the Applications of General Equilibrium models to
Economic History in Piggott aand Whalley Eds
Jorgenson D and K Yun 1984 Tax Policy and Capital Allocation Harvard Inslitute of
bull
I
I f
Economic Research WP 1107
Judd K 1965 ShonmiddotRun Analysis of Fiscal Policy in a Simple Perfect Foresight Model
Journal of poJiHcal 5CCJiW1l pp 298middot319 r
Upton D J Potbc J Sachsa nd L Summers 1982 Multiple Shooting in Rational
Expectations Models fuoometrlc 50 1329middot1333
Manne A ed 1985 fconQrnjc Eoullibrium Model FOrIDujaljQo and Solution Mathematical
Pogrammlng Study 23 Amsterdam NorthmiddotHolland
Manne A 1985 Or the Formulation and Solution 01 Economic Equilibrium Models In A
Manne ed ECQOQlli-Jr~ibmiddotum ~tQdel EQlJulailon and Solution Malhematical Programming
Study 23 Amsterdam NorthmiddotHolland
Mansur A and J Whalley 1984 Numerical Specification of Applied General Equilibrium
Models in Scarf H J Shoven Eds Aoplied General Eouillbrium AnaiysectIsect Cambridge Universily
Press Cambridge
Merrill 0 1972 Applications and Extensions of an Algorithm thaI Computes FixedmiddotPoints of
Cenain Upper Semimiddotcominuou Point to Set Mappings PhD Disserlation Unlvel$ity 01 Michigan
Negishi T 1960 Welfar Economics and Exislence I an Equnlbrlum for a Competitive
Economy MelroeCOOQ11ka 12 92middot97
Pereira A 1985 On the Existence and Uniqueness of Optimal Ou1put Path lor CRTS Firms in
Adjustment Costs Technologies Mimeo Stanford University
Pereira A 1985a Consumer Expectations and Ihe COSI 01 UllUty In an Intertemporat
Framework Mlmeo Stanford University
Pereira A 1985b Corporale Tax Integrallon and Government Deficits A Dynamic General
Equilibrium Anaiysls Mlmeo Stanford University
Pereir A 1987 On Ihe Computation of the Users Costs of Stocks Ecooomi (forthcoming)
47
Pereira A bull 1987b Equal Yield Alternatives and Government Oeficlts Mimeof Stanford
University
Pereira A 1987c A Dynamic Applied General ~uHlbrium Model for Tax Policy Evaluation
Comlletamp Docurnent2tion Mimeo University of California San Diego (in progress)
Pereira A 1987d Should the investment Tax Credit be Rei(ltroduced Mlmeo University
of California San Diego (ir progreSs)
Piggott J and J Whalley eds 1985 tJew QeyeQpments in Applied General EQuilibrium
bullenalysjs Cambridge Cambridge University Press
Preckel Pbull 1985 Alternative algorithms for computing economic equilibria In A Manne Ed
Radnor R 19amp3 Equllibrium under Uncertainty In K Arrow M Intrilligaor ods
I ibull
Robinson S 1 g86 Mullisectorai Models of Oeveloping CountrieS a Survey Oivision of
Agricullure and Resources UC Borkeley WP 401
Scarf H 1967 The approximation of fixed points of a continuous mapping ~MM Jouwal of
epplid Mathornallcs 15 pp 1328middot43
Scarf H with T Hansen 1973 00 the Computation of EconomiC Eguilibria New Haven Yale
University Press
Scarf H J Shoven Eds 1984 61l1llied Genera Elujlibrium 6nalyssect Cambridge University
~ Press Cambridge
Shoven J 1976 The Incidence and Efficiency Effects of Taxe on Income from Capital
Jurnal of ramie1 Economy 84 pp 1261-83
Shoven J 1983 Applied General Equilibrium Tax Modemng IMF Slaff PaoerS 30
Shoven J and LB Simon 1986 Share Repurchases and Acquisitions An Analysis of which
Firms Participate NBER Working Paper No 2243
bull
48
IShaven J and J Whalley 1972 ft General Equmbrium Calculation of lhe Effects of
Differential Taxation of Incoma from Capital in the US Jauroe of public Econgmics 1
pp281middot321 (
Shoven J bull and J Whalley 1973 General Equilibrium with Tax$s a computation procedure
and an existence pr00f ReviEhi of Economic Studies 40 pp 475middot490
Shoven J J Whalley 1977 Equal Yield Tax Alternatives A General Equilibrium
Computa~ui0nal Tfchnique J~HJrnal of PUblic EconomIcs 8 pp 211~224 bull
Shaven J J Whalley 1984 Applied GeneralmiddotEquilibrlum MOdels of Taxalion and
International Tr8de An Introduction and Survey Journal of Economic literature 23t pp
1007middot1051
Slemrod J 1980 A General Equilibrium Model of Capital Income Taxation Ph D bullDissertation Harvard Univnrsity
Slemrod J 1983 A General Equilibrium Model of Taxation with Endogenous Financial
Behavior in M Feldstein edbull Behavioral Simulation Methods In Tax Policy Analysis Chicago
UniVerSity of Chicago Press
Stone J 1985 Sequential Optimization and Complementarily Techniques for computing
Economic equilibrium in A Manne Ed
Summers L 1981 Taxation and Corporate Investment a q-Theory Approach BroQkinas
Summers L 1985 Taxation and the Size and ComposHion 01 the Capital Stock An Asset Price
Approach Mlmeo Harvard University
Whalley J 1975 A Geneal Equiloibrium Assessment of the 1973 United Kingdom tax
reform Economica 42 pp 139middot61
1JLTIllOS JIQll~)ltG PAPERS PUBLICADOS bull
I
I I
j
112 70 - LUCENA iJiCIIJo ~ IIEnvironrilent Honitoring and Orqanization Structure T (Setellibro 1987)
nI 71 - L~ENA Diogo A Note on the Representation of Information Stxueture (SeteJObro 1987)
nQ 72 - LUCENA D1CgO Envuorment 111)nitoring and Organization
Structure II (Setelilhro 1987)
ril 73 - AlmmE3 Antonio Pais e CAsPAR Yitor bull Tributacao Inoentivos e InvesUlllentos Analise Qualitativa noYellibro 19S7)
n2 74 - liAlltOSA rose Pedro Opt1lllal lIage R1gidity A SUggestedtlthodology to Teet tb Theory and an Application (Outubro 1987)
nI 75 - CABRAL L 11 B Three lIotes on SYJIIlletric Gemes witt itsYJIIlletric Equilibri (NoveJObro 1987)
n2 76 - LUCENlt D10gOTo Search or Not to Search (Fevereiro1967)
nQ 17 - SA Jorge asconcelos euroI A Theory of SynergyB (Fevereiro1987)
n2 76 - VILARES tanuei Jose Os Be IntenaMios IlIIportados ColIIO Factor de Produao (Julho 1987)
nQ 79 - SA J~rge VasCQncel03 e HoY to Copete and Co~unicate in tature Industrial Products_ (aio 1988)
n2 80 - ROIl Rafael Learning and Capacity Expansion in a New larket Under Uncertainty (Fevereiro 1988)
n2 81 PEREIRA Alfredo llarvilo Survey of Dynamic Computational General EquilibriUlll Hodel for Tax Policy Evaluation (Outubro 1987)
Qualquer in ormacao sabre o 1IIorlltlng Papers ja publlcados sera prestada pelo Secretariado de Apoio aos Docentes podendo os ~e3OSf
ser adquirido5 ne Secyo de Vlndas da Faculdade de EconollLia UNL na Travessa Esteviio Pinto Cltpolide - 1000 LISllOA
In GouJder (19851 the 9Vernmenl maximizes a static social welfare ~nctio subject to a-
bull
20 I i
I
I I I
I bull
balanced budget conSirtlL This follows the optimal alocatkm of government expenditures alpng
the lines of Baliard-FHeiton-Shoven-Whalley (1985) Ttle model Is generallzea to allow for
exogenous changes in HIe time path of government expenditures Two financing alternatives are
consjdeft~d and contr8s1ed EldditionaJ tax revenues and bond issuance
Pereira (198Gb 1137d) auempts to address both the incorporation of deficits and the
determination of govement expenditures The path of government expenditures and the path of
delicilSlsurpluses (ano therefore the path lor debt) are endogenously and optimally determined The
government is seen as mltJximizing n intertemporal social welfare function given the 13x structure
Optimization is subject to a sequence of recursive equations of motion reflecting the evolution of the
ptlbllc debt allowing fDr government budget imbalances It should be stressathat this specification
01 the government cOMtcalnt is equivalent to the specificaUon in Auerbaeh-Kotlikoff (1983 1987)
when no liquidity constraints affect -gove~nment behavior
The extra richness in the treatment of government behavior allows tfte examination of
government debt poices in a truly dynamic setting Also financial crowding out effects induced by
government deficits can be analyzed (see Auerbach-Kottlikoff (1987) Feltenstein (1986) and
Perer (198Gb 1987e))
34 Foreign sector
Most of the open economy models follow the assumptions of balanced trade with import and
export net demands characterized by constant elasticities along the line of Ballardmiddot
-Fullerton-Shaven-Whalley (1965) Such is the case of Salfard (1963) Ballard-Goulder
(19S5) and Goulder-Sllmmers (1987)_ In Feltenstin (1985) the rest of the world is treated as
21
an addjlion~ consumer group
Bovcoberg (1986) develops a model In which two economies lIr considered each following
intertempoml perlect foresight paths These economies meet in the international forum Their
trade relaroMhips are characterized by yearly balanced trade accounts
None of the new generation of dynamic cae tax models has yet IncorPorated the lnternational
capital flows as done in aoulder-Shoven-Whalley (1983) for the earlier Ballarrshy
-Fullertonmiddot Shaven-Whalley (1985) model This is unfortunate as there is an important
Intertmporal aspect to international lending The first al1empts along Ihese lines are due to
Andersson (1987) and Erlich-Glnsburgh-Heyden (1987) Andersson (1987) in his model of the
Swedish economy adopts a closed economy appraoch in which rates of return in the domesJic
economy are largely determined by the international capital markets Fidinterest rate induce
Intlnatlonal capital flows which determine and finance the International lrade imbalance In
lurn In ErlichGinsburghHeydenmiddot (1987) foreign trade Is generaled according to an
intertemporal Irade welfare lunction with constant import and export elasticities In the shortmiddotrun
they allow inlernational trade inbalances which generate capilal ftows to the domestic households In
the long run however trade balance is assumed
35 The need for financial markets
A dynamic economic Wucture not only provides the ideal environment 10 model many features of
economic behavior it also permits one to Incorporate the financial side of the economy If the
government is allowed to run deficits the question 01 deficll financing automatically follows If
investment is optimally dewmined and returns to capital are different aClOSS sectors problems of
investment financing arise If there are seeral final1cfal assets In the economymiddot government
4-
I i
ibull
I I I I
I
j
22
bonds privpte bonds and equity (or simply physical capital installed inmiddot different sectors) the
problem of allocation of saving among assets with potentially different returns arises
The papers surveyed var greatly with respect to l~e extent of their attention to the financial
side of the economy At Clne extreme are the models in ~anardmiddotFuliertonmiddotShoven-Whalley (1985)
Andersson (1987) Ballltld (1983) Ballard-Goulder (1985) Bovenberg (1985 1986) and
Erlich-Gjnsburgh-Heyden C19B which are devoted exclusively to the real side of the economy
In turn Auerbach-KotliKoff (1984 1987) Feltenstein (1984 1986) and Goulde (1985)
allow for government debt In these models saving finances changes in government debt and physical
capital Private and publlc 2ssets are perceived by the households as perfect substitutes The
allocation of saving merely adjusts to the relative demands for funds
Feltenstein (1984 1986) is the only model surveyed which introduces money Government deficits are financed by issuing money and bonds according to an exogenously given rule Money is
demanded by consumers for transaction motives and an exogenously given fraction as a store of value
On the other hand government bonds and physical capital are the vehicles for the intertemporal
transfer of wealth
Summers (1985) and GOlldermiddotSummers (1987) introduce a whole menu of financial assets shy
firm specific equity capital Different assets earn different rates of retum However such rates
are equal up to constant and exogenous sector speCific risk premia Therefore the introduction of
constant exogenous risk premia as helpful as it may be in the context of calibration does not solve
the main issue of the non-optimality of the allocation of saving Also talking about risk premia in a
deterministic context is somewhat unsatisfactory
Pereira (1986b 1987d) also introduces a whole menu of assets private and public bonds and
firm specific equity However all the assets are expected to yield the same rate of return and
therefore perc~jved as perfect substitutes The different asset types allow consideration of
23
exogenOU$~8tUequjty triO dividendlrc~enlion rufes and therefore seve-ral sources of investment
financing bond equity and retained earnings
The nonmiddotcptimalHy of the allocation of saving and the absence or exogenelty of corporate
financial rules is _a reftectlon of the limitations of the determInistic approach Either in a context of
perfect anticipation oi the future prices Of in general within the t$atm of point price
expectations all the papers follow a deterministic approach Under such circumstances consumers
either expect diHerent rales of return (inclusive of risk premium) across assets in which case
they wilt buy only one asset (that with highest rate) Of they expect equal rat~s of relurn in which
case Ihey are indifferent about the asset composition of Iheir portfoliO There is no way of
Iradingmiddotoff rales of return and risks to obtain an optimal interiof solution to the problem of the
allocation of saving
The most advanced contribution in the modelling of saving allocation in a CGE setting is due to
Slemrod (1980 1983) in the context of a static one-period model In his model consumers act
according to a two stage separable decision process They Hrsl decide on how much to save Then
they decide on the allocation of saving according to an Indirect uttlity function dependent on the rates
of ratum and variances offered by the different assets in the eoonomy The sourCe of riskiness in the
ampConory comes from an uncertain marginal product of capital On the other hand aside from
portfolio decisions the res of the economy is insulated from uncertainty
The most complete contribution in terms of the treatment of the corporate financial rules is
FulienonmiddotGordon (1983) In a variant 01 Ihe BailardmiddotFulienon-ShovenmiddotWhelley (1985) model
Fullerton and Gordon have capital intensity and optimal financial decisions Jointly deter~ined
through a two siage process The cosl of financing capital is minimized by trading off the tax
advantages of debt against the epected real bankrupcy coSls inherenl with high debVequity ratios
Given the optimal debtequity ratio the level of inveSlment Is chosen such that at Ihe margin the
24 l
bull
rerurn pn ~uity equas the return on bonds plus an exogenous risk premium
36 Market assumptions equilibrium and expectations
Virtually all of the papers arc characterized by Walrasian market clearing assumptions All
markets are perlect1y competi1ive Atomistic competition among agents 1s assumed even thoughmiddot only
a flnile number of agents is considered Virtually no mark~t disequilibria or price stickiness afe
considered The only exception is ErlichmiddotGinsburghmiddotHeyden (1987) bull In thei model of the Belgium
economy I the _wage rate is fixed in the short-run Therefore in the shOrlwrun desequilibrium in
the labor market will generate endogenous unemployment However in lh$ long-run all prices
Including the wage rata are Ilexible and accordingly ali markets ciear
Given the dynamic nature of behavlor in the economy markelmiddotcfearing prices in each period
depend on expectations of fture prices and on tax variables In the economy There are essentially
two ways at interpreting the economic equilibrium in such a dynamic conte~bull If future prices are
perfectly anticipated (io expectations are self-fulfilling) a perfect foresight equilibrium
prevaHs Then future actions are merely the implementation of current decisions for future
periods However jf ptica expecla110ns are not perfect (Le agents make mistakes w1th re$pect 10
future prices) then a temporary or shortrun equilibrium prevails Markets ciear and clearing
prices depend on fu1ure price expectations Current plans about the future are typically not
precisely implemented They will be revised as more or better information becomes available to the
economic agents See Grandmant (1982) for a comprehensive survey 01 the temporary equilibrium
literature
Wilh the exception of Gould (1985) and Pereira (19S6b 19S7d) all the models surveyed
adopt th~ concept of a perfect foresight equilibrium In turn BaliardmiddotGouldr (1965) considers a
J
2S
flexible am9unt of foresight In terms of the number of years over which price movements are ~
foreseen Pereiras model (1986b 1987b)) is flexible in a somewhat different way in that it can
indude any range of foresight from myopia to perfect foresight The choice between the perfct foresight and lemporary equilibrium is ultimately to be made on
philosophic grounds It can be argued that less than perfect expectations Imply that agents are
irralional in some way (see Auerbach-Kctiikoff (1987 p 10) However the reverse argument can
be made One can question wiether agents are reaHy rationar and perfectly knowledgable about
future prices
Recent evidene of Balird (1987) Ballard-Goulder (19851 Goulder (19851 and Pereira
(1986b 1987d) suggests that the choice in modemng expectations is an important one They show
that the degree of foresight inl0 thc future (ranging from perfect foresight to-myopic expectations)
may have dramatic impacts on the pollcy conclusions of Ihe model Accordingly the best research
strategy may b to design models which are flexible enough 10 allow for different rules regarding the
formation of expectations With the exception of the articles just mentioned sensitivity analysis
bullhave previously not been performed along this dimension
In terms of implementation the two concepts of equilibrium - perfect foresight and temporary
equilibrium have different implications The dimensionality of the equ1ibrium soll1ion algorIthm
is involved Suppose we have a model with ten markets to be run for a period of 50 years Aside
from normalization a perfect foresight model implies computing prices In 500 dimensions while a
-t~mporary equilibrium model requires sorving SO equilibria each in ten dimensions Given that
computational speed often varies with the cube of the number of dimensions the lemporary
equilibrium formulation is potentially strikingly more feasible However Ballard (1987) and
Goulder-Ballard (1985) have devoloped techniques to greatly speed the computation of a perfect
foresight equilibrium
26
j
I
4 Jrnplemenlalion and Issues on policy evaluallon
41 Calibration and equilibrium comparlslon
CGE models at typically parameterized by the use of a calibration procedure Some parameters
are exogenously given However some crucial parameters are determined in such a way that the
model replicates the data for a given base year See MansurmiddotWhalleY 11984) for an eXlensive
discussion of 1~i$ issue
CalibratIon in a dynamic context is generaliy interpreted as requiring two properties First
replication of base year data is required Second the model is parameterized to simulate an
intertemporal balanced growth path when Ihe base policy is maintaineq This Is the approach
followed by Ballard (1983) BallardmiddotGoulder (1985) Goulder (1985) and SummersmiddotGoulde
(1986) It follows the practice of BallardmiddotFuliertonmiddotShovenmiddotWhalley (1985) with their
sequential cquilibrium dynamics
Other authors Auerbach-Kotlikof (1983 1984 1987) Boyenberg (1984 1985 1986)
and Pereira (198Gb 1987d) follow whal we cali a qualitalive calibration The slructural
palametars are exogenously chosen so that lh economy follows a reasonable path into Ihe fulure
ThaI has 10 do wllh the fact thaI given the recursive nature 01 the dynamic economy and aside tom
such structural parameters only iniUa stock values are needed to run these models Given initial
conditions on the stocks of say private wealth capltal and government debt agents will optimize
and Ihereby generate a lirst round of net demands and equilibrium conditions In lurn the
equilibrium prices will determine the evolution of the stock variables into the next perIod
There are several potential problems with calibration in the conlext of dynamic models The
assumption of a steadymiddotstate growth ~ath in the base case can be questioned First while
27 I Ibullbull
steadymiddotstaU is a possibity it certainly is not the only meaningfull solution to dynamic models
Even in the case of a perfect foresight equilibrium the model implies an equilibrium path which i may or may not involve colanced growth The model not the modeller should dictate the nature 01 I themiddotbase case path Second in the context of a temporarY equilibrium path a steady state solutlcn is Inol a likely model QutcOn1tl In fact unlessmiddot expectations are stattc short run behavior consistent
with a steady-state evolution will in general not be generated On the other hand if static I expectations are self~fufjili[1g we have in fact a perfect foresight model Third even the base year
replication requirement may cause problems In Ihe cOntexl of temporary equilibrium Any
calibration parameter would be conditional on expectation rules which is probably an undesirable
feature
The nalur of the two-requirement calibration strategy is very muchin the spirit of the bull
traditional design of the comparision of alternative equilibria comparis_ion between a steady-state
base case on one hand and alternative paths including a transition period and a finat sleadyyenstate on
the other hand Pereira (198Gb 1987d) compare different (not necessarily steadystate)
equilibrium paths The arguments against such a procedure are based on the idea that the impact of
policy changes can be observed most easily since all departures from the steadymiddotstate can be
attributed to the altemalive polioy On the other hand the base case so dEfined as a steadymiddotstate is
consist~nl with previous work in a less dynamtc setting and Iherefore allows a common standard
for comparing model results
42 Computation algorithms
We are still al a stage in which basically each author uses a different computational technique
Th development of dynamic models - in parlicular wilh adjustment costs andior perfect foresight
bull
28
has corresponded with the decline in the use of fixed point algorithms In -fact given the relative
arge dimensions inevitably involved such algorithms lend to be very inefficient at the best and
chen prohibitively stow See Stone (1985) and Preckel (1965) for a comparative assessment of
different computation techniques Among the models surveyed only Ball1rdFulle(ton~
-ShavenmiddotWhalley (1985) and Feltensteln 1985 use Merrills variant of the fixed point
algorithm 1echnique
AuerbachmiddotKoHikoff (1963 1984 1967) follow a three stage procedure They first compute a
base case steady-state then a revised case steady state and finally transition path for the economy
berveen these two steady-states tn all stages a Gauss-Seidel iterative procedure is used
Ballard (1982) Ballard-Goulder (1965) Goulder (HiSS) and Summers-Goulder (1986)
~
use a method developed by Ballard and Goutder which is similar in many awects to the FairmiddotTaylQr bull
(1983) algorl1hm Short-run equilibria are calculated (using Merrills algorithm) parametric on
nrice expeC~lions The model is then iterated to generate self-fulfilling intertemporaf expectations
and the corresponding perfect foresight equilibrium In a relatively similar approach Andersson
1987 uses a simulation program SIMNON developed In the University of Lund This program
can handle two-paint boundary problems In a fashion consistent with the multiple shooting
algcrlthm (see Lipton-Poterba-Sachs-Eummers (1982))
Boyenbergs computational approach (1985 1985) differs from the other models surveyed in
that he relies heavily on analytical techniques Computations are done by using a dynamic version of
Johansons linearization method Being essentially determined by the continuous time nature of the
I I
f
model thiS linearization model has the disadvantage of confining the analysls to infinltesinal changes
around the base case equilibrium (see Bovenberg (1985) p 53) Pereira (1986b 1987d) uses an optimization algorithm NPSOL - developped by
rGillMurray-Saunders-Wrlght (1986)) This algorithm is used 10 compute the sequence of
I
29
Shortmiddotrun temporary equilibrium which makes up the intertemporal equilibrium path The
equilibrium conditions are seen as nonlinear equality constraints in the minimization of an artlcial
objective function The prices are normalized to the unit simplex by an additional linear equality I I 1 bull constraint
II ErlichmiddotGinsburgh-Heyden (1987) follow a unique approach in that they use a variant of the
optimization technique introduced by Negishl (1960) The economic equilibrium can be generated i
as a solution of a mathemallcal program Ihe objective function of which is a weighted sum of the
utility functions of the various agents while tho constraints set consists of the market dearing 1
conditions Ginsburgh-Heyden (1985) have extended Negishis resulllO tho case of downward price
lrigidities I
Finally the paper by Jorgonson-Yun (1984) is also unique in tIlat it is the only I
econometrically estimated model Different blocks for the consumption and production Side of the
-
model are separately estimated to provide the necessary structural parameters i
I The diversity of computation techniques is yet another indicator of the exploratory nature of the
body of literature surveyed in this article
43 Equal yield comparisions
The link between a base case and counteriactual simulations Is usually provided by the concept of
equal yield Shaven-Whaley (1977) discuss the meaning 01 equal yield in a general equilibrium
context when government is not allowed to run defICits Equal yield is interpreted to moan constant
public utility Government base case utility is maintained in the counferfactual experiments With
balanced budgets the concepl of equal yield is unambiguous The new equUibrium prices and the
balanced budge condition will del ermine the minimal expendilure and taxes needed to maintain base
30
case publi~ utility Accordingly in general equal yield is inconsistent with equal nominal tax
revenue Some change in tax revenue Is necessary Different tax replacement schemes ate
considered to assure that enough tax revenue is collected This is the approach essentially followed
by most 01 the papers surveyed Only Feltenstein (1985) Goulder (198) and Jorgenson-Yun
(1984) chose not to follow an equal yield strategy
The question is of how to Interprete lila concept of equal yield when the government is alowed to
fUn deflcLts The optimal leve of expenditure for base ease public trliUty can now be taAinanced
bond financed or financed by a mix of bonds and taxation We have several versions of equal yield In
particular equal yield may now be consistent with equal tax revenUE Furthermore some meaSure
ltgtf financial crowding out effects of government deficns can be inferred from the comparislon of the
several equal yield alternatives These issues are extensively discussed ~Pereira (1987b)
44 Price expectations and the dynamic generalization of compensation
Indicators
The sum of equivalent and compensation variations over households is the most widely used
aggregated measure of efficiency gains or losses In the context of $teady~state eomparisions andor
when complete future markets exist andlor perfect foresighl is assumed there are no difficulties
associated with the use of the standard Hicksian indicators In fact correct future prices are known
in these cases
If expectations are not selfmiddot fulfilling andlor future markels are not available a dynamic
generalization is necessary 8allardmiddotGoulder (1985) provide some steps in that direction by
defining an indicator that accomodates periods far into the future when households do not have
perfect foresight but a steady~state prevails On he other hand this issue is more fundamental in
i
[
I
i I I I I [
I I
bullI
31
the contex~ of a Gner1 lemporary equilibrium ftamawor~ In such Circumstances Pereira
(1986a) dev~lops a cnamlc generalization which is obtained as the present discounted value of a i i
sequence of short-fUn optimal expenditure functions consistent with a base case expected future I stream of IlWities I
32
5 Empirical evidenee from selected policy Issues
Dynamic tax models have generated several important results which escaped Sialic modellers
The following is a seleCpoundd set of issuos whIch stress the marginal benefits of dynam~ modelHng of
economic behavior in innovltlttive areas of analysis The discussion of a consumption tax emphasizes
the benefits of dynamic tusehold behavior and intergenerational aspects The study of the
flliminatlon or the re~intToduction of the investsnemt tax credits is made meaningful by the dynamic
modelling of production betlavor Modelling governmenl in a dynamic conlexts allows the modeller
to study the irrrpact of government deflcits Finally a dynamic framework lets us appreCiate the I
relevance of consumer expectations in terms of the evaluation of policy alternatives
i
i
51 Consumption Tax
t
By Its very nature a consumption tax can not be adequately investigated with a static model bull
Fullerion-Shoven-Whaicy (1983) use Ihe model of the US economy s described in
Ballard-Fullerton-Shoven-Whalley (1935) to evaluate the movement from the currenl US tax Isys1em to a progressive consumption tax Since their modeJ Incorporates a Jaborleisure choice
where leisure is an untaxed commodlty their results reflect the fact that both the consumption tax
and Ihe present system are dslortionary
Concentrating on the rntertemporaf distortions they show that sheltering more saving from lhe
current US income tax could improve econornic efficiency even if marginal tax rate increases are
necessary in order to maintain government fevenue~ At first you have a revenue shortfall
However the economy moves to a higher sustatned growth path and uitimatefy lower tax rates
cangenerate the same revenue path Also wages increase in the long run with this policy The
bull
bullbull ~
33 shyt
i
I
I
present value of welfare gi[lS for a potiey of complete savin9 deduction wilh marginal tate
adjustments (consumption lax) is simulated to be around $500 billion to $600 billion 01 1973
dollars
They also investiga1e th 1ngth of ttme it takes th~ economy to adjust to these policy changes
Roughly they estimate the long run to b thirty years although this figure Is very sensitive to the
specification of the savings elasticity - a crucial parameler in this model
52 Investment T~x Credit
Goulder-$ummers (1 S87) address the impact of eliminating the lnvestmenl Tax Cradit (ITC) on
intersectoral capital formation and on economic growth Their model isYa1ticularly adequate to
addies$ this issu~ in that is postulates a forward looking investment m9we with adjustment costs
They show that the eliminating the ITC causes a reduction in the rale of investment Inveslment is
estimated to fall by about 7 in the short-run and by about 12 in Ihe new long run steady state
In lurn a previous policy anncuncement lowers Ihe overall al1ractiveness of Investment and leads 10
a downward shift of the Investrrent profile On the other hand the combined effect of a revenue
neutral simultaneous efimiraron of the ITC and reduction of the corporate tax rates is a long run
reduction ef the capital Sieck by 35 This pattern suggests that a revenue neutral increament in
both Ihe corporale tax rates and the inveSlment lax credits would be preferable 10 the formula
adopted in Ihe US Tax Reform Act of 1986
Pereira (1987d) addresses the impact of the ITe from Ihe oland point of Ihe curTenllax syslem
under the Tax Reform Act of 1986 Given Ihe concern over the potential depressiVe effects upon
savings and investment of the new tax system in conjunction with deficit reduC1ion mechanisms of
the GrammmiddotRudman type a pertinent question is should Ih lTC be e-introduced In the context
bull
34
-of his mod~1 of the US economy Pereira focuses on the trade-off between the distortion in the
intersectoral allocation of capital induced by the lTC the lower tax revenues it generates and
potential financial crowding-out effects on one hand and the positive effects of lowering the relative
price of new capital goods on the other hand Preliminary results confirm the qualitative results in
Goulder-Summers (1987) suggestting a welfare gain of about 2 of the present value of GNP in
the best scenario of absence of replacement taxes Furthermore preliminary results show an
important time pattern to welfare gains with the average benefits increasing the further you look
into the future This pattern is due to the presence of constraints to intersectoral mobility of capital
and inertia in the adjustment towards the optimal capital levels as reflected by the presence of
adjustment costs
I
I
bullAuerbach-Katlikaff (1gB) in the context of their intergenerational model of the US economy
consider the impact on savings capital formation and interest rates of deficit policies and balanced
budget increases in government consumption They conclude that deficit finance and government
consumption can significa-ntly crowd out capital formation and lower the welfare of future
generations However crowding out from deficit finance is a very slow process because it results
from increased government spending over potentially long horizons Also deficit policies may
substantially influence the longmiddotrun interest rates while leaving the short-run rates essentially
unaffected Finally and in opposition to the central role adjustment costs seem to play in both
53 Financial crowding outeff~cts of government deficits
Goulder-Summers and Pereiras models Auerbach-Kotlikoff suggest that the time path of interest
rates induced by a policy of deficit financing seems to be insensitive to the presence of adjustment
costs
bull
35
54 The role 01 consumer expectatlons
The expectations analysis has uncovered one of the benefits of dynamic modelling
Ballard-Goulder (1985) find that the appeal of adopting a consumption tax depends on the level of
foresight possessed by the consumers Furthermore additional foresight may be welfare worsening
This is a second best type of result This tends 10 occur under policies that lead to capital deepening
and declining rate of return to capital over time To the extent that consumers have more foresight
they will be beller equippod to anticipate the fall in the rental price of capital Consequently if a
saving incentive is enacted people save more with myopia than with perfect foresight Given the
exi~tence of taxes on capita and the discrepancy between the private an~--sccial returns to capital
the greater savings with myopia is beller socially Ballard-Goulder find that the welfare gains
from a consumption tax are reduced by about 10 when we move from myopia to farleet foresight
Therefore the attractivenness of adopting a consumption tax seems fairly [obust across these
foresight specifications
I
bull
36
6 Tbe power and weaknesses 01 dynamic modelling the example 01 corporale
tax Integrallon
The Issue of corporate tax integration allows us to sh~w the stre-ngths and weaknesses or the
dynamic approach Empirical evidence using CGE techniques indicates that depending on the precise
scheme of integration and the 1ax replacement methods integratiOn may have substantial effects In
Ihe work of FuliertonmiddotKingmiddotShoven-Whalley (1980 1981) total integration was found to yield an bull
annual static eHielency gain of $4 to $8 billion in 1973 dollars Simulated dynamic gains may be as
large as $695 billion or abcu 14 of the present value of tulUre consumption and leisure in the
US economy These gains result primarily from interindustry reallocalions of Investment and an
improved inter~emporal allocation of consumption
Mora recent work emphasizes hOw consumers asset portfolio decisions and firms finanCial
decisions affect the efficiency gains from integration Slemrod (1980) focusing on consumers
asset ponfoio deCisions finds static efficiency gains which are about twice -as Iarge as those
repOrted iiy Fulierton-King-ShovenWhalley (1961) FuUrlon-Gurdon (1983) focus on the
firms financial decisions They roporl efficiency gains of 6 of GNP from the elimination of the
tax distor1ions favoring debt However when they eliminate the corporate tax and repiace it with
increaSed persona income taxes addlHonal dls10t1ions are created in the optimal labot~leisure
decisions These distortions tend to dominate the analysis slJ(h that the overall effects of complete
integration are very modest Galpr-LuckmiddotTodr (1986) address simultaneously consumers
asset pOrlfollo decisions and the firms financial decisions They also find very modest efficiency
gains from integralion
The above results are important but may be severely biased There are severa aspects of
economic behavior and modelling crucial for the study of Income tax Integration whiCh have no been
i 37
captured innny of the above papers One aspect is the absence of government deficitsmiddot a balanced
budget is assumed It is true Ihal resource crowding OUI is caplured In these mOdels but financial
crowding out induced by government spending is nOI Second investment is not derived from
optimtztlcn behavior Investment behavior passively accomodates endogenous saving decisions As
a consequence the differential impacts of policies in the incentives to save and to invest are not
captured Aso full capital mobility across secors is assumed with instantaneous capital
adjustmenlS towards optimal levels This assumption rules qut different costs of capital across
seclors and therefore differentiated reactions to tax policies changes Finallyhe modelling of both
gov9rnmerlt deficits and of endogenous real and financial investment decisions necessitates the
~nsideraiion of a dynamilt framework and the introduction of financiaJ assets govemment bonds and
iprivate financial assets A dynamic framework also highlights the efficiency effects of Integration on r
ithe ptimal intertemporal decisions The above models are either sIalic (Slemrods and
GalpermiddotltJckemiddotTode(sj or a dynamic sequ~nce of otherwise Slatic model bullbull I Pereir a (1986b) develops a dynamic applied general equilibrium model bull to study the tbull
efficiency effects of integration on the growth and allocation of investmenl across sectors Special
Iattention is paid to the structural affeelS of resource and financial crowding oul induced by
gove mect spending and deficits and to the real and financial Investment reactions across industries
to both tax integration and flnancial crowding out His mode departs from previous work on income
lax Integraticn and for that matler from most 01 the CGIE literalure for tax policy evaluation in
several direcllons It encompasses an endogenous sequential equilibrium struelure founded on
dynamic behavior with flexible expectations Government deficits are optlmaUy determined
Investment decisions are tward looking and the resull of optimizing behavior Several financial
assetsmiddot public and privatemiddot are considered
Simulation resulls Indicale that the welfare gains from Integration under large deficits are at
32
best lJlodest when measured in terms of the GNP The elimination of the corporate tax and its I replacemenl by increased income lax rates yields long run benefits which are never larger than
2 of the presenl yalue ct telr consumption and leisure onder the previous tax regime and 1
under the current tax regime However the short run w~lare effects of full integration tend 10 be
very low and in in some scenarios even negative This is a flew intertemporar pattern of efficiency Ieffects which reflects an adjJsiment lag in the interindustry investment decisions due to the
existence of costs of adjustment On the of her hand partial integration achieved by excluding I dividends from the corporale lax base syslematically yields negalive efficlencyeffeclS In Ihe long Inm these can be as high as 3 of present value of the future value of consumption and leisure This
is a new second best effect suggesllng Ihat less than complete integration may have perverse I efficiency effects
The dynamic slructure of PerclrBs model snowing for an improved freatment of real investment
declsions and government deficts provides a potential setting for a more accurate measure of the
costs and benefits of inlegration The simulation resullS suggest lower benefits and an intertempora Ibull
pattern of growing benefits
Simulation results also clearly suggest robust negative effects from partial integration How
reliable is this result If dJvidends were deductible from the corporate lax base (as are interest
payments) the preferenlial Ireatmant of debl over equily would be ellminaled Corporallons should
be expecled 10 react by decreasing the optimal deWeqully rallo However corporale financial
gecisions are exogenously set In Pereiras model as in aft of the other models surveyed that go as far
as dealing with the issue
Also withdrawing dividends from Ihe corporate tax base is a way of eliminating Ihe double
taxation of dividends at belh Ihe corporate and personnal Income levels How will the optimal
aliocellon of household saving accomodate that change in Ihe relalive return to corporale equity
39 I
t However hOusehold portfolio decisions are exogenouly set in Pereiras model as in all of the other
models surveyeltj go as far as dealing with the issue
IIt is sate to say that the evaluation 01 the benefits 01 partial Integration as defined above are
sevrey underestima1eO Meanwhile the distortionSoenerated by the increase in the marginal
personal tax rates designed to ralse the revenue foregone by the dividend exclusion are fully
aCCltlunted for A good measure of the costs of integration together With a poor evaluation of the
benefils may very well be the reason why partial integration is Simulated 10 yield neg alive
eHikiency gains
On a different vein as most of Ihe model surveyed here Pereiras is a closed economy model It
is legitimate to wonder how the resullS would change If international capital flows were anowed
IThis almost certainly would affect financial crowding out since a good d~al of the capital inflows
could be used to finance public debt
For these various reasons the results should be treated as preliminary but they strongly I suggosl that the dynamic structure provides valuable new insights into the corporate tax integration SSU2
40
7 Concluding remarks I I I i
What has been acccmplished Ihis far The success of Ihe CGE research in laelding the challenge
of dynamic modelling can not be denied
I iGreat progress in the modelling of household behavior has been achieved Promising
developm~nts in the modeUing of production and government behavior have also been accomplished
Interesting innovations in the concept of equilibrium and computation techniques were discussed
The policy analysis was greatly enriched by considering transitional effects together with longer
run steadymiddotstate equilibrium paths generalized equal yield strategies accomodating different
financial crowding out impacts and generalized dynamic policy evaluation indicators
An important set of issues have been addresed by the models surveyed Traditional issues
focusing on the effects of capital taxation and consumption taxes have been pursued In terms of
capital taxation for example the intertemporal nature of the issue was enhanced by allowing an
optimal evolution of capital stock in the economy and forward looking optimal investment decisions
In Andersson (1987) GouldermiddotSummers (1987) and Pereira (1986b 1987d) this is coupled
with a improved treatment of several tax provisions like investment tax credits and depreciation
allowances In terms of the consumption taxes developments in the specification of Intertemporal
household behavior the introduction of overlapping generations and the modelling of
intergenerational links via bequest motives as well as a closer attention to demographic evolution
provided a much improved economic setting for the understanding of the several aspects of the
problem
Furthermore dynamic modelling has permitted the addressing of a set of new issues Policy
Issues ranging from dynamic tax policy analysis to the evaluation of exogenous changes in
government expenditures to the impact of different methods of financing government expenditure to
4 1
the importance of tinanciai croHrii1g out to 1he relevance of consumorS expectations in tne policy
results have been addressed
Despile all of the progress and in part due to such progress several avenues are wide open for
much neded additional research The following seem to be the most Interesting and promising
areasmiddot First the specincation of liquidity constraints to household behavior may help to obtain a
better understanding of policy chenges that affect directly the interest rates in the economy Second
major attention needs to be paid to the incorporafion of a roraign sector (with open capital markets)
Third some efforts are needed in terms of finding adequate computation techniques now that
dynamic modelling eally makes the curse of dimensionality worse than eVOrbull
The modelling of financial markets is the single most unsatisfactory speet of the dynamic
models surveyed here~ The problems associated with provlding a mo~ adequate treatment of
financial markets and in general endogenous and optimal financial deci$ions bull both at the corporate
and at the personal levelsmiddot have 10 be addressed That necessarily requires Introducing uncertainty
into the models To b adequate this must go well beyond the mere assumption of point expeetalions
with the whole array of modelling implementalion and evaluation problems thai generates
bull I i
42
REFERENCES
Abel A and O Blanchard 1983 An Intertemporal Model of Saving and Interest Econometrica
51 pp 675-92
~ Andersson K 1987 Tar-ation of Capital in Sweden - A General Equilibrium Model Ph D
Dissenation UnIversity of Lund Sweder
Auerbach A L KotlikoH 1983 National Savings Economic Welfare and the Structure of
Taxation in M Feldstein ed Behavioral Simulation Methods in Tax Policy- Analysis Chicago
University of Chicago Press
Auerbach A l Kotlikoff 1984 Social Security and the Economics of Demographic-
Transition in H Aaron and G Burtless eds Retirement and Economic Behavior Washington DC
Brookings Institution
Auerbach A L Kotlikoff 1987 Dynamic Fiscal Policy Cambridge Cambridge University
Press
Auerbach A L Kotlikof and J Skinner 1983 The Efficiency Gains from Dynamic Tax
Reform Internatiooaj Economic Reyiew 24 pp 81middot100
Ballard C 1983 Evaluation of the Consumption Tax with Dynamic General Equilibrium
Models Ph D Dissertation Stanford University
Ballard C 1987 Tax Policy and Consumer Foresight A General Equilibrium Simulation
Study Economic Inquiry 25 pp 267-284
Ballard C and L Goulder 1985 Consumption Taxes ForeSight and Welfare a Computatonal
General Equilibrium Analysis in J Piggott and J Whalley eds New Deyelooments in Apolied
General Equilibrium Analysis Cambridge Cambridge University Press
Ballard C D Fullerton J Shoven and J Whalley 1985 A General Equilibrium Model for
t 43
Tax Eoliey Eyalua~iQO NBER University or Chicago Press
Borges A (1987) bull A Survey of Compula1ional General Equilibnum Models CECD Economic
Slud~ 8
Bovenbig L 19B3 imperfect Capital Mobility in a Perfect Foresight Herberger Model
Mimea UC Br~eley
Bovenberg L 1984 Capitel Accumulation and Capital Immobility Qmiddottheory In a Dynamic General Equilibrium Framework PhD dissertation University of California at Berkeley t
Bovenberg L 19850 Dynamic General Equilibrium Tax Models with Adjustment Costs in A iManne ed bull 1985 EtQnomrc ECUjilbrium Model Formulation and Solution Mathematical
lProgramming Siudy 23 Arnsterdam NorthmiddotHoliand
Bovenberg L t 1986 Capital Income Taxation in Growing Open Eeondmies Journa of Pubtic
Eamp(omlcs 3t pp 347middot376 IBovenberg L ald W Keller 1981 Dynamics and Nonlinearlties in Applied General
Equilibrium Models Internal Report Mlmiddot81middot208 Department for Statistical Methods CBS bull
Vooiburg The Netherlands
Charnley C 1981 The Welfare Cost of C~pltal income Taxation In a Growing Economy
IJournal of Political ECQnQm~ 89 3 pp 468middot496 I
Charnley C 19810 Efficient Tax Reform in a Dynamic Model of General Equilibrium I Mimeo Yale University I Charnley C 1982 On the Optimal Taxation In Stylized Models 01 Inlartemporal General
Equilibrium Mimeo Yale University I Decaluw B and A Manens 1985 Pays en Oeveloppement at Modeles Calculables OEquilibre
General - Une Revue de al litterature empirique Monographie Sene 9 Volume 3 Centre de
Recherche et Developpemont Economlque Unrslt de Montreal
44
bull
I I
I
I
1
Erlich $ V Ginsburgh and L van der Heyden 1987 Wher~ do Real Wage Policies Lead
Belgium bull A General Equilibrium Analysis European Economic Reyiew (forthcoming)
air R and J Taylor 1983 Solution and Maximum Likelihood Estimation of Dynamic
Nonlinear Ratioant Expeclations Models Ecooometrica 51 pp 1169middot1186
Feltenstein A 1984 Money and Bonds In a Disaggregated Ec~nonYin Searl 1-1 J Shoven
Eds Aoolied ~eoeral EqJiliDCiulD tioalYsis Cambridge University Press Cambridge
Feltenstein A 19B6 A Dynamic General Equllibrlu Analysis of Financial Crowding Out
Theory with an Application to Australia Journal of Public Economics 31 No1
Fullerlon D R Gorden 1983 A Reexamination of Tax Distorllons in General Equilibrium
Models in M Feldstein Ed Behayora Simulation Methods in Tax pOlicy Analysis Chicago
University of Chipgo Press
Fvllerlon D J Shaven and J Whalley 1983 Replacing the US Income Tax with a
Progressive Consumption Tax JoulOal of Public ECQnomics 20 pp3middot23
Fullerton D Y Henderson J Shaven 1984 A COmparision of Methodologies in Empirical
General Equilibrium Models of Taxation in Searl H J Shoven Eds ~pQlied Geoaral Jguilibrium
lIoalyss Cambridge University Press Cambridge
Fullerton 0 A King J Shaven and J Whaney 1981 Corporate Tax Integration in the United
Statesmiddot a General Equiiibdufi1 Approach American Economic Review 71 pp~ 677~691
Galper H R Lucke E Tader 1985 Taxation portlor Choice and the Allocation of Capital A
General Equilibrium Approach Mimeo
Gill p W Murray M Saunders M Wright 1986 Users Guide for NPSOL (Version 40) A
Fortran Package for Nonlinear Programming T echnicsl Report SOL 86middot2 Deparlment of
Operations Research Stanford University
Glnsbur~h V and L van der Heyden 1985 General Equilibrium with Wage Rigidities An
45
AppJication to Belgium in A Manne ed t Economic EQumprit1m~ Made Formulation and $oJioo
Mathematical Programming Study 23 Amsterdam NorthmiddotHolland ~
Goulet J 1968 1 Adjustrent CO$1S in the Theory o(the Firm Reyiew of EconomiC Studies 35 bull
pp47middot55
Goulder L 1985 Tax-Financed versus BondmiddotFinancedGhanges in Governent Spending
bull Mimeo Harva(d UniversilYmiddot
Goulder l J Shaven and J Whalley 1983 Domestic Tax Policy and the Foreign Sector The
importance of altemative foreign policy formulations to results from a general equilibrium tax
analysis model In ~n methods in Ibe antico of income from caoia1 ed Martin
Feldstein Chicago University of Chicago riSS
Goulder l L Summers 1987 Tax Policy Asset Prices and 3rowtly ~
A General Equilibrium Analysis NBER Working Pper No 212B
Grandmont J 1982 Temporary General Equilibrium Theory Ch 20 in Handbook 01
Mathemalical Eccnornics Vol II NorthmiddotHoliand
Harberger A 1959 The Corporate Income Tax An empirical Appraisal In lax Revision
Comoendjum Voll House Cornmitee on Ways and Means Washington DC Govemment Printing
Office
Harberger A 1962 The Incidence of the Corporate Income Tax JOULn1 of Political
Economy 70 pp 215middot240
Harberger A 1964 Taxation Resource Aliocalrlon and Welfare fn lb Role of Dir bullbulll and
Indjresl Taxes io tbe Fede[al Reserve s~I~Ill Princeton Princeton Unlveresity Press
James J 1985 New Developments in the Applications of General Equilibrium models to
Economic History in Piggott aand Whalley Eds
Jorgenson D and K Yun 1984 Tax Policy and Capital Allocation Harvard Inslitute of
bull
I
I f
Economic Research WP 1107
Judd K 1965 ShonmiddotRun Analysis of Fiscal Policy in a Simple Perfect Foresight Model
Journal of poJiHcal 5CCJiW1l pp 298middot319 r
Upton D J Potbc J Sachsa nd L Summers 1982 Multiple Shooting in Rational
Expectations Models fuoometrlc 50 1329middot1333
Manne A ed 1985 fconQrnjc Eoullibrium Model FOrIDujaljQo and Solution Mathematical
Pogrammlng Study 23 Amsterdam NorthmiddotHolland
Manne A 1985 Or the Formulation and Solution 01 Economic Equilibrium Models In A
Manne ed ECQOQlli-Jr~ibmiddotum ~tQdel EQlJulailon and Solution Malhematical Programming
Study 23 Amsterdam NorthmiddotHolland
Mansur A and J Whalley 1984 Numerical Specification of Applied General Equilibrium
Models in Scarf H J Shoven Eds Aoplied General Eouillbrium AnaiysectIsect Cambridge Universily
Press Cambridge
Merrill 0 1972 Applications and Extensions of an Algorithm thaI Computes FixedmiddotPoints of
Cenain Upper Semimiddotcominuou Point to Set Mappings PhD Disserlation Unlvel$ity 01 Michigan
Negishi T 1960 Welfar Economics and Exislence I an Equnlbrlum for a Competitive
Economy MelroeCOOQ11ka 12 92middot97
Pereira A 1985 On the Existence and Uniqueness of Optimal Ou1put Path lor CRTS Firms in
Adjustment Costs Technologies Mimeo Stanford University
Pereira A 1985a Consumer Expectations and Ihe COSI 01 UllUty In an Intertemporat
Framework Mlmeo Stanford University
Pereira A 1985b Corporale Tax Integrallon and Government Deficits A Dynamic General
Equilibrium Anaiysls Mlmeo Stanford University
Pereir A 1987 On Ihe Computation of the Users Costs of Stocks Ecooomi (forthcoming)
47
Pereira A bull 1987b Equal Yield Alternatives and Government Oeficlts Mimeof Stanford
University
Pereira A 1987c A Dynamic Applied General ~uHlbrium Model for Tax Policy Evaluation
Comlletamp Docurnent2tion Mimeo University of California San Diego (in progress)
Pereira A 1987d Should the investment Tax Credit be Rei(ltroduced Mlmeo University
of California San Diego (ir progreSs)
Piggott J and J Whalley eds 1985 tJew QeyeQpments in Applied General EQuilibrium
bullenalysjs Cambridge Cambridge University Press
Preckel Pbull 1985 Alternative algorithms for computing economic equilibria In A Manne Ed
Radnor R 19amp3 Equllibrium under Uncertainty In K Arrow M Intrilligaor ods
I ibull
Robinson S 1 g86 Mullisectorai Models of Oeveloping CountrieS a Survey Oivision of
Agricullure and Resources UC Borkeley WP 401
Scarf H 1967 The approximation of fixed points of a continuous mapping ~MM Jouwal of
epplid Mathornallcs 15 pp 1328middot43
Scarf H with T Hansen 1973 00 the Computation of EconomiC Eguilibria New Haven Yale
University Press
Scarf H J Shoven Eds 1984 61l1llied Genera Elujlibrium 6nalyssect Cambridge University
~ Press Cambridge
Shoven J 1976 The Incidence and Efficiency Effects of Taxe on Income from Capital
Jurnal of ramie1 Economy 84 pp 1261-83
Shoven J 1983 Applied General Equilibrium Tax Modemng IMF Slaff PaoerS 30
Shoven J and LB Simon 1986 Share Repurchases and Acquisitions An Analysis of which
Firms Participate NBER Working Paper No 2243
bull
48
IShaven J and J Whalley 1972 ft General Equmbrium Calculation of lhe Effects of
Differential Taxation of Incoma from Capital in the US Jauroe of public Econgmics 1
pp281middot321 (
Shoven J bull and J Whalley 1973 General Equilibrium with Tax$s a computation procedure
and an existence pr00f ReviEhi of Economic Studies 40 pp 475middot490
Shoven J J Whalley 1977 Equal Yield Tax Alternatives A General Equilibrium
Computa~ui0nal Tfchnique J~HJrnal of PUblic EconomIcs 8 pp 211~224 bull
Shaven J J Whalley 1984 Applied GeneralmiddotEquilibrlum MOdels of Taxalion and
International Tr8de An Introduction and Survey Journal of Economic literature 23t pp
1007middot1051
Slemrod J 1980 A General Equilibrium Model of Capital Income Taxation Ph D bullDissertation Harvard Univnrsity
Slemrod J 1983 A General Equilibrium Model of Taxation with Endogenous Financial
Behavior in M Feldstein edbull Behavioral Simulation Methods In Tax Policy Analysis Chicago
UniVerSity of Chicago Press
Stone J 1985 Sequential Optimization and Complementarily Techniques for computing
Economic equilibrium in A Manne Ed
Summers L 1981 Taxation and Corporate Investment a q-Theory Approach BroQkinas
Summers L 1985 Taxation and the Size and ComposHion 01 the Capital Stock An Asset Price
Approach Mlmeo Harvard University
Whalley J 1975 A Geneal Equiloibrium Assessment of the 1973 United Kingdom tax
reform Economica 42 pp 139middot61
1JLTIllOS JIQll~)ltG PAPERS PUBLICADOS bull
I
I I
j
112 70 - LUCENA iJiCIIJo ~ IIEnvironrilent Honitoring and Orqanization Structure T (Setellibro 1987)
nI 71 - L~ENA Diogo A Note on the Representation of Information Stxueture (SeteJObro 1987)
nQ 72 - LUCENA D1CgO Envuorment 111)nitoring and Organization
Structure II (Setelilhro 1987)
ril 73 - AlmmE3 Antonio Pais e CAsPAR Yitor bull Tributacao Inoentivos e InvesUlllentos Analise Qualitativa noYellibro 19S7)
n2 74 - liAlltOSA rose Pedro Opt1lllal lIage R1gidity A SUggestedtlthodology to Teet tb Theory and an Application (Outubro 1987)
nI 75 - CABRAL L 11 B Three lIotes on SYJIIlletric Gemes witt itsYJIIlletric Equilibri (NoveJObro 1987)
n2 76 - LUCENlt D10gOTo Search or Not to Search (Fevereiro1967)
nQ 17 - SA Jorge asconcelos euroI A Theory of SynergyB (Fevereiro1987)
n2 76 - VILARES tanuei Jose Os Be IntenaMios IlIIportados ColIIO Factor de Produao (Julho 1987)
nQ 79 - SA J~rge VasCQncel03 e HoY to Copete and Co~unicate in tature Industrial Products_ (aio 1988)
n2 80 - ROIl Rafael Learning and Capacity Expansion in a New larket Under Uncertainty (Fevereiro 1988)
n2 81 PEREIRA Alfredo llarvilo Survey of Dynamic Computational General EquilibriUlll Hodel for Tax Policy Evaluation (Outubro 1987)
Qualquer in ormacao sabre o 1IIorlltlng Papers ja publlcados sera prestada pelo Secretariado de Apoio aos Docentes podendo os ~e3OSf
ser adquirido5 ne Secyo de Vlndas da Faculdade de EconollLia UNL na Travessa Esteviio Pinto Cltpolide - 1000 LISllOA
21
an addjlion~ consumer group
Bovcoberg (1986) develops a model In which two economies lIr considered each following
intertempoml perlect foresight paths These economies meet in the international forum Their
trade relaroMhips are characterized by yearly balanced trade accounts
None of the new generation of dynamic cae tax models has yet IncorPorated the lnternational
capital flows as done in aoulder-Shoven-Whalley (1983) for the earlier Ballarrshy
-Fullertonmiddot Shaven-Whalley (1985) model This is unfortunate as there is an important
Intertmporal aspect to international lending The first al1empts along Ihese lines are due to
Andersson (1987) and Erlich-Glnsburgh-Heyden (1987) Andersson (1987) in his model of the
Swedish economy adopts a closed economy appraoch in which rates of return in the domesJic
economy are largely determined by the international capital markets Fidinterest rate induce
Intlnatlonal capital flows which determine and finance the International lrade imbalance In
lurn In ErlichGinsburghHeydenmiddot (1987) foreign trade Is generaled according to an
intertemporal Irade welfare lunction with constant import and export elasticities In the shortmiddotrun
they allow inlernational trade inbalances which generate capilal ftows to the domestic households In
the long run however trade balance is assumed
35 The need for financial markets
A dynamic economic Wucture not only provides the ideal environment 10 model many features of
economic behavior it also permits one to Incorporate the financial side of the economy If the
government is allowed to run deficits the question 01 deficll financing automatically follows If
investment is optimally dewmined and returns to capital are different aClOSS sectors problems of
investment financing arise If there are seeral final1cfal assets In the economymiddot government
4-
I i
ibull
I I I I
I
j
22
bonds privpte bonds and equity (or simply physical capital installed inmiddot different sectors) the
problem of allocation of saving among assets with potentially different returns arises
The papers surveyed var greatly with respect to l~e extent of their attention to the financial
side of the economy At Clne extreme are the models in ~anardmiddotFuliertonmiddotShoven-Whalley (1985)
Andersson (1987) Ballltld (1983) Ballard-Goulder (1985) Bovenberg (1985 1986) and
Erlich-Gjnsburgh-Heyden C19B which are devoted exclusively to the real side of the economy
In turn Auerbach-KotliKoff (1984 1987) Feltenstein (1984 1986) and Goulde (1985)
allow for government debt In these models saving finances changes in government debt and physical
capital Private and publlc 2ssets are perceived by the households as perfect substitutes The
allocation of saving merely adjusts to the relative demands for funds
Feltenstein (1984 1986) is the only model surveyed which introduces money Government deficits are financed by issuing money and bonds according to an exogenously given rule Money is
demanded by consumers for transaction motives and an exogenously given fraction as a store of value
On the other hand government bonds and physical capital are the vehicles for the intertemporal
transfer of wealth
Summers (1985) and GOlldermiddotSummers (1987) introduce a whole menu of financial assets shy
firm specific equity capital Different assets earn different rates of retum However such rates
are equal up to constant and exogenous sector speCific risk premia Therefore the introduction of
constant exogenous risk premia as helpful as it may be in the context of calibration does not solve
the main issue of the non-optimality of the allocation of saving Also talking about risk premia in a
deterministic context is somewhat unsatisfactory
Pereira (1986b 1987d) also introduces a whole menu of assets private and public bonds and
firm specific equity However all the assets are expected to yield the same rate of return and
therefore perc~jved as perfect substitutes The different asset types allow consideration of
23
exogenOU$~8tUequjty triO dividendlrc~enlion rufes and therefore seve-ral sources of investment
financing bond equity and retained earnings
The nonmiddotcptimalHy of the allocation of saving and the absence or exogenelty of corporate
financial rules is _a reftectlon of the limitations of the determInistic approach Either in a context of
perfect anticipation oi the future prices Of in general within the t$atm of point price
expectations all the papers follow a deterministic approach Under such circumstances consumers
either expect diHerent rales of return (inclusive of risk premium) across assets in which case
they wilt buy only one asset (that with highest rate) Of they expect equal rat~s of relurn in which
case Ihey are indifferent about the asset composition of Iheir portfoliO There is no way of
Iradingmiddotoff rales of return and risks to obtain an optimal interiof solution to the problem of the
allocation of saving
The most advanced contribution in the modelling of saving allocation in a CGE setting is due to
Slemrod (1980 1983) in the context of a static one-period model In his model consumers act
according to a two stage separable decision process They Hrsl decide on how much to save Then
they decide on the allocation of saving according to an Indirect uttlity function dependent on the rates
of ratum and variances offered by the different assets in the eoonomy The sourCe of riskiness in the
ampConory comes from an uncertain marginal product of capital On the other hand aside from
portfolio decisions the res of the economy is insulated from uncertainty
The most complete contribution in terms of the treatment of the corporate financial rules is
FulienonmiddotGordon (1983) In a variant 01 Ihe BailardmiddotFulienon-ShovenmiddotWhelley (1985) model
Fullerton and Gordon have capital intensity and optimal financial decisions Jointly deter~ined
through a two siage process The cosl of financing capital is minimized by trading off the tax
advantages of debt against the epected real bankrupcy coSls inherenl with high debVequity ratios
Given the optimal debtequity ratio the level of inveSlment Is chosen such that at Ihe margin the
24 l
bull
rerurn pn ~uity equas the return on bonds plus an exogenous risk premium
36 Market assumptions equilibrium and expectations
Virtually all of the papers arc characterized by Walrasian market clearing assumptions All
markets are perlect1y competi1ive Atomistic competition among agents 1s assumed even thoughmiddot only
a flnile number of agents is considered Virtually no mark~t disequilibria or price stickiness afe
considered The only exception is ErlichmiddotGinsburghmiddotHeyden (1987) bull In thei model of the Belgium
economy I the _wage rate is fixed in the short-run Therefore in the shOrlwrun desequilibrium in
the labor market will generate endogenous unemployment However in lh$ long-run all prices
Including the wage rata are Ilexible and accordingly ali markets ciear
Given the dynamic nature of behavlor in the economy markelmiddotcfearing prices in each period
depend on expectations of fture prices and on tax variables In the economy There are essentially
two ways at interpreting the economic equilibrium in such a dynamic conte~bull If future prices are
perfectly anticipated (io expectations are self-fulfilling) a perfect foresight equilibrium
prevaHs Then future actions are merely the implementation of current decisions for future
periods However jf ptica expecla110ns are not perfect (Le agents make mistakes w1th re$pect 10
future prices) then a temporary or shortrun equilibrium prevails Markets ciear and clearing
prices depend on fu1ure price expectations Current plans about the future are typically not
precisely implemented They will be revised as more or better information becomes available to the
economic agents See Grandmant (1982) for a comprehensive survey 01 the temporary equilibrium
literature
Wilh the exception of Gould (1985) and Pereira (19S6b 19S7d) all the models surveyed
adopt th~ concept of a perfect foresight equilibrium In turn BaliardmiddotGouldr (1965) considers a
J
2S
flexible am9unt of foresight In terms of the number of years over which price movements are ~
foreseen Pereiras model (1986b 1987b)) is flexible in a somewhat different way in that it can
indude any range of foresight from myopia to perfect foresight The choice between the perfct foresight and lemporary equilibrium is ultimately to be made on
philosophic grounds It can be argued that less than perfect expectations Imply that agents are
irralional in some way (see Auerbach-Kctiikoff (1987 p 10) However the reverse argument can
be made One can question wiether agents are reaHy rationar and perfectly knowledgable about
future prices
Recent evidene of Balird (1987) Ballard-Goulder (19851 Goulder (19851 and Pereira
(1986b 1987d) suggests that the choice in modemng expectations is an important one They show
that the degree of foresight inl0 thc future (ranging from perfect foresight to-myopic expectations)
may have dramatic impacts on the pollcy conclusions of Ihe model Accordingly the best research
strategy may b to design models which are flexible enough 10 allow for different rules regarding the
formation of expectations With the exception of the articles just mentioned sensitivity analysis
bullhave previously not been performed along this dimension
In terms of implementation the two concepts of equilibrium - perfect foresight and temporary
equilibrium have different implications The dimensionality of the equ1ibrium soll1ion algorIthm
is involved Suppose we have a model with ten markets to be run for a period of 50 years Aside
from normalization a perfect foresight model implies computing prices In 500 dimensions while a
-t~mporary equilibrium model requires sorving SO equilibria each in ten dimensions Given that
computational speed often varies with the cube of the number of dimensions the lemporary
equilibrium formulation is potentially strikingly more feasible However Ballard (1987) and
Goulder-Ballard (1985) have devoloped techniques to greatly speed the computation of a perfect
foresight equilibrium
26
j
I
4 Jrnplemenlalion and Issues on policy evaluallon
41 Calibration and equilibrium comparlslon
CGE models at typically parameterized by the use of a calibration procedure Some parameters
are exogenously given However some crucial parameters are determined in such a way that the
model replicates the data for a given base year See MansurmiddotWhalleY 11984) for an eXlensive
discussion of 1~i$ issue
CalibratIon in a dynamic context is generaliy interpreted as requiring two properties First
replication of base year data is required Second the model is parameterized to simulate an
intertemporal balanced growth path when Ihe base policy is maintaineq This Is the approach
followed by Ballard (1983) BallardmiddotGoulder (1985) Goulder (1985) and SummersmiddotGoulde
(1986) It follows the practice of BallardmiddotFuliertonmiddotShovenmiddotWhalley (1985) with their
sequential cquilibrium dynamics
Other authors Auerbach-Kotlikof (1983 1984 1987) Boyenberg (1984 1985 1986)
and Pereira (198Gb 1987d) follow whal we cali a qualitalive calibration The slructural
palametars are exogenously chosen so that lh economy follows a reasonable path into Ihe fulure
ThaI has 10 do wllh the fact thaI given the recursive nature 01 the dynamic economy and aside tom
such structural parameters only iniUa stock values are needed to run these models Given initial
conditions on the stocks of say private wealth capltal and government debt agents will optimize
and Ihereby generate a lirst round of net demands and equilibrium conditions In lurn the
equilibrium prices will determine the evolution of the stock variables into the next perIod
There are several potential problems with calibration in the conlext of dynamic models The
assumption of a steadymiddotstate growth ~ath in the base case can be questioned First while
27 I Ibullbull
steadymiddotstaU is a possibity it certainly is not the only meaningfull solution to dynamic models
Even in the case of a perfect foresight equilibrium the model implies an equilibrium path which i may or may not involve colanced growth The model not the modeller should dictate the nature 01 I themiddotbase case path Second in the context of a temporarY equilibrium path a steady state solutlcn is Inol a likely model QutcOn1tl In fact unlessmiddot expectations are stattc short run behavior consistent
with a steady-state evolution will in general not be generated On the other hand if static I expectations are self~fufjili[1g we have in fact a perfect foresight model Third even the base year
replication requirement may cause problems In Ihe cOntexl of temporary equilibrium Any
calibration parameter would be conditional on expectation rules which is probably an undesirable
feature
The nalur of the two-requirement calibration strategy is very muchin the spirit of the bull
traditional design of the comparision of alternative equilibria comparis_ion between a steady-state
base case on one hand and alternative paths including a transition period and a finat sleadyyenstate on
the other hand Pereira (198Gb 1987d) compare different (not necessarily steadystate)
equilibrium paths The arguments against such a procedure are based on the idea that the impact of
policy changes can be observed most easily since all departures from the steadymiddotstate can be
attributed to the altemalive polioy On the other hand the base case so dEfined as a steadymiddotstate is
consist~nl with previous work in a less dynamtc setting and Iherefore allows a common standard
for comparing model results
42 Computation algorithms
We are still al a stage in which basically each author uses a different computational technique
Th development of dynamic models - in parlicular wilh adjustment costs andior perfect foresight
bull
28
has corresponded with the decline in the use of fixed point algorithms In -fact given the relative
arge dimensions inevitably involved such algorithms lend to be very inefficient at the best and
chen prohibitively stow See Stone (1985) and Preckel (1965) for a comparative assessment of
different computation techniques Among the models surveyed only Ball1rdFulle(ton~
-ShavenmiddotWhalley (1985) and Feltensteln 1985 use Merrills variant of the fixed point
algorithm 1echnique
AuerbachmiddotKoHikoff (1963 1984 1967) follow a three stage procedure They first compute a
base case steady-state then a revised case steady state and finally transition path for the economy
berveen these two steady-states tn all stages a Gauss-Seidel iterative procedure is used
Ballard (1982) Ballard-Goulder (1965) Goulder (HiSS) and Summers-Goulder (1986)
~
use a method developed by Ballard and Goutder which is similar in many awects to the FairmiddotTaylQr bull
(1983) algorl1hm Short-run equilibria are calculated (using Merrills algorithm) parametric on
nrice expeC~lions The model is then iterated to generate self-fulfilling intertemporaf expectations
and the corresponding perfect foresight equilibrium In a relatively similar approach Andersson
1987 uses a simulation program SIMNON developed In the University of Lund This program
can handle two-paint boundary problems In a fashion consistent with the multiple shooting
algcrlthm (see Lipton-Poterba-Sachs-Eummers (1982))
Boyenbergs computational approach (1985 1985) differs from the other models surveyed in
that he relies heavily on analytical techniques Computations are done by using a dynamic version of
Johansons linearization method Being essentially determined by the continuous time nature of the
I I
f
model thiS linearization model has the disadvantage of confining the analysls to infinltesinal changes
around the base case equilibrium (see Bovenberg (1985) p 53) Pereira (1986b 1987d) uses an optimization algorithm NPSOL - developped by
rGillMurray-Saunders-Wrlght (1986)) This algorithm is used 10 compute the sequence of
I
29
Shortmiddotrun temporary equilibrium which makes up the intertemporal equilibrium path The
equilibrium conditions are seen as nonlinear equality constraints in the minimization of an artlcial
objective function The prices are normalized to the unit simplex by an additional linear equality I I 1 bull constraint
II ErlichmiddotGinsburgh-Heyden (1987) follow a unique approach in that they use a variant of the
optimization technique introduced by Negishl (1960) The economic equilibrium can be generated i
as a solution of a mathemallcal program Ihe objective function of which is a weighted sum of the
utility functions of the various agents while tho constraints set consists of the market dearing 1
conditions Ginsburgh-Heyden (1985) have extended Negishis resulllO tho case of downward price
lrigidities I
Finally the paper by Jorgonson-Yun (1984) is also unique in tIlat it is the only I
econometrically estimated model Different blocks for the consumption and production Side of the
-
model are separately estimated to provide the necessary structural parameters i
I The diversity of computation techniques is yet another indicator of the exploratory nature of the
body of literature surveyed in this article
43 Equal yield comparisions
The link between a base case and counteriactual simulations Is usually provided by the concept of
equal yield Shaven-Whaley (1977) discuss the meaning 01 equal yield in a general equilibrium
context when government is not allowed to run defICits Equal yield is interpreted to moan constant
public utility Government base case utility is maintained in the counferfactual experiments With
balanced budgets the concepl of equal yield is unambiguous The new equUibrium prices and the
balanced budge condition will del ermine the minimal expendilure and taxes needed to maintain base
30
case publi~ utility Accordingly in general equal yield is inconsistent with equal nominal tax
revenue Some change in tax revenue Is necessary Different tax replacement schemes ate
considered to assure that enough tax revenue is collected This is the approach essentially followed
by most 01 the papers surveyed Only Feltenstein (1985) Goulder (198) and Jorgenson-Yun
(1984) chose not to follow an equal yield strategy
The question is of how to Interprete lila concept of equal yield when the government is alowed to
fUn deflcLts The optimal leve of expenditure for base ease public trliUty can now be taAinanced
bond financed or financed by a mix of bonds and taxation We have several versions of equal yield In
particular equal yield may now be consistent with equal tax revenUE Furthermore some meaSure
ltgtf financial crowding out effects of government deficns can be inferred from the comparislon of the
several equal yield alternatives These issues are extensively discussed ~Pereira (1987b)
44 Price expectations and the dynamic generalization of compensation
Indicators
The sum of equivalent and compensation variations over households is the most widely used
aggregated measure of efficiency gains or losses In the context of $teady~state eomparisions andor
when complete future markets exist andlor perfect foresighl is assumed there are no difficulties
associated with the use of the standard Hicksian indicators In fact correct future prices are known
in these cases
If expectations are not selfmiddot fulfilling andlor future markels are not available a dynamic
generalization is necessary 8allardmiddotGoulder (1985) provide some steps in that direction by
defining an indicator that accomodates periods far into the future when households do not have
perfect foresight but a steady~state prevails On he other hand this issue is more fundamental in
i
[
I
i I I I I [
I I
bullI
31
the contex~ of a Gner1 lemporary equilibrium ftamawor~ In such Circumstances Pereira
(1986a) dev~lops a cnamlc generalization which is obtained as the present discounted value of a i i
sequence of short-fUn optimal expenditure functions consistent with a base case expected future I stream of IlWities I
32
5 Empirical evidenee from selected policy Issues
Dynamic tax models have generated several important results which escaped Sialic modellers
The following is a seleCpoundd set of issuos whIch stress the marginal benefits of dynam~ modelHng of
economic behavior in innovltlttive areas of analysis The discussion of a consumption tax emphasizes
the benefits of dynamic tusehold behavior and intergenerational aspects The study of the
flliminatlon or the re~intToduction of the investsnemt tax credits is made meaningful by the dynamic
modelling of production betlavor Modelling governmenl in a dynamic conlexts allows the modeller
to study the irrrpact of government deflcits Finally a dynamic framework lets us appreCiate the I
relevance of consumer expectations in terms of the evaluation of policy alternatives
i
i
51 Consumption Tax
t
By Its very nature a consumption tax can not be adequately investigated with a static model bull
Fullerion-Shoven-Whaicy (1983) use Ihe model of the US economy s described in
Ballard-Fullerton-Shoven-Whalley (1935) to evaluate the movement from the currenl US tax Isys1em to a progressive consumption tax Since their modeJ Incorporates a Jaborleisure choice
where leisure is an untaxed commodlty their results reflect the fact that both the consumption tax
and Ihe present system are dslortionary
Concentrating on the rntertemporaf distortions they show that sheltering more saving from lhe
current US income tax could improve econornic efficiency even if marginal tax rate increases are
necessary in order to maintain government fevenue~ At first you have a revenue shortfall
However the economy moves to a higher sustatned growth path and uitimatefy lower tax rates
cangenerate the same revenue path Also wages increase in the long run with this policy The
bull
bullbull ~
33 shyt
i
I
I
present value of welfare gi[lS for a potiey of complete savin9 deduction wilh marginal tate
adjustments (consumption lax) is simulated to be around $500 billion to $600 billion 01 1973
dollars
They also investiga1e th 1ngth of ttme it takes th~ economy to adjust to these policy changes
Roughly they estimate the long run to b thirty years although this figure Is very sensitive to the
specification of the savings elasticity - a crucial parameler in this model
52 Investment T~x Credit
Goulder-$ummers (1 S87) address the impact of eliminating the lnvestmenl Tax Cradit (ITC) on
intersectoral capital formation and on economic growth Their model isYa1ticularly adequate to
addies$ this issu~ in that is postulates a forward looking investment m9we with adjustment costs
They show that the eliminating the ITC causes a reduction in the rale of investment Inveslment is
estimated to fall by about 7 in the short-run and by about 12 in Ihe new long run steady state
In lurn a previous policy anncuncement lowers Ihe overall al1ractiveness of Investment and leads 10
a downward shift of the Investrrent profile On the other hand the combined effect of a revenue
neutral simultaneous efimiraron of the ITC and reduction of the corporate tax rates is a long run
reduction ef the capital Sieck by 35 This pattern suggests that a revenue neutral increament in
both Ihe corporale tax rates and the inveSlment lax credits would be preferable 10 the formula
adopted in Ihe US Tax Reform Act of 1986
Pereira (1987d) addresses the impact of the ITe from Ihe oland point of Ihe curTenllax syslem
under the Tax Reform Act of 1986 Given Ihe concern over the potential depressiVe effects upon
savings and investment of the new tax system in conjunction with deficit reduC1ion mechanisms of
the GrammmiddotRudman type a pertinent question is should Ih lTC be e-introduced In the context
bull
34
-of his mod~1 of the US economy Pereira focuses on the trade-off between the distortion in the
intersectoral allocation of capital induced by the lTC the lower tax revenues it generates and
potential financial crowding-out effects on one hand and the positive effects of lowering the relative
price of new capital goods on the other hand Preliminary results confirm the qualitative results in
Goulder-Summers (1987) suggestting a welfare gain of about 2 of the present value of GNP in
the best scenario of absence of replacement taxes Furthermore preliminary results show an
important time pattern to welfare gains with the average benefits increasing the further you look
into the future This pattern is due to the presence of constraints to intersectoral mobility of capital
and inertia in the adjustment towards the optimal capital levels as reflected by the presence of
adjustment costs
I
I
bullAuerbach-Katlikaff (1gB) in the context of their intergenerational model of the US economy
consider the impact on savings capital formation and interest rates of deficit policies and balanced
budget increases in government consumption They conclude that deficit finance and government
consumption can significa-ntly crowd out capital formation and lower the welfare of future
generations However crowding out from deficit finance is a very slow process because it results
from increased government spending over potentially long horizons Also deficit policies may
substantially influence the longmiddotrun interest rates while leaving the short-run rates essentially
unaffected Finally and in opposition to the central role adjustment costs seem to play in both
53 Financial crowding outeff~cts of government deficits
Goulder-Summers and Pereiras models Auerbach-Kotlikoff suggest that the time path of interest
rates induced by a policy of deficit financing seems to be insensitive to the presence of adjustment
costs
bull
35
54 The role 01 consumer expectatlons
The expectations analysis has uncovered one of the benefits of dynamic modelling
Ballard-Goulder (1985) find that the appeal of adopting a consumption tax depends on the level of
foresight possessed by the consumers Furthermore additional foresight may be welfare worsening
This is a second best type of result This tends 10 occur under policies that lead to capital deepening
and declining rate of return to capital over time To the extent that consumers have more foresight
they will be beller equippod to anticipate the fall in the rental price of capital Consequently if a
saving incentive is enacted people save more with myopia than with perfect foresight Given the
exi~tence of taxes on capita and the discrepancy between the private an~--sccial returns to capital
the greater savings with myopia is beller socially Ballard-Goulder find that the welfare gains
from a consumption tax are reduced by about 10 when we move from myopia to farleet foresight
Therefore the attractivenness of adopting a consumption tax seems fairly [obust across these
foresight specifications
I
bull
36
6 Tbe power and weaknesses 01 dynamic modelling the example 01 corporale
tax Integrallon
The Issue of corporate tax integration allows us to sh~w the stre-ngths and weaknesses or the
dynamic approach Empirical evidence using CGE techniques indicates that depending on the precise
scheme of integration and the 1ax replacement methods integratiOn may have substantial effects In
Ihe work of FuliertonmiddotKingmiddotShoven-Whalley (1980 1981) total integration was found to yield an bull
annual static eHielency gain of $4 to $8 billion in 1973 dollars Simulated dynamic gains may be as
large as $695 billion or abcu 14 of the present value of tulUre consumption and leisure in the
US economy These gains result primarily from interindustry reallocalions of Investment and an
improved inter~emporal allocation of consumption
Mora recent work emphasizes hOw consumers asset portfolio decisions and firms finanCial
decisions affect the efficiency gains from integration Slemrod (1980) focusing on consumers
asset ponfoio deCisions finds static efficiency gains which are about twice -as Iarge as those
repOrted iiy Fulierton-King-ShovenWhalley (1961) FuUrlon-Gurdon (1983) focus on the
firms financial decisions They roporl efficiency gains of 6 of GNP from the elimination of the
tax distor1ions favoring debt However when they eliminate the corporate tax and repiace it with
increaSed persona income taxes addlHonal dls10t1ions are created in the optimal labot~leisure
decisions These distortions tend to dominate the analysis slJ(h that the overall effects of complete
integration are very modest Galpr-LuckmiddotTodr (1986) address simultaneously consumers
asset pOrlfollo decisions and the firms financial decisions They also find very modest efficiency
gains from integralion
The above results are important but may be severely biased There are severa aspects of
economic behavior and modelling crucial for the study of Income tax Integration whiCh have no been
i 37
captured innny of the above papers One aspect is the absence of government deficitsmiddot a balanced
budget is assumed It is true Ihal resource crowding OUI is caplured In these mOdels but financial
crowding out induced by government spending is nOI Second investment is not derived from
optimtztlcn behavior Investment behavior passively accomodates endogenous saving decisions As
a consequence the differential impacts of policies in the incentives to save and to invest are not
captured Aso full capital mobility across secors is assumed with instantaneous capital
adjustmenlS towards optimal levels This assumption rules qut different costs of capital across
seclors and therefore differentiated reactions to tax policies changes Finallyhe modelling of both
gov9rnmerlt deficits and of endogenous real and financial investment decisions necessitates the
~nsideraiion of a dynamilt framework and the introduction of financiaJ assets govemment bonds and
iprivate financial assets A dynamic framework also highlights the efficiency effects of Integration on r
ithe ptimal intertemporal decisions The above models are either sIalic (Slemrods and
GalpermiddotltJckemiddotTode(sj or a dynamic sequ~nce of otherwise Slatic model bullbull I Pereir a (1986b) develops a dynamic applied general equilibrium model bull to study the tbull
efficiency effects of integration on the growth and allocation of investmenl across sectors Special
Iattention is paid to the structural affeelS of resource and financial crowding oul induced by
gove mect spending and deficits and to the real and financial Investment reactions across industries
to both tax integration and flnancial crowding out His mode departs from previous work on income
lax Integraticn and for that matler from most 01 the CGIE literalure for tax policy evaluation in
several direcllons It encompasses an endogenous sequential equilibrium struelure founded on
dynamic behavior with flexible expectations Government deficits are optlmaUy determined
Investment decisions are tward looking and the resull of optimizing behavior Several financial
assetsmiddot public and privatemiddot are considered
Simulation resulls Indicale that the welfare gains from Integration under large deficits are at
32
best lJlodest when measured in terms of the GNP The elimination of the corporate tax and its I replacemenl by increased income lax rates yields long run benefits which are never larger than
2 of the presenl yalue ct telr consumption and leisure onder the previous tax regime and 1
under the current tax regime However the short run w~lare effects of full integration tend 10 be
very low and in in some scenarios even negative This is a flew intertemporar pattern of efficiency Ieffects which reflects an adjJsiment lag in the interindustry investment decisions due to the
existence of costs of adjustment On the of her hand partial integration achieved by excluding I dividends from the corporale lax base syslematically yields negalive efficlencyeffeclS In Ihe long Inm these can be as high as 3 of present value of the future value of consumption and leisure This
is a new second best effect suggesllng Ihat less than complete integration may have perverse I efficiency effects
The dynamic slructure of PerclrBs model snowing for an improved freatment of real investment
declsions and government deficts provides a potential setting for a more accurate measure of the
costs and benefits of inlegration The simulation resullS suggest lower benefits and an intertempora Ibull
pattern of growing benefits
Simulation results also clearly suggest robust negative effects from partial integration How
reliable is this result If dJvidends were deductible from the corporate lax base (as are interest
payments) the preferenlial Ireatmant of debl over equily would be ellminaled Corporallons should
be expecled 10 react by decreasing the optimal deWeqully rallo However corporale financial
gecisions are exogenously set In Pereiras model as in aft of the other models surveyed that go as far
as dealing with the issue
Also withdrawing dividends from Ihe corporate tax base is a way of eliminating Ihe double
taxation of dividends at belh Ihe corporate and personnal Income levels How will the optimal
aliocellon of household saving accomodate that change in Ihe relalive return to corporale equity
39 I
t However hOusehold portfolio decisions are exogenouly set in Pereiras model as in all of the other
models surveyeltj go as far as dealing with the issue
IIt is sate to say that the evaluation 01 the benefits 01 partial Integration as defined above are
sevrey underestima1eO Meanwhile the distortionSoenerated by the increase in the marginal
personal tax rates designed to ralse the revenue foregone by the dividend exclusion are fully
aCCltlunted for A good measure of the costs of integration together With a poor evaluation of the
benefils may very well be the reason why partial integration is Simulated 10 yield neg alive
eHikiency gains
On a different vein as most of Ihe model surveyed here Pereiras is a closed economy model It
is legitimate to wonder how the resullS would change If international capital flows were anowed
IThis almost certainly would affect financial crowding out since a good d~al of the capital inflows
could be used to finance public debt
For these various reasons the results should be treated as preliminary but they strongly I suggosl that the dynamic structure provides valuable new insights into the corporate tax integration SSU2
40
7 Concluding remarks I I I i
What has been acccmplished Ihis far The success of Ihe CGE research in laelding the challenge
of dynamic modelling can not be denied
I iGreat progress in the modelling of household behavior has been achieved Promising
developm~nts in the modeUing of production and government behavior have also been accomplished
Interesting innovations in the concept of equilibrium and computation techniques were discussed
The policy analysis was greatly enriched by considering transitional effects together with longer
run steadymiddotstate equilibrium paths generalized equal yield strategies accomodating different
financial crowding out impacts and generalized dynamic policy evaluation indicators
An important set of issues have been addresed by the models surveyed Traditional issues
focusing on the effects of capital taxation and consumption taxes have been pursued In terms of
capital taxation for example the intertemporal nature of the issue was enhanced by allowing an
optimal evolution of capital stock in the economy and forward looking optimal investment decisions
In Andersson (1987) GouldermiddotSummers (1987) and Pereira (1986b 1987d) this is coupled
with a improved treatment of several tax provisions like investment tax credits and depreciation
allowances In terms of the consumption taxes developments in the specification of Intertemporal
household behavior the introduction of overlapping generations and the modelling of
intergenerational links via bequest motives as well as a closer attention to demographic evolution
provided a much improved economic setting for the understanding of the several aspects of the
problem
Furthermore dynamic modelling has permitted the addressing of a set of new issues Policy
Issues ranging from dynamic tax policy analysis to the evaluation of exogenous changes in
government expenditures to the impact of different methods of financing government expenditure to
4 1
the importance of tinanciai croHrii1g out to 1he relevance of consumorS expectations in tne policy
results have been addressed
Despile all of the progress and in part due to such progress several avenues are wide open for
much neded additional research The following seem to be the most Interesting and promising
areasmiddot First the specincation of liquidity constraints to household behavior may help to obtain a
better understanding of policy chenges that affect directly the interest rates in the economy Second
major attention needs to be paid to the incorporafion of a roraign sector (with open capital markets)
Third some efforts are needed in terms of finding adequate computation techniques now that
dynamic modelling eally makes the curse of dimensionality worse than eVOrbull
The modelling of financial markets is the single most unsatisfactory speet of the dynamic
models surveyed here~ The problems associated with provlding a mo~ adequate treatment of
financial markets and in general endogenous and optimal financial deci$ions bull both at the corporate
and at the personal levelsmiddot have 10 be addressed That necessarily requires Introducing uncertainty
into the models To b adequate this must go well beyond the mere assumption of point expeetalions
with the whole array of modelling implementalion and evaluation problems thai generates
bull I i
42
REFERENCES
Abel A and O Blanchard 1983 An Intertemporal Model of Saving and Interest Econometrica
51 pp 675-92
~ Andersson K 1987 Tar-ation of Capital in Sweden - A General Equilibrium Model Ph D
Dissenation UnIversity of Lund Sweder
Auerbach A L KotlikoH 1983 National Savings Economic Welfare and the Structure of
Taxation in M Feldstein ed Behavioral Simulation Methods in Tax Policy- Analysis Chicago
University of Chicago Press
Auerbach A l Kotlikoff 1984 Social Security and the Economics of Demographic-
Transition in H Aaron and G Burtless eds Retirement and Economic Behavior Washington DC
Brookings Institution
Auerbach A L Kotlikoff 1987 Dynamic Fiscal Policy Cambridge Cambridge University
Press
Auerbach A L Kotlikof and J Skinner 1983 The Efficiency Gains from Dynamic Tax
Reform Internatiooaj Economic Reyiew 24 pp 81middot100
Ballard C 1983 Evaluation of the Consumption Tax with Dynamic General Equilibrium
Models Ph D Dissertation Stanford University
Ballard C 1987 Tax Policy and Consumer Foresight A General Equilibrium Simulation
Study Economic Inquiry 25 pp 267-284
Ballard C and L Goulder 1985 Consumption Taxes ForeSight and Welfare a Computatonal
General Equilibrium Analysis in J Piggott and J Whalley eds New Deyelooments in Apolied
General Equilibrium Analysis Cambridge Cambridge University Press
Ballard C D Fullerton J Shoven and J Whalley 1985 A General Equilibrium Model for
t 43
Tax Eoliey Eyalua~iQO NBER University or Chicago Press
Borges A (1987) bull A Survey of Compula1ional General Equilibnum Models CECD Economic
Slud~ 8
Bovenbig L 19B3 imperfect Capital Mobility in a Perfect Foresight Herberger Model
Mimea UC Br~eley
Bovenberg L 1984 Capitel Accumulation and Capital Immobility Qmiddottheory In a Dynamic General Equilibrium Framework PhD dissertation University of California at Berkeley t
Bovenberg L 19850 Dynamic General Equilibrium Tax Models with Adjustment Costs in A iManne ed bull 1985 EtQnomrc ECUjilbrium Model Formulation and Solution Mathematical
lProgramming Siudy 23 Arnsterdam NorthmiddotHoliand
Bovenberg L t 1986 Capital Income Taxation in Growing Open Eeondmies Journa of Pubtic
Eamp(omlcs 3t pp 347middot376 IBovenberg L ald W Keller 1981 Dynamics and Nonlinearlties in Applied General
Equilibrium Models Internal Report Mlmiddot81middot208 Department for Statistical Methods CBS bull
Vooiburg The Netherlands
Charnley C 1981 The Welfare Cost of C~pltal income Taxation In a Growing Economy
IJournal of Political ECQnQm~ 89 3 pp 468middot496 I
Charnley C 19810 Efficient Tax Reform in a Dynamic Model of General Equilibrium I Mimeo Yale University I Charnley C 1982 On the Optimal Taxation In Stylized Models 01 Inlartemporal General
Equilibrium Mimeo Yale University I Decaluw B and A Manens 1985 Pays en Oeveloppement at Modeles Calculables OEquilibre
General - Une Revue de al litterature empirique Monographie Sene 9 Volume 3 Centre de
Recherche et Developpemont Economlque Unrslt de Montreal
44
bull
I I
I
I
1
Erlich $ V Ginsburgh and L van der Heyden 1987 Wher~ do Real Wage Policies Lead
Belgium bull A General Equilibrium Analysis European Economic Reyiew (forthcoming)
air R and J Taylor 1983 Solution and Maximum Likelihood Estimation of Dynamic
Nonlinear Ratioant Expeclations Models Ecooometrica 51 pp 1169middot1186
Feltenstein A 1984 Money and Bonds In a Disaggregated Ec~nonYin Searl 1-1 J Shoven
Eds Aoolied ~eoeral EqJiliDCiulD tioalYsis Cambridge University Press Cambridge
Feltenstein A 19B6 A Dynamic General Equllibrlu Analysis of Financial Crowding Out
Theory with an Application to Australia Journal of Public Economics 31 No1
Fullerlon D R Gorden 1983 A Reexamination of Tax Distorllons in General Equilibrium
Models in M Feldstein Ed Behayora Simulation Methods in Tax pOlicy Analysis Chicago
University of Chipgo Press
Fvllerlon D J Shaven and J Whalley 1983 Replacing the US Income Tax with a
Progressive Consumption Tax JoulOal of Public ECQnomics 20 pp3middot23
Fullerton D Y Henderson J Shaven 1984 A COmparision of Methodologies in Empirical
General Equilibrium Models of Taxation in Searl H J Shoven Eds ~pQlied Geoaral Jguilibrium
lIoalyss Cambridge University Press Cambridge
Fullerton 0 A King J Shaven and J Whaney 1981 Corporate Tax Integration in the United
Statesmiddot a General Equiiibdufi1 Approach American Economic Review 71 pp~ 677~691
Galper H R Lucke E Tader 1985 Taxation portlor Choice and the Allocation of Capital A
General Equilibrium Approach Mimeo
Gill p W Murray M Saunders M Wright 1986 Users Guide for NPSOL (Version 40) A
Fortran Package for Nonlinear Programming T echnicsl Report SOL 86middot2 Deparlment of
Operations Research Stanford University
Glnsbur~h V and L van der Heyden 1985 General Equilibrium with Wage Rigidities An
45
AppJication to Belgium in A Manne ed t Economic EQumprit1m~ Made Formulation and $oJioo
Mathematical Programming Study 23 Amsterdam NorthmiddotHolland ~
Goulet J 1968 1 Adjustrent CO$1S in the Theory o(the Firm Reyiew of EconomiC Studies 35 bull
pp47middot55
Goulder L 1985 Tax-Financed versus BondmiddotFinancedGhanges in Governent Spending
bull Mimeo Harva(d UniversilYmiddot
Goulder l J Shaven and J Whalley 1983 Domestic Tax Policy and the Foreign Sector The
importance of altemative foreign policy formulations to results from a general equilibrium tax
analysis model In ~n methods in Ibe antico of income from caoia1 ed Martin
Feldstein Chicago University of Chicago riSS
Goulder l L Summers 1987 Tax Policy Asset Prices and 3rowtly ~
A General Equilibrium Analysis NBER Working Pper No 212B
Grandmont J 1982 Temporary General Equilibrium Theory Ch 20 in Handbook 01
Mathemalical Eccnornics Vol II NorthmiddotHoliand
Harberger A 1959 The Corporate Income Tax An empirical Appraisal In lax Revision
Comoendjum Voll House Cornmitee on Ways and Means Washington DC Govemment Printing
Office
Harberger A 1962 The Incidence of the Corporate Income Tax JOULn1 of Political
Economy 70 pp 215middot240
Harberger A 1964 Taxation Resource Aliocalrlon and Welfare fn lb Role of Dir bullbulll and
Indjresl Taxes io tbe Fede[al Reserve s~I~Ill Princeton Princeton Unlveresity Press
James J 1985 New Developments in the Applications of General Equilibrium models to
Economic History in Piggott aand Whalley Eds
Jorgenson D and K Yun 1984 Tax Policy and Capital Allocation Harvard Inslitute of
bull
I
I f
Economic Research WP 1107
Judd K 1965 ShonmiddotRun Analysis of Fiscal Policy in a Simple Perfect Foresight Model
Journal of poJiHcal 5CCJiW1l pp 298middot319 r
Upton D J Potbc J Sachsa nd L Summers 1982 Multiple Shooting in Rational
Expectations Models fuoometrlc 50 1329middot1333
Manne A ed 1985 fconQrnjc Eoullibrium Model FOrIDujaljQo and Solution Mathematical
Pogrammlng Study 23 Amsterdam NorthmiddotHolland
Manne A 1985 Or the Formulation and Solution 01 Economic Equilibrium Models In A
Manne ed ECQOQlli-Jr~ibmiddotum ~tQdel EQlJulailon and Solution Malhematical Programming
Study 23 Amsterdam NorthmiddotHolland
Mansur A and J Whalley 1984 Numerical Specification of Applied General Equilibrium
Models in Scarf H J Shoven Eds Aoplied General Eouillbrium AnaiysectIsect Cambridge Universily
Press Cambridge
Merrill 0 1972 Applications and Extensions of an Algorithm thaI Computes FixedmiddotPoints of
Cenain Upper Semimiddotcominuou Point to Set Mappings PhD Disserlation Unlvel$ity 01 Michigan
Negishi T 1960 Welfar Economics and Exislence I an Equnlbrlum for a Competitive
Economy MelroeCOOQ11ka 12 92middot97
Pereira A 1985 On the Existence and Uniqueness of Optimal Ou1put Path lor CRTS Firms in
Adjustment Costs Technologies Mimeo Stanford University
Pereira A 1985a Consumer Expectations and Ihe COSI 01 UllUty In an Intertemporat
Framework Mlmeo Stanford University
Pereira A 1985b Corporale Tax Integrallon and Government Deficits A Dynamic General
Equilibrium Anaiysls Mlmeo Stanford University
Pereir A 1987 On Ihe Computation of the Users Costs of Stocks Ecooomi (forthcoming)
47
Pereira A bull 1987b Equal Yield Alternatives and Government Oeficlts Mimeof Stanford
University
Pereira A 1987c A Dynamic Applied General ~uHlbrium Model for Tax Policy Evaluation
Comlletamp Docurnent2tion Mimeo University of California San Diego (in progress)
Pereira A 1987d Should the investment Tax Credit be Rei(ltroduced Mlmeo University
of California San Diego (ir progreSs)
Piggott J and J Whalley eds 1985 tJew QeyeQpments in Applied General EQuilibrium
bullenalysjs Cambridge Cambridge University Press
Preckel Pbull 1985 Alternative algorithms for computing economic equilibria In A Manne Ed
Radnor R 19amp3 Equllibrium under Uncertainty In K Arrow M Intrilligaor ods
I ibull
Robinson S 1 g86 Mullisectorai Models of Oeveloping CountrieS a Survey Oivision of
Agricullure and Resources UC Borkeley WP 401
Scarf H 1967 The approximation of fixed points of a continuous mapping ~MM Jouwal of
epplid Mathornallcs 15 pp 1328middot43
Scarf H with T Hansen 1973 00 the Computation of EconomiC Eguilibria New Haven Yale
University Press
Scarf H J Shoven Eds 1984 61l1llied Genera Elujlibrium 6nalyssect Cambridge University
~ Press Cambridge
Shoven J 1976 The Incidence and Efficiency Effects of Taxe on Income from Capital
Jurnal of ramie1 Economy 84 pp 1261-83
Shoven J 1983 Applied General Equilibrium Tax Modemng IMF Slaff PaoerS 30
Shoven J and LB Simon 1986 Share Repurchases and Acquisitions An Analysis of which
Firms Participate NBER Working Paper No 2243
bull
48
IShaven J and J Whalley 1972 ft General Equmbrium Calculation of lhe Effects of
Differential Taxation of Incoma from Capital in the US Jauroe of public Econgmics 1
pp281middot321 (
Shoven J bull and J Whalley 1973 General Equilibrium with Tax$s a computation procedure
and an existence pr00f ReviEhi of Economic Studies 40 pp 475middot490
Shoven J J Whalley 1977 Equal Yield Tax Alternatives A General Equilibrium
Computa~ui0nal Tfchnique J~HJrnal of PUblic EconomIcs 8 pp 211~224 bull
Shaven J J Whalley 1984 Applied GeneralmiddotEquilibrlum MOdels of Taxalion and
International Tr8de An Introduction and Survey Journal of Economic literature 23t pp
1007middot1051
Slemrod J 1980 A General Equilibrium Model of Capital Income Taxation Ph D bullDissertation Harvard Univnrsity
Slemrod J 1983 A General Equilibrium Model of Taxation with Endogenous Financial
Behavior in M Feldstein edbull Behavioral Simulation Methods In Tax Policy Analysis Chicago
UniVerSity of Chicago Press
Stone J 1985 Sequential Optimization and Complementarily Techniques for computing
Economic equilibrium in A Manne Ed
Summers L 1981 Taxation and Corporate Investment a q-Theory Approach BroQkinas
Summers L 1985 Taxation and the Size and ComposHion 01 the Capital Stock An Asset Price
Approach Mlmeo Harvard University
Whalley J 1975 A Geneal Equiloibrium Assessment of the 1973 United Kingdom tax
reform Economica 42 pp 139middot61
1JLTIllOS JIQll~)ltG PAPERS PUBLICADOS bull
I
I I
j
112 70 - LUCENA iJiCIIJo ~ IIEnvironrilent Honitoring and Orqanization Structure T (Setellibro 1987)
nI 71 - L~ENA Diogo A Note on the Representation of Information Stxueture (SeteJObro 1987)
nQ 72 - LUCENA D1CgO Envuorment 111)nitoring and Organization
Structure II (Setelilhro 1987)
ril 73 - AlmmE3 Antonio Pais e CAsPAR Yitor bull Tributacao Inoentivos e InvesUlllentos Analise Qualitativa noYellibro 19S7)
n2 74 - liAlltOSA rose Pedro Opt1lllal lIage R1gidity A SUggestedtlthodology to Teet tb Theory and an Application (Outubro 1987)
nI 75 - CABRAL L 11 B Three lIotes on SYJIIlletric Gemes witt itsYJIIlletric Equilibri (NoveJObro 1987)
n2 76 - LUCENlt D10gOTo Search or Not to Search (Fevereiro1967)
nQ 17 - SA Jorge asconcelos euroI A Theory of SynergyB (Fevereiro1987)
n2 76 - VILARES tanuei Jose Os Be IntenaMios IlIIportados ColIIO Factor de Produao (Julho 1987)
nQ 79 - SA J~rge VasCQncel03 e HoY to Copete and Co~unicate in tature Industrial Products_ (aio 1988)
n2 80 - ROIl Rafael Learning and Capacity Expansion in a New larket Under Uncertainty (Fevereiro 1988)
n2 81 PEREIRA Alfredo llarvilo Survey of Dynamic Computational General EquilibriUlll Hodel for Tax Policy Evaluation (Outubro 1987)
Qualquer in ormacao sabre o 1IIorlltlng Papers ja publlcados sera prestada pelo Secretariado de Apoio aos Docentes podendo os ~e3OSf
ser adquirido5 ne Secyo de Vlndas da Faculdade de EconollLia UNL na Travessa Esteviio Pinto Cltpolide - 1000 LISllOA
j
22
bonds privpte bonds and equity (or simply physical capital installed inmiddot different sectors) the
problem of allocation of saving among assets with potentially different returns arises
The papers surveyed var greatly with respect to l~e extent of their attention to the financial
side of the economy At Clne extreme are the models in ~anardmiddotFuliertonmiddotShoven-Whalley (1985)
Andersson (1987) Ballltld (1983) Ballard-Goulder (1985) Bovenberg (1985 1986) and
Erlich-Gjnsburgh-Heyden C19B which are devoted exclusively to the real side of the economy
In turn Auerbach-KotliKoff (1984 1987) Feltenstein (1984 1986) and Goulde (1985)
allow for government debt In these models saving finances changes in government debt and physical
capital Private and publlc 2ssets are perceived by the households as perfect substitutes The
allocation of saving merely adjusts to the relative demands for funds
Feltenstein (1984 1986) is the only model surveyed which introduces money Government deficits are financed by issuing money and bonds according to an exogenously given rule Money is
demanded by consumers for transaction motives and an exogenously given fraction as a store of value
On the other hand government bonds and physical capital are the vehicles for the intertemporal
transfer of wealth
Summers (1985) and GOlldermiddotSummers (1987) introduce a whole menu of financial assets shy
firm specific equity capital Different assets earn different rates of retum However such rates
are equal up to constant and exogenous sector speCific risk premia Therefore the introduction of
constant exogenous risk premia as helpful as it may be in the context of calibration does not solve
the main issue of the non-optimality of the allocation of saving Also talking about risk premia in a
deterministic context is somewhat unsatisfactory
Pereira (1986b 1987d) also introduces a whole menu of assets private and public bonds and
firm specific equity However all the assets are expected to yield the same rate of return and
therefore perc~jved as perfect substitutes The different asset types allow consideration of
23
exogenOU$~8tUequjty triO dividendlrc~enlion rufes and therefore seve-ral sources of investment
financing bond equity and retained earnings
The nonmiddotcptimalHy of the allocation of saving and the absence or exogenelty of corporate
financial rules is _a reftectlon of the limitations of the determInistic approach Either in a context of
perfect anticipation oi the future prices Of in general within the t$atm of point price
expectations all the papers follow a deterministic approach Under such circumstances consumers
either expect diHerent rales of return (inclusive of risk premium) across assets in which case
they wilt buy only one asset (that with highest rate) Of they expect equal rat~s of relurn in which
case Ihey are indifferent about the asset composition of Iheir portfoliO There is no way of
Iradingmiddotoff rales of return and risks to obtain an optimal interiof solution to the problem of the
allocation of saving
The most advanced contribution in the modelling of saving allocation in a CGE setting is due to
Slemrod (1980 1983) in the context of a static one-period model In his model consumers act
according to a two stage separable decision process They Hrsl decide on how much to save Then
they decide on the allocation of saving according to an Indirect uttlity function dependent on the rates
of ratum and variances offered by the different assets in the eoonomy The sourCe of riskiness in the
ampConory comes from an uncertain marginal product of capital On the other hand aside from
portfolio decisions the res of the economy is insulated from uncertainty
The most complete contribution in terms of the treatment of the corporate financial rules is
FulienonmiddotGordon (1983) In a variant 01 Ihe BailardmiddotFulienon-ShovenmiddotWhelley (1985) model
Fullerton and Gordon have capital intensity and optimal financial decisions Jointly deter~ined
through a two siage process The cosl of financing capital is minimized by trading off the tax
advantages of debt against the epected real bankrupcy coSls inherenl with high debVequity ratios
Given the optimal debtequity ratio the level of inveSlment Is chosen such that at Ihe margin the
24 l
bull
rerurn pn ~uity equas the return on bonds plus an exogenous risk premium
36 Market assumptions equilibrium and expectations
Virtually all of the papers arc characterized by Walrasian market clearing assumptions All
markets are perlect1y competi1ive Atomistic competition among agents 1s assumed even thoughmiddot only
a flnile number of agents is considered Virtually no mark~t disequilibria or price stickiness afe
considered The only exception is ErlichmiddotGinsburghmiddotHeyden (1987) bull In thei model of the Belgium
economy I the _wage rate is fixed in the short-run Therefore in the shOrlwrun desequilibrium in
the labor market will generate endogenous unemployment However in lh$ long-run all prices
Including the wage rata are Ilexible and accordingly ali markets ciear
Given the dynamic nature of behavlor in the economy markelmiddotcfearing prices in each period
depend on expectations of fture prices and on tax variables In the economy There are essentially
two ways at interpreting the economic equilibrium in such a dynamic conte~bull If future prices are
perfectly anticipated (io expectations are self-fulfilling) a perfect foresight equilibrium
prevaHs Then future actions are merely the implementation of current decisions for future
periods However jf ptica expecla110ns are not perfect (Le agents make mistakes w1th re$pect 10
future prices) then a temporary or shortrun equilibrium prevails Markets ciear and clearing
prices depend on fu1ure price expectations Current plans about the future are typically not
precisely implemented They will be revised as more or better information becomes available to the
economic agents See Grandmant (1982) for a comprehensive survey 01 the temporary equilibrium
literature
Wilh the exception of Gould (1985) and Pereira (19S6b 19S7d) all the models surveyed
adopt th~ concept of a perfect foresight equilibrium In turn BaliardmiddotGouldr (1965) considers a
J
2S
flexible am9unt of foresight In terms of the number of years over which price movements are ~
foreseen Pereiras model (1986b 1987b)) is flexible in a somewhat different way in that it can
indude any range of foresight from myopia to perfect foresight The choice between the perfct foresight and lemporary equilibrium is ultimately to be made on
philosophic grounds It can be argued that less than perfect expectations Imply that agents are
irralional in some way (see Auerbach-Kctiikoff (1987 p 10) However the reverse argument can
be made One can question wiether agents are reaHy rationar and perfectly knowledgable about
future prices
Recent evidene of Balird (1987) Ballard-Goulder (19851 Goulder (19851 and Pereira
(1986b 1987d) suggests that the choice in modemng expectations is an important one They show
that the degree of foresight inl0 thc future (ranging from perfect foresight to-myopic expectations)
may have dramatic impacts on the pollcy conclusions of Ihe model Accordingly the best research
strategy may b to design models which are flexible enough 10 allow for different rules regarding the
formation of expectations With the exception of the articles just mentioned sensitivity analysis
bullhave previously not been performed along this dimension
In terms of implementation the two concepts of equilibrium - perfect foresight and temporary
equilibrium have different implications The dimensionality of the equ1ibrium soll1ion algorIthm
is involved Suppose we have a model with ten markets to be run for a period of 50 years Aside
from normalization a perfect foresight model implies computing prices In 500 dimensions while a
-t~mporary equilibrium model requires sorving SO equilibria each in ten dimensions Given that
computational speed often varies with the cube of the number of dimensions the lemporary
equilibrium formulation is potentially strikingly more feasible However Ballard (1987) and
Goulder-Ballard (1985) have devoloped techniques to greatly speed the computation of a perfect
foresight equilibrium
26
j
I
4 Jrnplemenlalion and Issues on policy evaluallon
41 Calibration and equilibrium comparlslon
CGE models at typically parameterized by the use of a calibration procedure Some parameters
are exogenously given However some crucial parameters are determined in such a way that the
model replicates the data for a given base year See MansurmiddotWhalleY 11984) for an eXlensive
discussion of 1~i$ issue
CalibratIon in a dynamic context is generaliy interpreted as requiring two properties First
replication of base year data is required Second the model is parameterized to simulate an
intertemporal balanced growth path when Ihe base policy is maintaineq This Is the approach
followed by Ballard (1983) BallardmiddotGoulder (1985) Goulder (1985) and SummersmiddotGoulde
(1986) It follows the practice of BallardmiddotFuliertonmiddotShovenmiddotWhalley (1985) with their
sequential cquilibrium dynamics
Other authors Auerbach-Kotlikof (1983 1984 1987) Boyenberg (1984 1985 1986)
and Pereira (198Gb 1987d) follow whal we cali a qualitalive calibration The slructural
palametars are exogenously chosen so that lh economy follows a reasonable path into Ihe fulure
ThaI has 10 do wllh the fact thaI given the recursive nature 01 the dynamic economy and aside tom
such structural parameters only iniUa stock values are needed to run these models Given initial
conditions on the stocks of say private wealth capltal and government debt agents will optimize
and Ihereby generate a lirst round of net demands and equilibrium conditions In lurn the
equilibrium prices will determine the evolution of the stock variables into the next perIod
There are several potential problems with calibration in the conlext of dynamic models The
assumption of a steadymiddotstate growth ~ath in the base case can be questioned First while
27 I Ibullbull
steadymiddotstaU is a possibity it certainly is not the only meaningfull solution to dynamic models
Even in the case of a perfect foresight equilibrium the model implies an equilibrium path which i may or may not involve colanced growth The model not the modeller should dictate the nature 01 I themiddotbase case path Second in the context of a temporarY equilibrium path a steady state solutlcn is Inol a likely model QutcOn1tl In fact unlessmiddot expectations are stattc short run behavior consistent
with a steady-state evolution will in general not be generated On the other hand if static I expectations are self~fufjili[1g we have in fact a perfect foresight model Third even the base year
replication requirement may cause problems In Ihe cOntexl of temporary equilibrium Any
calibration parameter would be conditional on expectation rules which is probably an undesirable
feature
The nalur of the two-requirement calibration strategy is very muchin the spirit of the bull
traditional design of the comparision of alternative equilibria comparis_ion between a steady-state
base case on one hand and alternative paths including a transition period and a finat sleadyyenstate on
the other hand Pereira (198Gb 1987d) compare different (not necessarily steadystate)
equilibrium paths The arguments against such a procedure are based on the idea that the impact of
policy changes can be observed most easily since all departures from the steadymiddotstate can be
attributed to the altemalive polioy On the other hand the base case so dEfined as a steadymiddotstate is
consist~nl with previous work in a less dynamtc setting and Iherefore allows a common standard
for comparing model results
42 Computation algorithms
We are still al a stage in which basically each author uses a different computational technique
Th development of dynamic models - in parlicular wilh adjustment costs andior perfect foresight
bull
28
has corresponded with the decline in the use of fixed point algorithms In -fact given the relative
arge dimensions inevitably involved such algorithms lend to be very inefficient at the best and
chen prohibitively stow See Stone (1985) and Preckel (1965) for a comparative assessment of
different computation techniques Among the models surveyed only Ball1rdFulle(ton~
-ShavenmiddotWhalley (1985) and Feltensteln 1985 use Merrills variant of the fixed point
algorithm 1echnique
AuerbachmiddotKoHikoff (1963 1984 1967) follow a three stage procedure They first compute a
base case steady-state then a revised case steady state and finally transition path for the economy
berveen these two steady-states tn all stages a Gauss-Seidel iterative procedure is used
Ballard (1982) Ballard-Goulder (1965) Goulder (HiSS) and Summers-Goulder (1986)
~
use a method developed by Ballard and Goutder which is similar in many awects to the FairmiddotTaylQr bull
(1983) algorl1hm Short-run equilibria are calculated (using Merrills algorithm) parametric on
nrice expeC~lions The model is then iterated to generate self-fulfilling intertemporaf expectations
and the corresponding perfect foresight equilibrium In a relatively similar approach Andersson
1987 uses a simulation program SIMNON developed In the University of Lund This program
can handle two-paint boundary problems In a fashion consistent with the multiple shooting
algcrlthm (see Lipton-Poterba-Sachs-Eummers (1982))
Boyenbergs computational approach (1985 1985) differs from the other models surveyed in
that he relies heavily on analytical techniques Computations are done by using a dynamic version of
Johansons linearization method Being essentially determined by the continuous time nature of the
I I
f
model thiS linearization model has the disadvantage of confining the analysls to infinltesinal changes
around the base case equilibrium (see Bovenberg (1985) p 53) Pereira (1986b 1987d) uses an optimization algorithm NPSOL - developped by
rGillMurray-Saunders-Wrlght (1986)) This algorithm is used 10 compute the sequence of
I
29
Shortmiddotrun temporary equilibrium which makes up the intertemporal equilibrium path The
equilibrium conditions are seen as nonlinear equality constraints in the minimization of an artlcial
objective function The prices are normalized to the unit simplex by an additional linear equality I I 1 bull constraint
II ErlichmiddotGinsburgh-Heyden (1987) follow a unique approach in that they use a variant of the
optimization technique introduced by Negishl (1960) The economic equilibrium can be generated i
as a solution of a mathemallcal program Ihe objective function of which is a weighted sum of the
utility functions of the various agents while tho constraints set consists of the market dearing 1
conditions Ginsburgh-Heyden (1985) have extended Negishis resulllO tho case of downward price
lrigidities I
Finally the paper by Jorgonson-Yun (1984) is also unique in tIlat it is the only I
econometrically estimated model Different blocks for the consumption and production Side of the
-
model are separately estimated to provide the necessary structural parameters i
I The diversity of computation techniques is yet another indicator of the exploratory nature of the
body of literature surveyed in this article
43 Equal yield comparisions
The link between a base case and counteriactual simulations Is usually provided by the concept of
equal yield Shaven-Whaley (1977) discuss the meaning 01 equal yield in a general equilibrium
context when government is not allowed to run defICits Equal yield is interpreted to moan constant
public utility Government base case utility is maintained in the counferfactual experiments With
balanced budgets the concepl of equal yield is unambiguous The new equUibrium prices and the
balanced budge condition will del ermine the minimal expendilure and taxes needed to maintain base
30
case publi~ utility Accordingly in general equal yield is inconsistent with equal nominal tax
revenue Some change in tax revenue Is necessary Different tax replacement schemes ate
considered to assure that enough tax revenue is collected This is the approach essentially followed
by most 01 the papers surveyed Only Feltenstein (1985) Goulder (198) and Jorgenson-Yun
(1984) chose not to follow an equal yield strategy
The question is of how to Interprete lila concept of equal yield when the government is alowed to
fUn deflcLts The optimal leve of expenditure for base ease public trliUty can now be taAinanced
bond financed or financed by a mix of bonds and taxation We have several versions of equal yield In
particular equal yield may now be consistent with equal tax revenUE Furthermore some meaSure
ltgtf financial crowding out effects of government deficns can be inferred from the comparislon of the
several equal yield alternatives These issues are extensively discussed ~Pereira (1987b)
44 Price expectations and the dynamic generalization of compensation
Indicators
The sum of equivalent and compensation variations over households is the most widely used
aggregated measure of efficiency gains or losses In the context of $teady~state eomparisions andor
when complete future markets exist andlor perfect foresighl is assumed there are no difficulties
associated with the use of the standard Hicksian indicators In fact correct future prices are known
in these cases
If expectations are not selfmiddot fulfilling andlor future markels are not available a dynamic
generalization is necessary 8allardmiddotGoulder (1985) provide some steps in that direction by
defining an indicator that accomodates periods far into the future when households do not have
perfect foresight but a steady~state prevails On he other hand this issue is more fundamental in
i
[
I
i I I I I [
I I
bullI
31
the contex~ of a Gner1 lemporary equilibrium ftamawor~ In such Circumstances Pereira
(1986a) dev~lops a cnamlc generalization which is obtained as the present discounted value of a i i
sequence of short-fUn optimal expenditure functions consistent with a base case expected future I stream of IlWities I
32
5 Empirical evidenee from selected policy Issues
Dynamic tax models have generated several important results which escaped Sialic modellers
The following is a seleCpoundd set of issuos whIch stress the marginal benefits of dynam~ modelHng of
economic behavior in innovltlttive areas of analysis The discussion of a consumption tax emphasizes
the benefits of dynamic tusehold behavior and intergenerational aspects The study of the
flliminatlon or the re~intToduction of the investsnemt tax credits is made meaningful by the dynamic
modelling of production betlavor Modelling governmenl in a dynamic conlexts allows the modeller
to study the irrrpact of government deflcits Finally a dynamic framework lets us appreCiate the I
relevance of consumer expectations in terms of the evaluation of policy alternatives
i
i
51 Consumption Tax
t
By Its very nature a consumption tax can not be adequately investigated with a static model bull
Fullerion-Shoven-Whaicy (1983) use Ihe model of the US economy s described in
Ballard-Fullerton-Shoven-Whalley (1935) to evaluate the movement from the currenl US tax Isys1em to a progressive consumption tax Since their modeJ Incorporates a Jaborleisure choice
where leisure is an untaxed commodlty their results reflect the fact that both the consumption tax
and Ihe present system are dslortionary
Concentrating on the rntertemporaf distortions they show that sheltering more saving from lhe
current US income tax could improve econornic efficiency even if marginal tax rate increases are
necessary in order to maintain government fevenue~ At first you have a revenue shortfall
However the economy moves to a higher sustatned growth path and uitimatefy lower tax rates
cangenerate the same revenue path Also wages increase in the long run with this policy The
bull
bullbull ~
33 shyt
i
I
I
present value of welfare gi[lS for a potiey of complete savin9 deduction wilh marginal tate
adjustments (consumption lax) is simulated to be around $500 billion to $600 billion 01 1973
dollars
They also investiga1e th 1ngth of ttme it takes th~ economy to adjust to these policy changes
Roughly they estimate the long run to b thirty years although this figure Is very sensitive to the
specification of the savings elasticity - a crucial parameler in this model
52 Investment T~x Credit
Goulder-$ummers (1 S87) address the impact of eliminating the lnvestmenl Tax Cradit (ITC) on
intersectoral capital formation and on economic growth Their model isYa1ticularly adequate to
addies$ this issu~ in that is postulates a forward looking investment m9we with adjustment costs
They show that the eliminating the ITC causes a reduction in the rale of investment Inveslment is
estimated to fall by about 7 in the short-run and by about 12 in Ihe new long run steady state
In lurn a previous policy anncuncement lowers Ihe overall al1ractiveness of Investment and leads 10
a downward shift of the Investrrent profile On the other hand the combined effect of a revenue
neutral simultaneous efimiraron of the ITC and reduction of the corporate tax rates is a long run
reduction ef the capital Sieck by 35 This pattern suggests that a revenue neutral increament in
both Ihe corporale tax rates and the inveSlment lax credits would be preferable 10 the formula
adopted in Ihe US Tax Reform Act of 1986
Pereira (1987d) addresses the impact of the ITe from Ihe oland point of Ihe curTenllax syslem
under the Tax Reform Act of 1986 Given Ihe concern over the potential depressiVe effects upon
savings and investment of the new tax system in conjunction with deficit reduC1ion mechanisms of
the GrammmiddotRudman type a pertinent question is should Ih lTC be e-introduced In the context
bull
34
-of his mod~1 of the US economy Pereira focuses on the trade-off between the distortion in the
intersectoral allocation of capital induced by the lTC the lower tax revenues it generates and
potential financial crowding-out effects on one hand and the positive effects of lowering the relative
price of new capital goods on the other hand Preliminary results confirm the qualitative results in
Goulder-Summers (1987) suggestting a welfare gain of about 2 of the present value of GNP in
the best scenario of absence of replacement taxes Furthermore preliminary results show an
important time pattern to welfare gains with the average benefits increasing the further you look
into the future This pattern is due to the presence of constraints to intersectoral mobility of capital
and inertia in the adjustment towards the optimal capital levels as reflected by the presence of
adjustment costs
I
I
bullAuerbach-Katlikaff (1gB) in the context of their intergenerational model of the US economy
consider the impact on savings capital formation and interest rates of deficit policies and balanced
budget increases in government consumption They conclude that deficit finance and government
consumption can significa-ntly crowd out capital formation and lower the welfare of future
generations However crowding out from deficit finance is a very slow process because it results
from increased government spending over potentially long horizons Also deficit policies may
substantially influence the longmiddotrun interest rates while leaving the short-run rates essentially
unaffected Finally and in opposition to the central role adjustment costs seem to play in both
53 Financial crowding outeff~cts of government deficits
Goulder-Summers and Pereiras models Auerbach-Kotlikoff suggest that the time path of interest
rates induced by a policy of deficit financing seems to be insensitive to the presence of adjustment
costs
bull
35
54 The role 01 consumer expectatlons
The expectations analysis has uncovered one of the benefits of dynamic modelling
Ballard-Goulder (1985) find that the appeal of adopting a consumption tax depends on the level of
foresight possessed by the consumers Furthermore additional foresight may be welfare worsening
This is a second best type of result This tends 10 occur under policies that lead to capital deepening
and declining rate of return to capital over time To the extent that consumers have more foresight
they will be beller equippod to anticipate the fall in the rental price of capital Consequently if a
saving incentive is enacted people save more with myopia than with perfect foresight Given the
exi~tence of taxes on capita and the discrepancy between the private an~--sccial returns to capital
the greater savings with myopia is beller socially Ballard-Goulder find that the welfare gains
from a consumption tax are reduced by about 10 when we move from myopia to farleet foresight
Therefore the attractivenness of adopting a consumption tax seems fairly [obust across these
foresight specifications
I
bull
36
6 Tbe power and weaknesses 01 dynamic modelling the example 01 corporale
tax Integrallon
The Issue of corporate tax integration allows us to sh~w the stre-ngths and weaknesses or the
dynamic approach Empirical evidence using CGE techniques indicates that depending on the precise
scheme of integration and the 1ax replacement methods integratiOn may have substantial effects In
Ihe work of FuliertonmiddotKingmiddotShoven-Whalley (1980 1981) total integration was found to yield an bull
annual static eHielency gain of $4 to $8 billion in 1973 dollars Simulated dynamic gains may be as
large as $695 billion or abcu 14 of the present value of tulUre consumption and leisure in the
US economy These gains result primarily from interindustry reallocalions of Investment and an
improved inter~emporal allocation of consumption
Mora recent work emphasizes hOw consumers asset portfolio decisions and firms finanCial
decisions affect the efficiency gains from integration Slemrod (1980) focusing on consumers
asset ponfoio deCisions finds static efficiency gains which are about twice -as Iarge as those
repOrted iiy Fulierton-King-ShovenWhalley (1961) FuUrlon-Gurdon (1983) focus on the
firms financial decisions They roporl efficiency gains of 6 of GNP from the elimination of the
tax distor1ions favoring debt However when they eliminate the corporate tax and repiace it with
increaSed persona income taxes addlHonal dls10t1ions are created in the optimal labot~leisure
decisions These distortions tend to dominate the analysis slJ(h that the overall effects of complete
integration are very modest Galpr-LuckmiddotTodr (1986) address simultaneously consumers
asset pOrlfollo decisions and the firms financial decisions They also find very modest efficiency
gains from integralion
The above results are important but may be severely biased There are severa aspects of
economic behavior and modelling crucial for the study of Income tax Integration whiCh have no been
i 37
captured innny of the above papers One aspect is the absence of government deficitsmiddot a balanced
budget is assumed It is true Ihal resource crowding OUI is caplured In these mOdels but financial
crowding out induced by government spending is nOI Second investment is not derived from
optimtztlcn behavior Investment behavior passively accomodates endogenous saving decisions As
a consequence the differential impacts of policies in the incentives to save and to invest are not
captured Aso full capital mobility across secors is assumed with instantaneous capital
adjustmenlS towards optimal levels This assumption rules qut different costs of capital across
seclors and therefore differentiated reactions to tax policies changes Finallyhe modelling of both
gov9rnmerlt deficits and of endogenous real and financial investment decisions necessitates the
~nsideraiion of a dynamilt framework and the introduction of financiaJ assets govemment bonds and
iprivate financial assets A dynamic framework also highlights the efficiency effects of Integration on r
ithe ptimal intertemporal decisions The above models are either sIalic (Slemrods and
GalpermiddotltJckemiddotTode(sj or a dynamic sequ~nce of otherwise Slatic model bullbull I Pereir a (1986b) develops a dynamic applied general equilibrium model bull to study the tbull
efficiency effects of integration on the growth and allocation of investmenl across sectors Special
Iattention is paid to the structural affeelS of resource and financial crowding oul induced by
gove mect spending and deficits and to the real and financial Investment reactions across industries
to both tax integration and flnancial crowding out His mode departs from previous work on income
lax Integraticn and for that matler from most 01 the CGIE literalure for tax policy evaluation in
several direcllons It encompasses an endogenous sequential equilibrium struelure founded on
dynamic behavior with flexible expectations Government deficits are optlmaUy determined
Investment decisions are tward looking and the resull of optimizing behavior Several financial
assetsmiddot public and privatemiddot are considered
Simulation resulls Indicale that the welfare gains from Integration under large deficits are at
32
best lJlodest when measured in terms of the GNP The elimination of the corporate tax and its I replacemenl by increased income lax rates yields long run benefits which are never larger than
2 of the presenl yalue ct telr consumption and leisure onder the previous tax regime and 1
under the current tax regime However the short run w~lare effects of full integration tend 10 be
very low and in in some scenarios even negative This is a flew intertemporar pattern of efficiency Ieffects which reflects an adjJsiment lag in the interindustry investment decisions due to the
existence of costs of adjustment On the of her hand partial integration achieved by excluding I dividends from the corporale lax base syslematically yields negalive efficlencyeffeclS In Ihe long Inm these can be as high as 3 of present value of the future value of consumption and leisure This
is a new second best effect suggesllng Ihat less than complete integration may have perverse I efficiency effects
The dynamic slructure of PerclrBs model snowing for an improved freatment of real investment
declsions and government deficts provides a potential setting for a more accurate measure of the
costs and benefits of inlegration The simulation resullS suggest lower benefits and an intertempora Ibull
pattern of growing benefits
Simulation results also clearly suggest robust negative effects from partial integration How
reliable is this result If dJvidends were deductible from the corporate lax base (as are interest
payments) the preferenlial Ireatmant of debl over equily would be ellminaled Corporallons should
be expecled 10 react by decreasing the optimal deWeqully rallo However corporale financial
gecisions are exogenously set In Pereiras model as in aft of the other models surveyed that go as far
as dealing with the issue
Also withdrawing dividends from Ihe corporate tax base is a way of eliminating Ihe double
taxation of dividends at belh Ihe corporate and personnal Income levels How will the optimal
aliocellon of household saving accomodate that change in Ihe relalive return to corporale equity
39 I
t However hOusehold portfolio decisions are exogenouly set in Pereiras model as in all of the other
models surveyeltj go as far as dealing with the issue
IIt is sate to say that the evaluation 01 the benefits 01 partial Integration as defined above are
sevrey underestima1eO Meanwhile the distortionSoenerated by the increase in the marginal
personal tax rates designed to ralse the revenue foregone by the dividend exclusion are fully
aCCltlunted for A good measure of the costs of integration together With a poor evaluation of the
benefils may very well be the reason why partial integration is Simulated 10 yield neg alive
eHikiency gains
On a different vein as most of Ihe model surveyed here Pereiras is a closed economy model It
is legitimate to wonder how the resullS would change If international capital flows were anowed
IThis almost certainly would affect financial crowding out since a good d~al of the capital inflows
could be used to finance public debt
For these various reasons the results should be treated as preliminary but they strongly I suggosl that the dynamic structure provides valuable new insights into the corporate tax integration SSU2
40
7 Concluding remarks I I I i
What has been acccmplished Ihis far The success of Ihe CGE research in laelding the challenge
of dynamic modelling can not be denied
I iGreat progress in the modelling of household behavior has been achieved Promising
developm~nts in the modeUing of production and government behavior have also been accomplished
Interesting innovations in the concept of equilibrium and computation techniques were discussed
The policy analysis was greatly enriched by considering transitional effects together with longer
run steadymiddotstate equilibrium paths generalized equal yield strategies accomodating different
financial crowding out impacts and generalized dynamic policy evaluation indicators
An important set of issues have been addresed by the models surveyed Traditional issues
focusing on the effects of capital taxation and consumption taxes have been pursued In terms of
capital taxation for example the intertemporal nature of the issue was enhanced by allowing an
optimal evolution of capital stock in the economy and forward looking optimal investment decisions
In Andersson (1987) GouldermiddotSummers (1987) and Pereira (1986b 1987d) this is coupled
with a improved treatment of several tax provisions like investment tax credits and depreciation
allowances In terms of the consumption taxes developments in the specification of Intertemporal
household behavior the introduction of overlapping generations and the modelling of
intergenerational links via bequest motives as well as a closer attention to demographic evolution
provided a much improved economic setting for the understanding of the several aspects of the
problem
Furthermore dynamic modelling has permitted the addressing of a set of new issues Policy
Issues ranging from dynamic tax policy analysis to the evaluation of exogenous changes in
government expenditures to the impact of different methods of financing government expenditure to
4 1
the importance of tinanciai croHrii1g out to 1he relevance of consumorS expectations in tne policy
results have been addressed
Despile all of the progress and in part due to such progress several avenues are wide open for
much neded additional research The following seem to be the most Interesting and promising
areasmiddot First the specincation of liquidity constraints to household behavior may help to obtain a
better understanding of policy chenges that affect directly the interest rates in the economy Second
major attention needs to be paid to the incorporafion of a roraign sector (with open capital markets)
Third some efforts are needed in terms of finding adequate computation techniques now that
dynamic modelling eally makes the curse of dimensionality worse than eVOrbull
The modelling of financial markets is the single most unsatisfactory speet of the dynamic
models surveyed here~ The problems associated with provlding a mo~ adequate treatment of
financial markets and in general endogenous and optimal financial deci$ions bull both at the corporate
and at the personal levelsmiddot have 10 be addressed That necessarily requires Introducing uncertainty
into the models To b adequate this must go well beyond the mere assumption of point expeetalions
with the whole array of modelling implementalion and evaluation problems thai generates
bull I i
42
REFERENCES
Abel A and O Blanchard 1983 An Intertemporal Model of Saving and Interest Econometrica
51 pp 675-92
~ Andersson K 1987 Tar-ation of Capital in Sweden - A General Equilibrium Model Ph D
Dissenation UnIversity of Lund Sweder
Auerbach A L KotlikoH 1983 National Savings Economic Welfare and the Structure of
Taxation in M Feldstein ed Behavioral Simulation Methods in Tax Policy- Analysis Chicago
University of Chicago Press
Auerbach A l Kotlikoff 1984 Social Security and the Economics of Demographic-
Transition in H Aaron and G Burtless eds Retirement and Economic Behavior Washington DC
Brookings Institution
Auerbach A L Kotlikoff 1987 Dynamic Fiscal Policy Cambridge Cambridge University
Press
Auerbach A L Kotlikof and J Skinner 1983 The Efficiency Gains from Dynamic Tax
Reform Internatiooaj Economic Reyiew 24 pp 81middot100
Ballard C 1983 Evaluation of the Consumption Tax with Dynamic General Equilibrium
Models Ph D Dissertation Stanford University
Ballard C 1987 Tax Policy and Consumer Foresight A General Equilibrium Simulation
Study Economic Inquiry 25 pp 267-284
Ballard C and L Goulder 1985 Consumption Taxes ForeSight and Welfare a Computatonal
General Equilibrium Analysis in J Piggott and J Whalley eds New Deyelooments in Apolied
General Equilibrium Analysis Cambridge Cambridge University Press
Ballard C D Fullerton J Shoven and J Whalley 1985 A General Equilibrium Model for
t 43
Tax Eoliey Eyalua~iQO NBER University or Chicago Press
Borges A (1987) bull A Survey of Compula1ional General Equilibnum Models CECD Economic
Slud~ 8
Bovenbig L 19B3 imperfect Capital Mobility in a Perfect Foresight Herberger Model
Mimea UC Br~eley
Bovenberg L 1984 Capitel Accumulation and Capital Immobility Qmiddottheory In a Dynamic General Equilibrium Framework PhD dissertation University of California at Berkeley t
Bovenberg L 19850 Dynamic General Equilibrium Tax Models with Adjustment Costs in A iManne ed bull 1985 EtQnomrc ECUjilbrium Model Formulation and Solution Mathematical
lProgramming Siudy 23 Arnsterdam NorthmiddotHoliand
Bovenberg L t 1986 Capital Income Taxation in Growing Open Eeondmies Journa of Pubtic
Eamp(omlcs 3t pp 347middot376 IBovenberg L ald W Keller 1981 Dynamics and Nonlinearlties in Applied General
Equilibrium Models Internal Report Mlmiddot81middot208 Department for Statistical Methods CBS bull
Vooiburg The Netherlands
Charnley C 1981 The Welfare Cost of C~pltal income Taxation In a Growing Economy
IJournal of Political ECQnQm~ 89 3 pp 468middot496 I
Charnley C 19810 Efficient Tax Reform in a Dynamic Model of General Equilibrium I Mimeo Yale University I Charnley C 1982 On the Optimal Taxation In Stylized Models 01 Inlartemporal General
Equilibrium Mimeo Yale University I Decaluw B and A Manens 1985 Pays en Oeveloppement at Modeles Calculables OEquilibre
General - Une Revue de al litterature empirique Monographie Sene 9 Volume 3 Centre de
Recherche et Developpemont Economlque Unrslt de Montreal
44
bull
I I
I
I
1
Erlich $ V Ginsburgh and L van der Heyden 1987 Wher~ do Real Wage Policies Lead
Belgium bull A General Equilibrium Analysis European Economic Reyiew (forthcoming)
air R and J Taylor 1983 Solution and Maximum Likelihood Estimation of Dynamic
Nonlinear Ratioant Expeclations Models Ecooometrica 51 pp 1169middot1186
Feltenstein A 1984 Money and Bonds In a Disaggregated Ec~nonYin Searl 1-1 J Shoven
Eds Aoolied ~eoeral EqJiliDCiulD tioalYsis Cambridge University Press Cambridge
Feltenstein A 19B6 A Dynamic General Equllibrlu Analysis of Financial Crowding Out
Theory with an Application to Australia Journal of Public Economics 31 No1
Fullerlon D R Gorden 1983 A Reexamination of Tax Distorllons in General Equilibrium
Models in M Feldstein Ed Behayora Simulation Methods in Tax pOlicy Analysis Chicago
University of Chipgo Press
Fvllerlon D J Shaven and J Whalley 1983 Replacing the US Income Tax with a
Progressive Consumption Tax JoulOal of Public ECQnomics 20 pp3middot23
Fullerton D Y Henderson J Shaven 1984 A COmparision of Methodologies in Empirical
General Equilibrium Models of Taxation in Searl H J Shoven Eds ~pQlied Geoaral Jguilibrium
lIoalyss Cambridge University Press Cambridge
Fullerton 0 A King J Shaven and J Whaney 1981 Corporate Tax Integration in the United
Statesmiddot a General Equiiibdufi1 Approach American Economic Review 71 pp~ 677~691
Galper H R Lucke E Tader 1985 Taxation portlor Choice and the Allocation of Capital A
General Equilibrium Approach Mimeo
Gill p W Murray M Saunders M Wright 1986 Users Guide for NPSOL (Version 40) A
Fortran Package for Nonlinear Programming T echnicsl Report SOL 86middot2 Deparlment of
Operations Research Stanford University
Glnsbur~h V and L van der Heyden 1985 General Equilibrium with Wage Rigidities An
45
AppJication to Belgium in A Manne ed t Economic EQumprit1m~ Made Formulation and $oJioo
Mathematical Programming Study 23 Amsterdam NorthmiddotHolland ~
Goulet J 1968 1 Adjustrent CO$1S in the Theory o(the Firm Reyiew of EconomiC Studies 35 bull
pp47middot55
Goulder L 1985 Tax-Financed versus BondmiddotFinancedGhanges in Governent Spending
bull Mimeo Harva(d UniversilYmiddot
Goulder l J Shaven and J Whalley 1983 Domestic Tax Policy and the Foreign Sector The
importance of altemative foreign policy formulations to results from a general equilibrium tax
analysis model In ~n methods in Ibe antico of income from caoia1 ed Martin
Feldstein Chicago University of Chicago riSS
Goulder l L Summers 1987 Tax Policy Asset Prices and 3rowtly ~
A General Equilibrium Analysis NBER Working Pper No 212B
Grandmont J 1982 Temporary General Equilibrium Theory Ch 20 in Handbook 01
Mathemalical Eccnornics Vol II NorthmiddotHoliand
Harberger A 1959 The Corporate Income Tax An empirical Appraisal In lax Revision
Comoendjum Voll House Cornmitee on Ways and Means Washington DC Govemment Printing
Office
Harberger A 1962 The Incidence of the Corporate Income Tax JOULn1 of Political
Economy 70 pp 215middot240
Harberger A 1964 Taxation Resource Aliocalrlon and Welfare fn lb Role of Dir bullbulll and
Indjresl Taxes io tbe Fede[al Reserve s~I~Ill Princeton Princeton Unlveresity Press
James J 1985 New Developments in the Applications of General Equilibrium models to
Economic History in Piggott aand Whalley Eds
Jorgenson D and K Yun 1984 Tax Policy and Capital Allocation Harvard Inslitute of
bull
I
I f
Economic Research WP 1107
Judd K 1965 ShonmiddotRun Analysis of Fiscal Policy in a Simple Perfect Foresight Model
Journal of poJiHcal 5CCJiW1l pp 298middot319 r
Upton D J Potbc J Sachsa nd L Summers 1982 Multiple Shooting in Rational
Expectations Models fuoometrlc 50 1329middot1333
Manne A ed 1985 fconQrnjc Eoullibrium Model FOrIDujaljQo and Solution Mathematical
Pogrammlng Study 23 Amsterdam NorthmiddotHolland
Manne A 1985 Or the Formulation and Solution 01 Economic Equilibrium Models In A
Manne ed ECQOQlli-Jr~ibmiddotum ~tQdel EQlJulailon and Solution Malhematical Programming
Study 23 Amsterdam NorthmiddotHolland
Mansur A and J Whalley 1984 Numerical Specification of Applied General Equilibrium
Models in Scarf H J Shoven Eds Aoplied General Eouillbrium AnaiysectIsect Cambridge Universily
Press Cambridge
Merrill 0 1972 Applications and Extensions of an Algorithm thaI Computes FixedmiddotPoints of
Cenain Upper Semimiddotcominuou Point to Set Mappings PhD Disserlation Unlvel$ity 01 Michigan
Negishi T 1960 Welfar Economics and Exislence I an Equnlbrlum for a Competitive
Economy MelroeCOOQ11ka 12 92middot97
Pereira A 1985 On the Existence and Uniqueness of Optimal Ou1put Path lor CRTS Firms in
Adjustment Costs Technologies Mimeo Stanford University
Pereira A 1985a Consumer Expectations and Ihe COSI 01 UllUty In an Intertemporat
Framework Mlmeo Stanford University
Pereira A 1985b Corporale Tax Integrallon and Government Deficits A Dynamic General
Equilibrium Anaiysls Mlmeo Stanford University
Pereir A 1987 On Ihe Computation of the Users Costs of Stocks Ecooomi (forthcoming)
47
Pereira A bull 1987b Equal Yield Alternatives and Government Oeficlts Mimeof Stanford
University
Pereira A 1987c A Dynamic Applied General ~uHlbrium Model for Tax Policy Evaluation
Comlletamp Docurnent2tion Mimeo University of California San Diego (in progress)
Pereira A 1987d Should the investment Tax Credit be Rei(ltroduced Mlmeo University
of California San Diego (ir progreSs)
Piggott J and J Whalley eds 1985 tJew QeyeQpments in Applied General EQuilibrium
bullenalysjs Cambridge Cambridge University Press
Preckel Pbull 1985 Alternative algorithms for computing economic equilibria In A Manne Ed
Radnor R 19amp3 Equllibrium under Uncertainty In K Arrow M Intrilligaor ods
I ibull
Robinson S 1 g86 Mullisectorai Models of Oeveloping CountrieS a Survey Oivision of
Agricullure and Resources UC Borkeley WP 401
Scarf H 1967 The approximation of fixed points of a continuous mapping ~MM Jouwal of
epplid Mathornallcs 15 pp 1328middot43
Scarf H with T Hansen 1973 00 the Computation of EconomiC Eguilibria New Haven Yale
University Press
Scarf H J Shoven Eds 1984 61l1llied Genera Elujlibrium 6nalyssect Cambridge University
~ Press Cambridge
Shoven J 1976 The Incidence and Efficiency Effects of Taxe on Income from Capital
Jurnal of ramie1 Economy 84 pp 1261-83
Shoven J 1983 Applied General Equilibrium Tax Modemng IMF Slaff PaoerS 30
Shoven J and LB Simon 1986 Share Repurchases and Acquisitions An Analysis of which
Firms Participate NBER Working Paper No 2243
bull
48
IShaven J and J Whalley 1972 ft General Equmbrium Calculation of lhe Effects of
Differential Taxation of Incoma from Capital in the US Jauroe of public Econgmics 1
pp281middot321 (
Shoven J bull and J Whalley 1973 General Equilibrium with Tax$s a computation procedure
and an existence pr00f ReviEhi of Economic Studies 40 pp 475middot490
Shoven J J Whalley 1977 Equal Yield Tax Alternatives A General Equilibrium
Computa~ui0nal Tfchnique J~HJrnal of PUblic EconomIcs 8 pp 211~224 bull
Shaven J J Whalley 1984 Applied GeneralmiddotEquilibrlum MOdels of Taxalion and
International Tr8de An Introduction and Survey Journal of Economic literature 23t pp
1007middot1051
Slemrod J 1980 A General Equilibrium Model of Capital Income Taxation Ph D bullDissertation Harvard Univnrsity
Slemrod J 1983 A General Equilibrium Model of Taxation with Endogenous Financial
Behavior in M Feldstein edbull Behavioral Simulation Methods In Tax Policy Analysis Chicago
UniVerSity of Chicago Press
Stone J 1985 Sequential Optimization and Complementarily Techniques for computing
Economic equilibrium in A Manne Ed
Summers L 1981 Taxation and Corporate Investment a q-Theory Approach BroQkinas
Summers L 1985 Taxation and the Size and ComposHion 01 the Capital Stock An Asset Price
Approach Mlmeo Harvard University
Whalley J 1975 A Geneal Equiloibrium Assessment of the 1973 United Kingdom tax
reform Economica 42 pp 139middot61
1JLTIllOS JIQll~)ltG PAPERS PUBLICADOS bull
I
I I
j
112 70 - LUCENA iJiCIIJo ~ IIEnvironrilent Honitoring and Orqanization Structure T (Setellibro 1987)
nI 71 - L~ENA Diogo A Note on the Representation of Information Stxueture (SeteJObro 1987)
nQ 72 - LUCENA D1CgO Envuorment 111)nitoring and Organization
Structure II (Setelilhro 1987)
ril 73 - AlmmE3 Antonio Pais e CAsPAR Yitor bull Tributacao Inoentivos e InvesUlllentos Analise Qualitativa noYellibro 19S7)
n2 74 - liAlltOSA rose Pedro Opt1lllal lIage R1gidity A SUggestedtlthodology to Teet tb Theory and an Application (Outubro 1987)
nI 75 - CABRAL L 11 B Three lIotes on SYJIIlletric Gemes witt itsYJIIlletric Equilibri (NoveJObro 1987)
n2 76 - LUCENlt D10gOTo Search or Not to Search (Fevereiro1967)
nQ 17 - SA Jorge asconcelos euroI A Theory of SynergyB (Fevereiro1987)
n2 76 - VILARES tanuei Jose Os Be IntenaMios IlIIportados ColIIO Factor de Produao (Julho 1987)
nQ 79 - SA J~rge VasCQncel03 e HoY to Copete and Co~unicate in tature Industrial Products_ (aio 1988)
n2 80 - ROIl Rafael Learning and Capacity Expansion in a New larket Under Uncertainty (Fevereiro 1988)
n2 81 PEREIRA Alfredo llarvilo Survey of Dynamic Computational General EquilibriUlll Hodel for Tax Policy Evaluation (Outubro 1987)
Qualquer in ormacao sabre o 1IIorlltlng Papers ja publlcados sera prestada pelo Secretariado de Apoio aos Docentes podendo os ~e3OSf
ser adquirido5 ne Secyo de Vlndas da Faculdade de EconollLia UNL na Travessa Esteviio Pinto Cltpolide - 1000 LISllOA
23
exogenOU$~8tUequjty triO dividendlrc~enlion rufes and therefore seve-ral sources of investment
financing bond equity and retained earnings
The nonmiddotcptimalHy of the allocation of saving and the absence or exogenelty of corporate
financial rules is _a reftectlon of the limitations of the determInistic approach Either in a context of
perfect anticipation oi the future prices Of in general within the t$atm of point price
expectations all the papers follow a deterministic approach Under such circumstances consumers
either expect diHerent rales of return (inclusive of risk premium) across assets in which case
they wilt buy only one asset (that with highest rate) Of they expect equal rat~s of relurn in which
case Ihey are indifferent about the asset composition of Iheir portfoliO There is no way of
Iradingmiddotoff rales of return and risks to obtain an optimal interiof solution to the problem of the
allocation of saving
The most advanced contribution in the modelling of saving allocation in a CGE setting is due to
Slemrod (1980 1983) in the context of a static one-period model In his model consumers act
according to a two stage separable decision process They Hrsl decide on how much to save Then
they decide on the allocation of saving according to an Indirect uttlity function dependent on the rates
of ratum and variances offered by the different assets in the eoonomy The sourCe of riskiness in the
ampConory comes from an uncertain marginal product of capital On the other hand aside from
portfolio decisions the res of the economy is insulated from uncertainty
The most complete contribution in terms of the treatment of the corporate financial rules is
FulienonmiddotGordon (1983) In a variant 01 Ihe BailardmiddotFulienon-ShovenmiddotWhelley (1985) model
Fullerton and Gordon have capital intensity and optimal financial decisions Jointly deter~ined
through a two siage process The cosl of financing capital is minimized by trading off the tax
advantages of debt against the epected real bankrupcy coSls inherenl with high debVequity ratios
Given the optimal debtequity ratio the level of inveSlment Is chosen such that at Ihe margin the
24 l
bull
rerurn pn ~uity equas the return on bonds plus an exogenous risk premium
36 Market assumptions equilibrium and expectations
Virtually all of the papers arc characterized by Walrasian market clearing assumptions All
markets are perlect1y competi1ive Atomistic competition among agents 1s assumed even thoughmiddot only
a flnile number of agents is considered Virtually no mark~t disequilibria or price stickiness afe
considered The only exception is ErlichmiddotGinsburghmiddotHeyden (1987) bull In thei model of the Belgium
economy I the _wage rate is fixed in the short-run Therefore in the shOrlwrun desequilibrium in
the labor market will generate endogenous unemployment However in lh$ long-run all prices
Including the wage rata are Ilexible and accordingly ali markets ciear
Given the dynamic nature of behavlor in the economy markelmiddotcfearing prices in each period
depend on expectations of fture prices and on tax variables In the economy There are essentially
two ways at interpreting the economic equilibrium in such a dynamic conte~bull If future prices are
perfectly anticipated (io expectations are self-fulfilling) a perfect foresight equilibrium
prevaHs Then future actions are merely the implementation of current decisions for future
periods However jf ptica expecla110ns are not perfect (Le agents make mistakes w1th re$pect 10
future prices) then a temporary or shortrun equilibrium prevails Markets ciear and clearing
prices depend on fu1ure price expectations Current plans about the future are typically not
precisely implemented They will be revised as more or better information becomes available to the
economic agents See Grandmant (1982) for a comprehensive survey 01 the temporary equilibrium
literature
Wilh the exception of Gould (1985) and Pereira (19S6b 19S7d) all the models surveyed
adopt th~ concept of a perfect foresight equilibrium In turn BaliardmiddotGouldr (1965) considers a
J
2S
flexible am9unt of foresight In terms of the number of years over which price movements are ~
foreseen Pereiras model (1986b 1987b)) is flexible in a somewhat different way in that it can
indude any range of foresight from myopia to perfect foresight The choice between the perfct foresight and lemporary equilibrium is ultimately to be made on
philosophic grounds It can be argued that less than perfect expectations Imply that agents are
irralional in some way (see Auerbach-Kctiikoff (1987 p 10) However the reverse argument can
be made One can question wiether agents are reaHy rationar and perfectly knowledgable about
future prices
Recent evidene of Balird (1987) Ballard-Goulder (19851 Goulder (19851 and Pereira
(1986b 1987d) suggests that the choice in modemng expectations is an important one They show
that the degree of foresight inl0 thc future (ranging from perfect foresight to-myopic expectations)
may have dramatic impacts on the pollcy conclusions of Ihe model Accordingly the best research
strategy may b to design models which are flexible enough 10 allow for different rules regarding the
formation of expectations With the exception of the articles just mentioned sensitivity analysis
bullhave previously not been performed along this dimension
In terms of implementation the two concepts of equilibrium - perfect foresight and temporary
equilibrium have different implications The dimensionality of the equ1ibrium soll1ion algorIthm
is involved Suppose we have a model with ten markets to be run for a period of 50 years Aside
from normalization a perfect foresight model implies computing prices In 500 dimensions while a
-t~mporary equilibrium model requires sorving SO equilibria each in ten dimensions Given that
computational speed often varies with the cube of the number of dimensions the lemporary
equilibrium formulation is potentially strikingly more feasible However Ballard (1987) and
Goulder-Ballard (1985) have devoloped techniques to greatly speed the computation of a perfect
foresight equilibrium
26
j
I
4 Jrnplemenlalion and Issues on policy evaluallon
41 Calibration and equilibrium comparlslon
CGE models at typically parameterized by the use of a calibration procedure Some parameters
are exogenously given However some crucial parameters are determined in such a way that the
model replicates the data for a given base year See MansurmiddotWhalleY 11984) for an eXlensive
discussion of 1~i$ issue
CalibratIon in a dynamic context is generaliy interpreted as requiring two properties First
replication of base year data is required Second the model is parameterized to simulate an
intertemporal balanced growth path when Ihe base policy is maintaineq This Is the approach
followed by Ballard (1983) BallardmiddotGoulder (1985) Goulder (1985) and SummersmiddotGoulde
(1986) It follows the practice of BallardmiddotFuliertonmiddotShovenmiddotWhalley (1985) with their
sequential cquilibrium dynamics
Other authors Auerbach-Kotlikof (1983 1984 1987) Boyenberg (1984 1985 1986)
and Pereira (198Gb 1987d) follow whal we cali a qualitalive calibration The slructural
palametars are exogenously chosen so that lh economy follows a reasonable path into Ihe fulure
ThaI has 10 do wllh the fact thaI given the recursive nature 01 the dynamic economy and aside tom
such structural parameters only iniUa stock values are needed to run these models Given initial
conditions on the stocks of say private wealth capltal and government debt agents will optimize
and Ihereby generate a lirst round of net demands and equilibrium conditions In lurn the
equilibrium prices will determine the evolution of the stock variables into the next perIod
There are several potential problems with calibration in the conlext of dynamic models The
assumption of a steadymiddotstate growth ~ath in the base case can be questioned First while
27 I Ibullbull
steadymiddotstaU is a possibity it certainly is not the only meaningfull solution to dynamic models
Even in the case of a perfect foresight equilibrium the model implies an equilibrium path which i may or may not involve colanced growth The model not the modeller should dictate the nature 01 I themiddotbase case path Second in the context of a temporarY equilibrium path a steady state solutlcn is Inol a likely model QutcOn1tl In fact unlessmiddot expectations are stattc short run behavior consistent
with a steady-state evolution will in general not be generated On the other hand if static I expectations are self~fufjili[1g we have in fact a perfect foresight model Third even the base year
replication requirement may cause problems In Ihe cOntexl of temporary equilibrium Any
calibration parameter would be conditional on expectation rules which is probably an undesirable
feature
The nalur of the two-requirement calibration strategy is very muchin the spirit of the bull
traditional design of the comparision of alternative equilibria comparis_ion between a steady-state
base case on one hand and alternative paths including a transition period and a finat sleadyyenstate on
the other hand Pereira (198Gb 1987d) compare different (not necessarily steadystate)
equilibrium paths The arguments against such a procedure are based on the idea that the impact of
policy changes can be observed most easily since all departures from the steadymiddotstate can be
attributed to the altemalive polioy On the other hand the base case so dEfined as a steadymiddotstate is
consist~nl with previous work in a less dynamtc setting and Iherefore allows a common standard
for comparing model results
42 Computation algorithms
We are still al a stage in which basically each author uses a different computational technique
Th development of dynamic models - in parlicular wilh adjustment costs andior perfect foresight
bull
28
has corresponded with the decline in the use of fixed point algorithms In -fact given the relative
arge dimensions inevitably involved such algorithms lend to be very inefficient at the best and
chen prohibitively stow See Stone (1985) and Preckel (1965) for a comparative assessment of
different computation techniques Among the models surveyed only Ball1rdFulle(ton~
-ShavenmiddotWhalley (1985) and Feltensteln 1985 use Merrills variant of the fixed point
algorithm 1echnique
AuerbachmiddotKoHikoff (1963 1984 1967) follow a three stage procedure They first compute a
base case steady-state then a revised case steady state and finally transition path for the economy
berveen these two steady-states tn all stages a Gauss-Seidel iterative procedure is used
Ballard (1982) Ballard-Goulder (1965) Goulder (HiSS) and Summers-Goulder (1986)
~
use a method developed by Ballard and Goutder which is similar in many awects to the FairmiddotTaylQr bull
(1983) algorl1hm Short-run equilibria are calculated (using Merrills algorithm) parametric on
nrice expeC~lions The model is then iterated to generate self-fulfilling intertemporaf expectations
and the corresponding perfect foresight equilibrium In a relatively similar approach Andersson
1987 uses a simulation program SIMNON developed In the University of Lund This program
can handle two-paint boundary problems In a fashion consistent with the multiple shooting
algcrlthm (see Lipton-Poterba-Sachs-Eummers (1982))
Boyenbergs computational approach (1985 1985) differs from the other models surveyed in
that he relies heavily on analytical techniques Computations are done by using a dynamic version of
Johansons linearization method Being essentially determined by the continuous time nature of the
I I
f
model thiS linearization model has the disadvantage of confining the analysls to infinltesinal changes
around the base case equilibrium (see Bovenberg (1985) p 53) Pereira (1986b 1987d) uses an optimization algorithm NPSOL - developped by
rGillMurray-Saunders-Wrlght (1986)) This algorithm is used 10 compute the sequence of
I
29
Shortmiddotrun temporary equilibrium which makes up the intertemporal equilibrium path The
equilibrium conditions are seen as nonlinear equality constraints in the minimization of an artlcial
objective function The prices are normalized to the unit simplex by an additional linear equality I I 1 bull constraint
II ErlichmiddotGinsburgh-Heyden (1987) follow a unique approach in that they use a variant of the
optimization technique introduced by Negishl (1960) The economic equilibrium can be generated i
as a solution of a mathemallcal program Ihe objective function of which is a weighted sum of the
utility functions of the various agents while tho constraints set consists of the market dearing 1
conditions Ginsburgh-Heyden (1985) have extended Negishis resulllO tho case of downward price
lrigidities I
Finally the paper by Jorgonson-Yun (1984) is also unique in tIlat it is the only I
econometrically estimated model Different blocks for the consumption and production Side of the
-
model are separately estimated to provide the necessary structural parameters i
I The diversity of computation techniques is yet another indicator of the exploratory nature of the
body of literature surveyed in this article
43 Equal yield comparisions
The link between a base case and counteriactual simulations Is usually provided by the concept of
equal yield Shaven-Whaley (1977) discuss the meaning 01 equal yield in a general equilibrium
context when government is not allowed to run defICits Equal yield is interpreted to moan constant
public utility Government base case utility is maintained in the counferfactual experiments With
balanced budgets the concepl of equal yield is unambiguous The new equUibrium prices and the
balanced budge condition will del ermine the minimal expendilure and taxes needed to maintain base
30
case publi~ utility Accordingly in general equal yield is inconsistent with equal nominal tax
revenue Some change in tax revenue Is necessary Different tax replacement schemes ate
considered to assure that enough tax revenue is collected This is the approach essentially followed
by most 01 the papers surveyed Only Feltenstein (1985) Goulder (198) and Jorgenson-Yun
(1984) chose not to follow an equal yield strategy
The question is of how to Interprete lila concept of equal yield when the government is alowed to
fUn deflcLts The optimal leve of expenditure for base ease public trliUty can now be taAinanced
bond financed or financed by a mix of bonds and taxation We have several versions of equal yield In
particular equal yield may now be consistent with equal tax revenUE Furthermore some meaSure
ltgtf financial crowding out effects of government deficns can be inferred from the comparislon of the
several equal yield alternatives These issues are extensively discussed ~Pereira (1987b)
44 Price expectations and the dynamic generalization of compensation
Indicators
The sum of equivalent and compensation variations over households is the most widely used
aggregated measure of efficiency gains or losses In the context of $teady~state eomparisions andor
when complete future markets exist andlor perfect foresighl is assumed there are no difficulties
associated with the use of the standard Hicksian indicators In fact correct future prices are known
in these cases
If expectations are not selfmiddot fulfilling andlor future markels are not available a dynamic
generalization is necessary 8allardmiddotGoulder (1985) provide some steps in that direction by
defining an indicator that accomodates periods far into the future when households do not have
perfect foresight but a steady~state prevails On he other hand this issue is more fundamental in
i
[
I
i I I I I [
I I
bullI
31
the contex~ of a Gner1 lemporary equilibrium ftamawor~ In such Circumstances Pereira
(1986a) dev~lops a cnamlc generalization which is obtained as the present discounted value of a i i
sequence of short-fUn optimal expenditure functions consistent with a base case expected future I stream of IlWities I
32
5 Empirical evidenee from selected policy Issues
Dynamic tax models have generated several important results which escaped Sialic modellers
The following is a seleCpoundd set of issuos whIch stress the marginal benefits of dynam~ modelHng of
economic behavior in innovltlttive areas of analysis The discussion of a consumption tax emphasizes
the benefits of dynamic tusehold behavior and intergenerational aspects The study of the
flliminatlon or the re~intToduction of the investsnemt tax credits is made meaningful by the dynamic
modelling of production betlavor Modelling governmenl in a dynamic conlexts allows the modeller
to study the irrrpact of government deflcits Finally a dynamic framework lets us appreCiate the I
relevance of consumer expectations in terms of the evaluation of policy alternatives
i
i
51 Consumption Tax
t
By Its very nature a consumption tax can not be adequately investigated with a static model bull
Fullerion-Shoven-Whaicy (1983) use Ihe model of the US economy s described in
Ballard-Fullerton-Shoven-Whalley (1935) to evaluate the movement from the currenl US tax Isys1em to a progressive consumption tax Since their modeJ Incorporates a Jaborleisure choice
where leisure is an untaxed commodlty their results reflect the fact that both the consumption tax
and Ihe present system are dslortionary
Concentrating on the rntertemporaf distortions they show that sheltering more saving from lhe
current US income tax could improve econornic efficiency even if marginal tax rate increases are
necessary in order to maintain government fevenue~ At first you have a revenue shortfall
However the economy moves to a higher sustatned growth path and uitimatefy lower tax rates
cangenerate the same revenue path Also wages increase in the long run with this policy The
bull
bullbull ~
33 shyt
i
I
I
present value of welfare gi[lS for a potiey of complete savin9 deduction wilh marginal tate
adjustments (consumption lax) is simulated to be around $500 billion to $600 billion 01 1973
dollars
They also investiga1e th 1ngth of ttme it takes th~ economy to adjust to these policy changes
Roughly they estimate the long run to b thirty years although this figure Is very sensitive to the
specification of the savings elasticity - a crucial parameler in this model
52 Investment T~x Credit
Goulder-$ummers (1 S87) address the impact of eliminating the lnvestmenl Tax Cradit (ITC) on
intersectoral capital formation and on economic growth Their model isYa1ticularly adequate to
addies$ this issu~ in that is postulates a forward looking investment m9we with adjustment costs
They show that the eliminating the ITC causes a reduction in the rale of investment Inveslment is
estimated to fall by about 7 in the short-run and by about 12 in Ihe new long run steady state
In lurn a previous policy anncuncement lowers Ihe overall al1ractiveness of Investment and leads 10
a downward shift of the Investrrent profile On the other hand the combined effect of a revenue
neutral simultaneous efimiraron of the ITC and reduction of the corporate tax rates is a long run
reduction ef the capital Sieck by 35 This pattern suggests that a revenue neutral increament in
both Ihe corporale tax rates and the inveSlment lax credits would be preferable 10 the formula
adopted in Ihe US Tax Reform Act of 1986
Pereira (1987d) addresses the impact of the ITe from Ihe oland point of Ihe curTenllax syslem
under the Tax Reform Act of 1986 Given Ihe concern over the potential depressiVe effects upon
savings and investment of the new tax system in conjunction with deficit reduC1ion mechanisms of
the GrammmiddotRudman type a pertinent question is should Ih lTC be e-introduced In the context
bull
34
-of his mod~1 of the US economy Pereira focuses on the trade-off between the distortion in the
intersectoral allocation of capital induced by the lTC the lower tax revenues it generates and
potential financial crowding-out effects on one hand and the positive effects of lowering the relative
price of new capital goods on the other hand Preliminary results confirm the qualitative results in
Goulder-Summers (1987) suggestting a welfare gain of about 2 of the present value of GNP in
the best scenario of absence of replacement taxes Furthermore preliminary results show an
important time pattern to welfare gains with the average benefits increasing the further you look
into the future This pattern is due to the presence of constraints to intersectoral mobility of capital
and inertia in the adjustment towards the optimal capital levels as reflected by the presence of
adjustment costs
I
I
bullAuerbach-Katlikaff (1gB) in the context of their intergenerational model of the US economy
consider the impact on savings capital formation and interest rates of deficit policies and balanced
budget increases in government consumption They conclude that deficit finance and government
consumption can significa-ntly crowd out capital formation and lower the welfare of future
generations However crowding out from deficit finance is a very slow process because it results
from increased government spending over potentially long horizons Also deficit policies may
substantially influence the longmiddotrun interest rates while leaving the short-run rates essentially
unaffected Finally and in opposition to the central role adjustment costs seem to play in both
53 Financial crowding outeff~cts of government deficits
Goulder-Summers and Pereiras models Auerbach-Kotlikoff suggest that the time path of interest
rates induced by a policy of deficit financing seems to be insensitive to the presence of adjustment
costs
bull
35
54 The role 01 consumer expectatlons
The expectations analysis has uncovered one of the benefits of dynamic modelling
Ballard-Goulder (1985) find that the appeal of adopting a consumption tax depends on the level of
foresight possessed by the consumers Furthermore additional foresight may be welfare worsening
This is a second best type of result This tends 10 occur under policies that lead to capital deepening
and declining rate of return to capital over time To the extent that consumers have more foresight
they will be beller equippod to anticipate the fall in the rental price of capital Consequently if a
saving incentive is enacted people save more with myopia than with perfect foresight Given the
exi~tence of taxes on capita and the discrepancy between the private an~--sccial returns to capital
the greater savings with myopia is beller socially Ballard-Goulder find that the welfare gains
from a consumption tax are reduced by about 10 when we move from myopia to farleet foresight
Therefore the attractivenness of adopting a consumption tax seems fairly [obust across these
foresight specifications
I
bull
36
6 Tbe power and weaknesses 01 dynamic modelling the example 01 corporale
tax Integrallon
The Issue of corporate tax integration allows us to sh~w the stre-ngths and weaknesses or the
dynamic approach Empirical evidence using CGE techniques indicates that depending on the precise
scheme of integration and the 1ax replacement methods integratiOn may have substantial effects In
Ihe work of FuliertonmiddotKingmiddotShoven-Whalley (1980 1981) total integration was found to yield an bull
annual static eHielency gain of $4 to $8 billion in 1973 dollars Simulated dynamic gains may be as
large as $695 billion or abcu 14 of the present value of tulUre consumption and leisure in the
US economy These gains result primarily from interindustry reallocalions of Investment and an
improved inter~emporal allocation of consumption
Mora recent work emphasizes hOw consumers asset portfolio decisions and firms finanCial
decisions affect the efficiency gains from integration Slemrod (1980) focusing on consumers
asset ponfoio deCisions finds static efficiency gains which are about twice -as Iarge as those
repOrted iiy Fulierton-King-ShovenWhalley (1961) FuUrlon-Gurdon (1983) focus on the
firms financial decisions They roporl efficiency gains of 6 of GNP from the elimination of the
tax distor1ions favoring debt However when they eliminate the corporate tax and repiace it with
increaSed persona income taxes addlHonal dls10t1ions are created in the optimal labot~leisure
decisions These distortions tend to dominate the analysis slJ(h that the overall effects of complete
integration are very modest Galpr-LuckmiddotTodr (1986) address simultaneously consumers
asset pOrlfollo decisions and the firms financial decisions They also find very modest efficiency
gains from integralion
The above results are important but may be severely biased There are severa aspects of
economic behavior and modelling crucial for the study of Income tax Integration whiCh have no been
i 37
captured innny of the above papers One aspect is the absence of government deficitsmiddot a balanced
budget is assumed It is true Ihal resource crowding OUI is caplured In these mOdels but financial
crowding out induced by government spending is nOI Second investment is not derived from
optimtztlcn behavior Investment behavior passively accomodates endogenous saving decisions As
a consequence the differential impacts of policies in the incentives to save and to invest are not
captured Aso full capital mobility across secors is assumed with instantaneous capital
adjustmenlS towards optimal levels This assumption rules qut different costs of capital across
seclors and therefore differentiated reactions to tax policies changes Finallyhe modelling of both
gov9rnmerlt deficits and of endogenous real and financial investment decisions necessitates the
~nsideraiion of a dynamilt framework and the introduction of financiaJ assets govemment bonds and
iprivate financial assets A dynamic framework also highlights the efficiency effects of Integration on r
ithe ptimal intertemporal decisions The above models are either sIalic (Slemrods and
GalpermiddotltJckemiddotTode(sj or a dynamic sequ~nce of otherwise Slatic model bullbull I Pereir a (1986b) develops a dynamic applied general equilibrium model bull to study the tbull
efficiency effects of integration on the growth and allocation of investmenl across sectors Special
Iattention is paid to the structural affeelS of resource and financial crowding oul induced by
gove mect spending and deficits and to the real and financial Investment reactions across industries
to both tax integration and flnancial crowding out His mode departs from previous work on income
lax Integraticn and for that matler from most 01 the CGIE literalure for tax policy evaluation in
several direcllons It encompasses an endogenous sequential equilibrium struelure founded on
dynamic behavior with flexible expectations Government deficits are optlmaUy determined
Investment decisions are tward looking and the resull of optimizing behavior Several financial
assetsmiddot public and privatemiddot are considered
Simulation resulls Indicale that the welfare gains from Integration under large deficits are at
32
best lJlodest when measured in terms of the GNP The elimination of the corporate tax and its I replacemenl by increased income lax rates yields long run benefits which are never larger than
2 of the presenl yalue ct telr consumption and leisure onder the previous tax regime and 1
under the current tax regime However the short run w~lare effects of full integration tend 10 be
very low and in in some scenarios even negative This is a flew intertemporar pattern of efficiency Ieffects which reflects an adjJsiment lag in the interindustry investment decisions due to the
existence of costs of adjustment On the of her hand partial integration achieved by excluding I dividends from the corporale lax base syslematically yields negalive efficlencyeffeclS In Ihe long Inm these can be as high as 3 of present value of the future value of consumption and leisure This
is a new second best effect suggesllng Ihat less than complete integration may have perverse I efficiency effects
The dynamic slructure of PerclrBs model snowing for an improved freatment of real investment
declsions and government deficts provides a potential setting for a more accurate measure of the
costs and benefits of inlegration The simulation resullS suggest lower benefits and an intertempora Ibull
pattern of growing benefits
Simulation results also clearly suggest robust negative effects from partial integration How
reliable is this result If dJvidends were deductible from the corporate lax base (as are interest
payments) the preferenlial Ireatmant of debl over equily would be ellminaled Corporallons should
be expecled 10 react by decreasing the optimal deWeqully rallo However corporale financial
gecisions are exogenously set In Pereiras model as in aft of the other models surveyed that go as far
as dealing with the issue
Also withdrawing dividends from Ihe corporate tax base is a way of eliminating Ihe double
taxation of dividends at belh Ihe corporate and personnal Income levels How will the optimal
aliocellon of household saving accomodate that change in Ihe relalive return to corporale equity
39 I
t However hOusehold portfolio decisions are exogenouly set in Pereiras model as in all of the other
models surveyeltj go as far as dealing with the issue
IIt is sate to say that the evaluation 01 the benefits 01 partial Integration as defined above are
sevrey underestima1eO Meanwhile the distortionSoenerated by the increase in the marginal
personal tax rates designed to ralse the revenue foregone by the dividend exclusion are fully
aCCltlunted for A good measure of the costs of integration together With a poor evaluation of the
benefils may very well be the reason why partial integration is Simulated 10 yield neg alive
eHikiency gains
On a different vein as most of Ihe model surveyed here Pereiras is a closed economy model It
is legitimate to wonder how the resullS would change If international capital flows were anowed
IThis almost certainly would affect financial crowding out since a good d~al of the capital inflows
could be used to finance public debt
For these various reasons the results should be treated as preliminary but they strongly I suggosl that the dynamic structure provides valuable new insights into the corporate tax integration SSU2
40
7 Concluding remarks I I I i
What has been acccmplished Ihis far The success of Ihe CGE research in laelding the challenge
of dynamic modelling can not be denied
I iGreat progress in the modelling of household behavior has been achieved Promising
developm~nts in the modeUing of production and government behavior have also been accomplished
Interesting innovations in the concept of equilibrium and computation techniques were discussed
The policy analysis was greatly enriched by considering transitional effects together with longer
run steadymiddotstate equilibrium paths generalized equal yield strategies accomodating different
financial crowding out impacts and generalized dynamic policy evaluation indicators
An important set of issues have been addresed by the models surveyed Traditional issues
focusing on the effects of capital taxation and consumption taxes have been pursued In terms of
capital taxation for example the intertemporal nature of the issue was enhanced by allowing an
optimal evolution of capital stock in the economy and forward looking optimal investment decisions
In Andersson (1987) GouldermiddotSummers (1987) and Pereira (1986b 1987d) this is coupled
with a improved treatment of several tax provisions like investment tax credits and depreciation
allowances In terms of the consumption taxes developments in the specification of Intertemporal
household behavior the introduction of overlapping generations and the modelling of
intergenerational links via bequest motives as well as a closer attention to demographic evolution
provided a much improved economic setting for the understanding of the several aspects of the
problem
Furthermore dynamic modelling has permitted the addressing of a set of new issues Policy
Issues ranging from dynamic tax policy analysis to the evaluation of exogenous changes in
government expenditures to the impact of different methods of financing government expenditure to
4 1
the importance of tinanciai croHrii1g out to 1he relevance of consumorS expectations in tne policy
results have been addressed
Despile all of the progress and in part due to such progress several avenues are wide open for
much neded additional research The following seem to be the most Interesting and promising
areasmiddot First the specincation of liquidity constraints to household behavior may help to obtain a
better understanding of policy chenges that affect directly the interest rates in the economy Second
major attention needs to be paid to the incorporafion of a roraign sector (with open capital markets)
Third some efforts are needed in terms of finding adequate computation techniques now that
dynamic modelling eally makes the curse of dimensionality worse than eVOrbull
The modelling of financial markets is the single most unsatisfactory speet of the dynamic
models surveyed here~ The problems associated with provlding a mo~ adequate treatment of
financial markets and in general endogenous and optimal financial deci$ions bull both at the corporate
and at the personal levelsmiddot have 10 be addressed That necessarily requires Introducing uncertainty
into the models To b adequate this must go well beyond the mere assumption of point expeetalions
with the whole array of modelling implementalion and evaluation problems thai generates
bull I i
42
REFERENCES
Abel A and O Blanchard 1983 An Intertemporal Model of Saving and Interest Econometrica
51 pp 675-92
~ Andersson K 1987 Tar-ation of Capital in Sweden - A General Equilibrium Model Ph D
Dissenation UnIversity of Lund Sweder
Auerbach A L KotlikoH 1983 National Savings Economic Welfare and the Structure of
Taxation in M Feldstein ed Behavioral Simulation Methods in Tax Policy- Analysis Chicago
University of Chicago Press
Auerbach A l Kotlikoff 1984 Social Security and the Economics of Demographic-
Transition in H Aaron and G Burtless eds Retirement and Economic Behavior Washington DC
Brookings Institution
Auerbach A L Kotlikoff 1987 Dynamic Fiscal Policy Cambridge Cambridge University
Press
Auerbach A L Kotlikof and J Skinner 1983 The Efficiency Gains from Dynamic Tax
Reform Internatiooaj Economic Reyiew 24 pp 81middot100
Ballard C 1983 Evaluation of the Consumption Tax with Dynamic General Equilibrium
Models Ph D Dissertation Stanford University
Ballard C 1987 Tax Policy and Consumer Foresight A General Equilibrium Simulation
Study Economic Inquiry 25 pp 267-284
Ballard C and L Goulder 1985 Consumption Taxes ForeSight and Welfare a Computatonal
General Equilibrium Analysis in J Piggott and J Whalley eds New Deyelooments in Apolied
General Equilibrium Analysis Cambridge Cambridge University Press
Ballard C D Fullerton J Shoven and J Whalley 1985 A General Equilibrium Model for
t 43
Tax Eoliey Eyalua~iQO NBER University or Chicago Press
Borges A (1987) bull A Survey of Compula1ional General Equilibnum Models CECD Economic
Slud~ 8
Bovenbig L 19B3 imperfect Capital Mobility in a Perfect Foresight Herberger Model
Mimea UC Br~eley
Bovenberg L 1984 Capitel Accumulation and Capital Immobility Qmiddottheory In a Dynamic General Equilibrium Framework PhD dissertation University of California at Berkeley t
Bovenberg L 19850 Dynamic General Equilibrium Tax Models with Adjustment Costs in A iManne ed bull 1985 EtQnomrc ECUjilbrium Model Formulation and Solution Mathematical
lProgramming Siudy 23 Arnsterdam NorthmiddotHoliand
Bovenberg L t 1986 Capital Income Taxation in Growing Open Eeondmies Journa of Pubtic
Eamp(omlcs 3t pp 347middot376 IBovenberg L ald W Keller 1981 Dynamics and Nonlinearlties in Applied General
Equilibrium Models Internal Report Mlmiddot81middot208 Department for Statistical Methods CBS bull
Vooiburg The Netherlands
Charnley C 1981 The Welfare Cost of C~pltal income Taxation In a Growing Economy
IJournal of Political ECQnQm~ 89 3 pp 468middot496 I
Charnley C 19810 Efficient Tax Reform in a Dynamic Model of General Equilibrium I Mimeo Yale University I Charnley C 1982 On the Optimal Taxation In Stylized Models 01 Inlartemporal General
Equilibrium Mimeo Yale University I Decaluw B and A Manens 1985 Pays en Oeveloppement at Modeles Calculables OEquilibre
General - Une Revue de al litterature empirique Monographie Sene 9 Volume 3 Centre de
Recherche et Developpemont Economlque Unrslt de Montreal
44
bull
I I
I
I
1
Erlich $ V Ginsburgh and L van der Heyden 1987 Wher~ do Real Wage Policies Lead
Belgium bull A General Equilibrium Analysis European Economic Reyiew (forthcoming)
air R and J Taylor 1983 Solution and Maximum Likelihood Estimation of Dynamic
Nonlinear Ratioant Expeclations Models Ecooometrica 51 pp 1169middot1186
Feltenstein A 1984 Money and Bonds In a Disaggregated Ec~nonYin Searl 1-1 J Shoven
Eds Aoolied ~eoeral EqJiliDCiulD tioalYsis Cambridge University Press Cambridge
Feltenstein A 19B6 A Dynamic General Equllibrlu Analysis of Financial Crowding Out
Theory with an Application to Australia Journal of Public Economics 31 No1
Fullerlon D R Gorden 1983 A Reexamination of Tax Distorllons in General Equilibrium
Models in M Feldstein Ed Behayora Simulation Methods in Tax pOlicy Analysis Chicago
University of Chipgo Press
Fvllerlon D J Shaven and J Whalley 1983 Replacing the US Income Tax with a
Progressive Consumption Tax JoulOal of Public ECQnomics 20 pp3middot23
Fullerton D Y Henderson J Shaven 1984 A COmparision of Methodologies in Empirical
General Equilibrium Models of Taxation in Searl H J Shoven Eds ~pQlied Geoaral Jguilibrium
lIoalyss Cambridge University Press Cambridge
Fullerton 0 A King J Shaven and J Whaney 1981 Corporate Tax Integration in the United
Statesmiddot a General Equiiibdufi1 Approach American Economic Review 71 pp~ 677~691
Galper H R Lucke E Tader 1985 Taxation portlor Choice and the Allocation of Capital A
General Equilibrium Approach Mimeo
Gill p W Murray M Saunders M Wright 1986 Users Guide for NPSOL (Version 40) A
Fortran Package for Nonlinear Programming T echnicsl Report SOL 86middot2 Deparlment of
Operations Research Stanford University
Glnsbur~h V and L van der Heyden 1985 General Equilibrium with Wage Rigidities An
45
AppJication to Belgium in A Manne ed t Economic EQumprit1m~ Made Formulation and $oJioo
Mathematical Programming Study 23 Amsterdam NorthmiddotHolland ~
Goulet J 1968 1 Adjustrent CO$1S in the Theory o(the Firm Reyiew of EconomiC Studies 35 bull
pp47middot55
Goulder L 1985 Tax-Financed versus BondmiddotFinancedGhanges in Governent Spending
bull Mimeo Harva(d UniversilYmiddot
Goulder l J Shaven and J Whalley 1983 Domestic Tax Policy and the Foreign Sector The
importance of altemative foreign policy formulations to results from a general equilibrium tax
analysis model In ~n methods in Ibe antico of income from caoia1 ed Martin
Feldstein Chicago University of Chicago riSS
Goulder l L Summers 1987 Tax Policy Asset Prices and 3rowtly ~
A General Equilibrium Analysis NBER Working Pper No 212B
Grandmont J 1982 Temporary General Equilibrium Theory Ch 20 in Handbook 01
Mathemalical Eccnornics Vol II NorthmiddotHoliand
Harberger A 1959 The Corporate Income Tax An empirical Appraisal In lax Revision
Comoendjum Voll House Cornmitee on Ways and Means Washington DC Govemment Printing
Office
Harberger A 1962 The Incidence of the Corporate Income Tax JOULn1 of Political
Economy 70 pp 215middot240
Harberger A 1964 Taxation Resource Aliocalrlon and Welfare fn lb Role of Dir bullbulll and
Indjresl Taxes io tbe Fede[al Reserve s~I~Ill Princeton Princeton Unlveresity Press
James J 1985 New Developments in the Applications of General Equilibrium models to
Economic History in Piggott aand Whalley Eds
Jorgenson D and K Yun 1984 Tax Policy and Capital Allocation Harvard Inslitute of
bull
I
I f
Economic Research WP 1107
Judd K 1965 ShonmiddotRun Analysis of Fiscal Policy in a Simple Perfect Foresight Model
Journal of poJiHcal 5CCJiW1l pp 298middot319 r
Upton D J Potbc J Sachsa nd L Summers 1982 Multiple Shooting in Rational
Expectations Models fuoometrlc 50 1329middot1333
Manne A ed 1985 fconQrnjc Eoullibrium Model FOrIDujaljQo and Solution Mathematical
Pogrammlng Study 23 Amsterdam NorthmiddotHolland
Manne A 1985 Or the Formulation and Solution 01 Economic Equilibrium Models In A
Manne ed ECQOQlli-Jr~ibmiddotum ~tQdel EQlJulailon and Solution Malhematical Programming
Study 23 Amsterdam NorthmiddotHolland
Mansur A and J Whalley 1984 Numerical Specification of Applied General Equilibrium
Models in Scarf H J Shoven Eds Aoplied General Eouillbrium AnaiysectIsect Cambridge Universily
Press Cambridge
Merrill 0 1972 Applications and Extensions of an Algorithm thaI Computes FixedmiddotPoints of
Cenain Upper Semimiddotcominuou Point to Set Mappings PhD Disserlation Unlvel$ity 01 Michigan
Negishi T 1960 Welfar Economics and Exislence I an Equnlbrlum for a Competitive
Economy MelroeCOOQ11ka 12 92middot97
Pereira A 1985 On the Existence and Uniqueness of Optimal Ou1put Path lor CRTS Firms in
Adjustment Costs Technologies Mimeo Stanford University
Pereira A 1985a Consumer Expectations and Ihe COSI 01 UllUty In an Intertemporat
Framework Mlmeo Stanford University
Pereira A 1985b Corporale Tax Integrallon and Government Deficits A Dynamic General
Equilibrium Anaiysls Mlmeo Stanford University
Pereir A 1987 On Ihe Computation of the Users Costs of Stocks Ecooomi (forthcoming)
47
Pereira A bull 1987b Equal Yield Alternatives and Government Oeficlts Mimeof Stanford
University
Pereira A 1987c A Dynamic Applied General ~uHlbrium Model for Tax Policy Evaluation
Comlletamp Docurnent2tion Mimeo University of California San Diego (in progress)
Pereira A 1987d Should the investment Tax Credit be Rei(ltroduced Mlmeo University
of California San Diego (ir progreSs)
Piggott J and J Whalley eds 1985 tJew QeyeQpments in Applied General EQuilibrium
bullenalysjs Cambridge Cambridge University Press
Preckel Pbull 1985 Alternative algorithms for computing economic equilibria In A Manne Ed
Radnor R 19amp3 Equllibrium under Uncertainty In K Arrow M Intrilligaor ods
I ibull
Robinson S 1 g86 Mullisectorai Models of Oeveloping CountrieS a Survey Oivision of
Agricullure and Resources UC Borkeley WP 401
Scarf H 1967 The approximation of fixed points of a continuous mapping ~MM Jouwal of
epplid Mathornallcs 15 pp 1328middot43
Scarf H with T Hansen 1973 00 the Computation of EconomiC Eguilibria New Haven Yale
University Press
Scarf H J Shoven Eds 1984 61l1llied Genera Elujlibrium 6nalyssect Cambridge University
~ Press Cambridge
Shoven J 1976 The Incidence and Efficiency Effects of Taxe on Income from Capital
Jurnal of ramie1 Economy 84 pp 1261-83
Shoven J 1983 Applied General Equilibrium Tax Modemng IMF Slaff PaoerS 30
Shoven J and LB Simon 1986 Share Repurchases and Acquisitions An Analysis of which
Firms Participate NBER Working Paper No 2243
bull
48
IShaven J and J Whalley 1972 ft General Equmbrium Calculation of lhe Effects of
Differential Taxation of Incoma from Capital in the US Jauroe of public Econgmics 1
pp281middot321 (
Shoven J bull and J Whalley 1973 General Equilibrium with Tax$s a computation procedure
and an existence pr00f ReviEhi of Economic Studies 40 pp 475middot490
Shoven J J Whalley 1977 Equal Yield Tax Alternatives A General Equilibrium
Computa~ui0nal Tfchnique J~HJrnal of PUblic EconomIcs 8 pp 211~224 bull
Shaven J J Whalley 1984 Applied GeneralmiddotEquilibrlum MOdels of Taxalion and
International Tr8de An Introduction and Survey Journal of Economic literature 23t pp
1007middot1051
Slemrod J 1980 A General Equilibrium Model of Capital Income Taxation Ph D bullDissertation Harvard Univnrsity
Slemrod J 1983 A General Equilibrium Model of Taxation with Endogenous Financial
Behavior in M Feldstein edbull Behavioral Simulation Methods In Tax Policy Analysis Chicago
UniVerSity of Chicago Press
Stone J 1985 Sequential Optimization and Complementarily Techniques for computing
Economic equilibrium in A Manne Ed
Summers L 1981 Taxation and Corporate Investment a q-Theory Approach BroQkinas
Summers L 1985 Taxation and the Size and ComposHion 01 the Capital Stock An Asset Price
Approach Mlmeo Harvard University
Whalley J 1975 A Geneal Equiloibrium Assessment of the 1973 United Kingdom tax
reform Economica 42 pp 139middot61
1JLTIllOS JIQll~)ltG PAPERS PUBLICADOS bull
I
I I
j
112 70 - LUCENA iJiCIIJo ~ IIEnvironrilent Honitoring and Orqanization Structure T (Setellibro 1987)
nI 71 - L~ENA Diogo A Note on the Representation of Information Stxueture (SeteJObro 1987)
nQ 72 - LUCENA D1CgO Envuorment 111)nitoring and Organization
Structure II (Setelilhro 1987)
ril 73 - AlmmE3 Antonio Pais e CAsPAR Yitor bull Tributacao Inoentivos e InvesUlllentos Analise Qualitativa noYellibro 19S7)
n2 74 - liAlltOSA rose Pedro Opt1lllal lIage R1gidity A SUggestedtlthodology to Teet tb Theory and an Application (Outubro 1987)
nI 75 - CABRAL L 11 B Three lIotes on SYJIIlletric Gemes witt itsYJIIlletric Equilibri (NoveJObro 1987)
n2 76 - LUCENlt D10gOTo Search or Not to Search (Fevereiro1967)
nQ 17 - SA Jorge asconcelos euroI A Theory of SynergyB (Fevereiro1987)
n2 76 - VILARES tanuei Jose Os Be IntenaMios IlIIportados ColIIO Factor de Produao (Julho 1987)
nQ 79 - SA J~rge VasCQncel03 e HoY to Copete and Co~unicate in tature Industrial Products_ (aio 1988)
n2 80 - ROIl Rafael Learning and Capacity Expansion in a New larket Under Uncertainty (Fevereiro 1988)
n2 81 PEREIRA Alfredo llarvilo Survey of Dynamic Computational General EquilibriUlll Hodel for Tax Policy Evaluation (Outubro 1987)
Qualquer in ormacao sabre o 1IIorlltlng Papers ja publlcados sera prestada pelo Secretariado de Apoio aos Docentes podendo os ~e3OSf
ser adquirido5 ne Secyo de Vlndas da Faculdade de EconollLia UNL na Travessa Esteviio Pinto Cltpolide - 1000 LISllOA
24 l
bull
rerurn pn ~uity equas the return on bonds plus an exogenous risk premium
36 Market assumptions equilibrium and expectations
Virtually all of the papers arc characterized by Walrasian market clearing assumptions All
markets are perlect1y competi1ive Atomistic competition among agents 1s assumed even thoughmiddot only
a flnile number of agents is considered Virtually no mark~t disequilibria or price stickiness afe
considered The only exception is ErlichmiddotGinsburghmiddotHeyden (1987) bull In thei model of the Belgium
economy I the _wage rate is fixed in the short-run Therefore in the shOrlwrun desequilibrium in
the labor market will generate endogenous unemployment However in lh$ long-run all prices
Including the wage rata are Ilexible and accordingly ali markets ciear
Given the dynamic nature of behavlor in the economy markelmiddotcfearing prices in each period
depend on expectations of fture prices and on tax variables In the economy There are essentially
two ways at interpreting the economic equilibrium in such a dynamic conte~bull If future prices are
perfectly anticipated (io expectations are self-fulfilling) a perfect foresight equilibrium
prevaHs Then future actions are merely the implementation of current decisions for future
periods However jf ptica expecla110ns are not perfect (Le agents make mistakes w1th re$pect 10
future prices) then a temporary or shortrun equilibrium prevails Markets ciear and clearing
prices depend on fu1ure price expectations Current plans about the future are typically not
precisely implemented They will be revised as more or better information becomes available to the
economic agents See Grandmant (1982) for a comprehensive survey 01 the temporary equilibrium
literature
Wilh the exception of Gould (1985) and Pereira (19S6b 19S7d) all the models surveyed
adopt th~ concept of a perfect foresight equilibrium In turn BaliardmiddotGouldr (1965) considers a
J
2S
flexible am9unt of foresight In terms of the number of years over which price movements are ~
foreseen Pereiras model (1986b 1987b)) is flexible in a somewhat different way in that it can
indude any range of foresight from myopia to perfect foresight The choice between the perfct foresight and lemporary equilibrium is ultimately to be made on
philosophic grounds It can be argued that less than perfect expectations Imply that agents are
irralional in some way (see Auerbach-Kctiikoff (1987 p 10) However the reverse argument can
be made One can question wiether agents are reaHy rationar and perfectly knowledgable about
future prices
Recent evidene of Balird (1987) Ballard-Goulder (19851 Goulder (19851 and Pereira
(1986b 1987d) suggests that the choice in modemng expectations is an important one They show
that the degree of foresight inl0 thc future (ranging from perfect foresight to-myopic expectations)
may have dramatic impacts on the pollcy conclusions of Ihe model Accordingly the best research
strategy may b to design models which are flexible enough 10 allow for different rules regarding the
formation of expectations With the exception of the articles just mentioned sensitivity analysis
bullhave previously not been performed along this dimension
In terms of implementation the two concepts of equilibrium - perfect foresight and temporary
equilibrium have different implications The dimensionality of the equ1ibrium soll1ion algorIthm
is involved Suppose we have a model with ten markets to be run for a period of 50 years Aside
from normalization a perfect foresight model implies computing prices In 500 dimensions while a
-t~mporary equilibrium model requires sorving SO equilibria each in ten dimensions Given that
computational speed often varies with the cube of the number of dimensions the lemporary
equilibrium formulation is potentially strikingly more feasible However Ballard (1987) and
Goulder-Ballard (1985) have devoloped techniques to greatly speed the computation of a perfect
foresight equilibrium
26
j
I
4 Jrnplemenlalion and Issues on policy evaluallon
41 Calibration and equilibrium comparlslon
CGE models at typically parameterized by the use of a calibration procedure Some parameters
are exogenously given However some crucial parameters are determined in such a way that the
model replicates the data for a given base year See MansurmiddotWhalleY 11984) for an eXlensive
discussion of 1~i$ issue
CalibratIon in a dynamic context is generaliy interpreted as requiring two properties First
replication of base year data is required Second the model is parameterized to simulate an
intertemporal balanced growth path when Ihe base policy is maintaineq This Is the approach
followed by Ballard (1983) BallardmiddotGoulder (1985) Goulder (1985) and SummersmiddotGoulde
(1986) It follows the practice of BallardmiddotFuliertonmiddotShovenmiddotWhalley (1985) with their
sequential cquilibrium dynamics
Other authors Auerbach-Kotlikof (1983 1984 1987) Boyenberg (1984 1985 1986)
and Pereira (198Gb 1987d) follow whal we cali a qualitalive calibration The slructural
palametars are exogenously chosen so that lh economy follows a reasonable path into Ihe fulure
ThaI has 10 do wllh the fact thaI given the recursive nature 01 the dynamic economy and aside tom
such structural parameters only iniUa stock values are needed to run these models Given initial
conditions on the stocks of say private wealth capltal and government debt agents will optimize
and Ihereby generate a lirst round of net demands and equilibrium conditions In lurn the
equilibrium prices will determine the evolution of the stock variables into the next perIod
There are several potential problems with calibration in the conlext of dynamic models The
assumption of a steadymiddotstate growth ~ath in the base case can be questioned First while
27 I Ibullbull
steadymiddotstaU is a possibity it certainly is not the only meaningfull solution to dynamic models
Even in the case of a perfect foresight equilibrium the model implies an equilibrium path which i may or may not involve colanced growth The model not the modeller should dictate the nature 01 I themiddotbase case path Second in the context of a temporarY equilibrium path a steady state solutlcn is Inol a likely model QutcOn1tl In fact unlessmiddot expectations are stattc short run behavior consistent
with a steady-state evolution will in general not be generated On the other hand if static I expectations are self~fufjili[1g we have in fact a perfect foresight model Third even the base year
replication requirement may cause problems In Ihe cOntexl of temporary equilibrium Any
calibration parameter would be conditional on expectation rules which is probably an undesirable
feature
The nalur of the two-requirement calibration strategy is very muchin the spirit of the bull
traditional design of the comparision of alternative equilibria comparis_ion between a steady-state
base case on one hand and alternative paths including a transition period and a finat sleadyyenstate on
the other hand Pereira (198Gb 1987d) compare different (not necessarily steadystate)
equilibrium paths The arguments against such a procedure are based on the idea that the impact of
policy changes can be observed most easily since all departures from the steadymiddotstate can be
attributed to the altemalive polioy On the other hand the base case so dEfined as a steadymiddotstate is
consist~nl with previous work in a less dynamtc setting and Iherefore allows a common standard
for comparing model results
42 Computation algorithms
We are still al a stage in which basically each author uses a different computational technique
Th development of dynamic models - in parlicular wilh adjustment costs andior perfect foresight
bull
28
has corresponded with the decline in the use of fixed point algorithms In -fact given the relative
arge dimensions inevitably involved such algorithms lend to be very inefficient at the best and
chen prohibitively stow See Stone (1985) and Preckel (1965) for a comparative assessment of
different computation techniques Among the models surveyed only Ball1rdFulle(ton~
-ShavenmiddotWhalley (1985) and Feltensteln 1985 use Merrills variant of the fixed point
algorithm 1echnique
AuerbachmiddotKoHikoff (1963 1984 1967) follow a three stage procedure They first compute a
base case steady-state then a revised case steady state and finally transition path for the economy
berveen these two steady-states tn all stages a Gauss-Seidel iterative procedure is used
Ballard (1982) Ballard-Goulder (1965) Goulder (HiSS) and Summers-Goulder (1986)
~
use a method developed by Ballard and Goutder which is similar in many awects to the FairmiddotTaylQr bull
(1983) algorl1hm Short-run equilibria are calculated (using Merrills algorithm) parametric on
nrice expeC~lions The model is then iterated to generate self-fulfilling intertemporaf expectations
and the corresponding perfect foresight equilibrium In a relatively similar approach Andersson
1987 uses a simulation program SIMNON developed In the University of Lund This program
can handle two-paint boundary problems In a fashion consistent with the multiple shooting
algcrlthm (see Lipton-Poterba-Sachs-Eummers (1982))
Boyenbergs computational approach (1985 1985) differs from the other models surveyed in
that he relies heavily on analytical techniques Computations are done by using a dynamic version of
Johansons linearization method Being essentially determined by the continuous time nature of the
I I
f
model thiS linearization model has the disadvantage of confining the analysls to infinltesinal changes
around the base case equilibrium (see Bovenberg (1985) p 53) Pereira (1986b 1987d) uses an optimization algorithm NPSOL - developped by
rGillMurray-Saunders-Wrlght (1986)) This algorithm is used 10 compute the sequence of
I
29
Shortmiddotrun temporary equilibrium which makes up the intertemporal equilibrium path The
equilibrium conditions are seen as nonlinear equality constraints in the minimization of an artlcial
objective function The prices are normalized to the unit simplex by an additional linear equality I I 1 bull constraint
II ErlichmiddotGinsburgh-Heyden (1987) follow a unique approach in that they use a variant of the
optimization technique introduced by Negishl (1960) The economic equilibrium can be generated i
as a solution of a mathemallcal program Ihe objective function of which is a weighted sum of the
utility functions of the various agents while tho constraints set consists of the market dearing 1
conditions Ginsburgh-Heyden (1985) have extended Negishis resulllO tho case of downward price
lrigidities I
Finally the paper by Jorgonson-Yun (1984) is also unique in tIlat it is the only I
econometrically estimated model Different blocks for the consumption and production Side of the
-
model are separately estimated to provide the necessary structural parameters i
I The diversity of computation techniques is yet another indicator of the exploratory nature of the
body of literature surveyed in this article
43 Equal yield comparisions
The link between a base case and counteriactual simulations Is usually provided by the concept of
equal yield Shaven-Whaley (1977) discuss the meaning 01 equal yield in a general equilibrium
context when government is not allowed to run defICits Equal yield is interpreted to moan constant
public utility Government base case utility is maintained in the counferfactual experiments With
balanced budgets the concepl of equal yield is unambiguous The new equUibrium prices and the
balanced budge condition will del ermine the minimal expendilure and taxes needed to maintain base
30
case publi~ utility Accordingly in general equal yield is inconsistent with equal nominal tax
revenue Some change in tax revenue Is necessary Different tax replacement schemes ate
considered to assure that enough tax revenue is collected This is the approach essentially followed
by most 01 the papers surveyed Only Feltenstein (1985) Goulder (198) and Jorgenson-Yun
(1984) chose not to follow an equal yield strategy
The question is of how to Interprete lila concept of equal yield when the government is alowed to
fUn deflcLts The optimal leve of expenditure for base ease public trliUty can now be taAinanced
bond financed or financed by a mix of bonds and taxation We have several versions of equal yield In
particular equal yield may now be consistent with equal tax revenUE Furthermore some meaSure
ltgtf financial crowding out effects of government deficns can be inferred from the comparislon of the
several equal yield alternatives These issues are extensively discussed ~Pereira (1987b)
44 Price expectations and the dynamic generalization of compensation
Indicators
The sum of equivalent and compensation variations over households is the most widely used
aggregated measure of efficiency gains or losses In the context of $teady~state eomparisions andor
when complete future markets exist andlor perfect foresighl is assumed there are no difficulties
associated with the use of the standard Hicksian indicators In fact correct future prices are known
in these cases
If expectations are not selfmiddot fulfilling andlor future markels are not available a dynamic
generalization is necessary 8allardmiddotGoulder (1985) provide some steps in that direction by
defining an indicator that accomodates periods far into the future when households do not have
perfect foresight but a steady~state prevails On he other hand this issue is more fundamental in
i
[
I
i I I I I [
I I
bullI
31
the contex~ of a Gner1 lemporary equilibrium ftamawor~ In such Circumstances Pereira
(1986a) dev~lops a cnamlc generalization which is obtained as the present discounted value of a i i
sequence of short-fUn optimal expenditure functions consistent with a base case expected future I stream of IlWities I
32
5 Empirical evidenee from selected policy Issues
Dynamic tax models have generated several important results which escaped Sialic modellers
The following is a seleCpoundd set of issuos whIch stress the marginal benefits of dynam~ modelHng of
economic behavior in innovltlttive areas of analysis The discussion of a consumption tax emphasizes
the benefits of dynamic tusehold behavior and intergenerational aspects The study of the
flliminatlon or the re~intToduction of the investsnemt tax credits is made meaningful by the dynamic
modelling of production betlavor Modelling governmenl in a dynamic conlexts allows the modeller
to study the irrrpact of government deflcits Finally a dynamic framework lets us appreCiate the I
relevance of consumer expectations in terms of the evaluation of policy alternatives
i
i
51 Consumption Tax
t
By Its very nature a consumption tax can not be adequately investigated with a static model bull
Fullerion-Shoven-Whaicy (1983) use Ihe model of the US economy s described in
Ballard-Fullerton-Shoven-Whalley (1935) to evaluate the movement from the currenl US tax Isys1em to a progressive consumption tax Since their modeJ Incorporates a Jaborleisure choice
where leisure is an untaxed commodlty their results reflect the fact that both the consumption tax
and Ihe present system are dslortionary
Concentrating on the rntertemporaf distortions they show that sheltering more saving from lhe
current US income tax could improve econornic efficiency even if marginal tax rate increases are
necessary in order to maintain government fevenue~ At first you have a revenue shortfall
However the economy moves to a higher sustatned growth path and uitimatefy lower tax rates
cangenerate the same revenue path Also wages increase in the long run with this policy The
bull
bullbull ~
33 shyt
i
I
I
present value of welfare gi[lS for a potiey of complete savin9 deduction wilh marginal tate
adjustments (consumption lax) is simulated to be around $500 billion to $600 billion 01 1973
dollars
They also investiga1e th 1ngth of ttme it takes th~ economy to adjust to these policy changes
Roughly they estimate the long run to b thirty years although this figure Is very sensitive to the
specification of the savings elasticity - a crucial parameler in this model
52 Investment T~x Credit
Goulder-$ummers (1 S87) address the impact of eliminating the lnvestmenl Tax Cradit (ITC) on
intersectoral capital formation and on economic growth Their model isYa1ticularly adequate to
addies$ this issu~ in that is postulates a forward looking investment m9we with adjustment costs
They show that the eliminating the ITC causes a reduction in the rale of investment Inveslment is
estimated to fall by about 7 in the short-run and by about 12 in Ihe new long run steady state
In lurn a previous policy anncuncement lowers Ihe overall al1ractiveness of Investment and leads 10
a downward shift of the Investrrent profile On the other hand the combined effect of a revenue
neutral simultaneous efimiraron of the ITC and reduction of the corporate tax rates is a long run
reduction ef the capital Sieck by 35 This pattern suggests that a revenue neutral increament in
both Ihe corporale tax rates and the inveSlment lax credits would be preferable 10 the formula
adopted in Ihe US Tax Reform Act of 1986
Pereira (1987d) addresses the impact of the ITe from Ihe oland point of Ihe curTenllax syslem
under the Tax Reform Act of 1986 Given Ihe concern over the potential depressiVe effects upon
savings and investment of the new tax system in conjunction with deficit reduC1ion mechanisms of
the GrammmiddotRudman type a pertinent question is should Ih lTC be e-introduced In the context
bull
34
-of his mod~1 of the US economy Pereira focuses on the trade-off between the distortion in the
intersectoral allocation of capital induced by the lTC the lower tax revenues it generates and
potential financial crowding-out effects on one hand and the positive effects of lowering the relative
price of new capital goods on the other hand Preliminary results confirm the qualitative results in
Goulder-Summers (1987) suggestting a welfare gain of about 2 of the present value of GNP in
the best scenario of absence of replacement taxes Furthermore preliminary results show an
important time pattern to welfare gains with the average benefits increasing the further you look
into the future This pattern is due to the presence of constraints to intersectoral mobility of capital
and inertia in the adjustment towards the optimal capital levels as reflected by the presence of
adjustment costs
I
I
bullAuerbach-Katlikaff (1gB) in the context of their intergenerational model of the US economy
consider the impact on savings capital formation and interest rates of deficit policies and balanced
budget increases in government consumption They conclude that deficit finance and government
consumption can significa-ntly crowd out capital formation and lower the welfare of future
generations However crowding out from deficit finance is a very slow process because it results
from increased government spending over potentially long horizons Also deficit policies may
substantially influence the longmiddotrun interest rates while leaving the short-run rates essentially
unaffected Finally and in opposition to the central role adjustment costs seem to play in both
53 Financial crowding outeff~cts of government deficits
Goulder-Summers and Pereiras models Auerbach-Kotlikoff suggest that the time path of interest
rates induced by a policy of deficit financing seems to be insensitive to the presence of adjustment
costs
bull
35
54 The role 01 consumer expectatlons
The expectations analysis has uncovered one of the benefits of dynamic modelling
Ballard-Goulder (1985) find that the appeal of adopting a consumption tax depends on the level of
foresight possessed by the consumers Furthermore additional foresight may be welfare worsening
This is a second best type of result This tends 10 occur under policies that lead to capital deepening
and declining rate of return to capital over time To the extent that consumers have more foresight
they will be beller equippod to anticipate the fall in the rental price of capital Consequently if a
saving incentive is enacted people save more with myopia than with perfect foresight Given the
exi~tence of taxes on capita and the discrepancy between the private an~--sccial returns to capital
the greater savings with myopia is beller socially Ballard-Goulder find that the welfare gains
from a consumption tax are reduced by about 10 when we move from myopia to farleet foresight
Therefore the attractivenness of adopting a consumption tax seems fairly [obust across these
foresight specifications
I
bull
36
6 Tbe power and weaknesses 01 dynamic modelling the example 01 corporale
tax Integrallon
The Issue of corporate tax integration allows us to sh~w the stre-ngths and weaknesses or the
dynamic approach Empirical evidence using CGE techniques indicates that depending on the precise
scheme of integration and the 1ax replacement methods integratiOn may have substantial effects In
Ihe work of FuliertonmiddotKingmiddotShoven-Whalley (1980 1981) total integration was found to yield an bull
annual static eHielency gain of $4 to $8 billion in 1973 dollars Simulated dynamic gains may be as
large as $695 billion or abcu 14 of the present value of tulUre consumption and leisure in the
US economy These gains result primarily from interindustry reallocalions of Investment and an
improved inter~emporal allocation of consumption
Mora recent work emphasizes hOw consumers asset portfolio decisions and firms finanCial
decisions affect the efficiency gains from integration Slemrod (1980) focusing on consumers
asset ponfoio deCisions finds static efficiency gains which are about twice -as Iarge as those
repOrted iiy Fulierton-King-ShovenWhalley (1961) FuUrlon-Gurdon (1983) focus on the
firms financial decisions They roporl efficiency gains of 6 of GNP from the elimination of the
tax distor1ions favoring debt However when they eliminate the corporate tax and repiace it with
increaSed persona income taxes addlHonal dls10t1ions are created in the optimal labot~leisure
decisions These distortions tend to dominate the analysis slJ(h that the overall effects of complete
integration are very modest Galpr-LuckmiddotTodr (1986) address simultaneously consumers
asset pOrlfollo decisions and the firms financial decisions They also find very modest efficiency
gains from integralion
The above results are important but may be severely biased There are severa aspects of
economic behavior and modelling crucial for the study of Income tax Integration whiCh have no been
i 37
captured innny of the above papers One aspect is the absence of government deficitsmiddot a balanced
budget is assumed It is true Ihal resource crowding OUI is caplured In these mOdels but financial
crowding out induced by government spending is nOI Second investment is not derived from
optimtztlcn behavior Investment behavior passively accomodates endogenous saving decisions As
a consequence the differential impacts of policies in the incentives to save and to invest are not
captured Aso full capital mobility across secors is assumed with instantaneous capital
adjustmenlS towards optimal levels This assumption rules qut different costs of capital across
seclors and therefore differentiated reactions to tax policies changes Finallyhe modelling of both
gov9rnmerlt deficits and of endogenous real and financial investment decisions necessitates the
~nsideraiion of a dynamilt framework and the introduction of financiaJ assets govemment bonds and
iprivate financial assets A dynamic framework also highlights the efficiency effects of Integration on r
ithe ptimal intertemporal decisions The above models are either sIalic (Slemrods and
GalpermiddotltJckemiddotTode(sj or a dynamic sequ~nce of otherwise Slatic model bullbull I Pereir a (1986b) develops a dynamic applied general equilibrium model bull to study the tbull
efficiency effects of integration on the growth and allocation of investmenl across sectors Special
Iattention is paid to the structural affeelS of resource and financial crowding oul induced by
gove mect spending and deficits and to the real and financial Investment reactions across industries
to both tax integration and flnancial crowding out His mode departs from previous work on income
lax Integraticn and for that matler from most 01 the CGIE literalure for tax policy evaluation in
several direcllons It encompasses an endogenous sequential equilibrium struelure founded on
dynamic behavior with flexible expectations Government deficits are optlmaUy determined
Investment decisions are tward looking and the resull of optimizing behavior Several financial
assetsmiddot public and privatemiddot are considered
Simulation resulls Indicale that the welfare gains from Integration under large deficits are at
32
best lJlodest when measured in terms of the GNP The elimination of the corporate tax and its I replacemenl by increased income lax rates yields long run benefits which are never larger than
2 of the presenl yalue ct telr consumption and leisure onder the previous tax regime and 1
under the current tax regime However the short run w~lare effects of full integration tend 10 be
very low and in in some scenarios even negative This is a flew intertemporar pattern of efficiency Ieffects which reflects an adjJsiment lag in the interindustry investment decisions due to the
existence of costs of adjustment On the of her hand partial integration achieved by excluding I dividends from the corporale lax base syslematically yields negalive efficlencyeffeclS In Ihe long Inm these can be as high as 3 of present value of the future value of consumption and leisure This
is a new second best effect suggesllng Ihat less than complete integration may have perverse I efficiency effects
The dynamic slructure of PerclrBs model snowing for an improved freatment of real investment
declsions and government deficts provides a potential setting for a more accurate measure of the
costs and benefits of inlegration The simulation resullS suggest lower benefits and an intertempora Ibull
pattern of growing benefits
Simulation results also clearly suggest robust negative effects from partial integration How
reliable is this result If dJvidends were deductible from the corporate lax base (as are interest
payments) the preferenlial Ireatmant of debl over equily would be ellminaled Corporallons should
be expecled 10 react by decreasing the optimal deWeqully rallo However corporale financial
gecisions are exogenously set In Pereiras model as in aft of the other models surveyed that go as far
as dealing with the issue
Also withdrawing dividends from Ihe corporate tax base is a way of eliminating Ihe double
taxation of dividends at belh Ihe corporate and personnal Income levels How will the optimal
aliocellon of household saving accomodate that change in Ihe relalive return to corporale equity
39 I
t However hOusehold portfolio decisions are exogenouly set in Pereiras model as in all of the other
models surveyeltj go as far as dealing with the issue
IIt is sate to say that the evaluation 01 the benefits 01 partial Integration as defined above are
sevrey underestima1eO Meanwhile the distortionSoenerated by the increase in the marginal
personal tax rates designed to ralse the revenue foregone by the dividend exclusion are fully
aCCltlunted for A good measure of the costs of integration together With a poor evaluation of the
benefils may very well be the reason why partial integration is Simulated 10 yield neg alive
eHikiency gains
On a different vein as most of Ihe model surveyed here Pereiras is a closed economy model It
is legitimate to wonder how the resullS would change If international capital flows were anowed
IThis almost certainly would affect financial crowding out since a good d~al of the capital inflows
could be used to finance public debt
For these various reasons the results should be treated as preliminary but they strongly I suggosl that the dynamic structure provides valuable new insights into the corporate tax integration SSU2
40
7 Concluding remarks I I I i
What has been acccmplished Ihis far The success of Ihe CGE research in laelding the challenge
of dynamic modelling can not be denied
I iGreat progress in the modelling of household behavior has been achieved Promising
developm~nts in the modeUing of production and government behavior have also been accomplished
Interesting innovations in the concept of equilibrium and computation techniques were discussed
The policy analysis was greatly enriched by considering transitional effects together with longer
run steadymiddotstate equilibrium paths generalized equal yield strategies accomodating different
financial crowding out impacts and generalized dynamic policy evaluation indicators
An important set of issues have been addresed by the models surveyed Traditional issues
focusing on the effects of capital taxation and consumption taxes have been pursued In terms of
capital taxation for example the intertemporal nature of the issue was enhanced by allowing an
optimal evolution of capital stock in the economy and forward looking optimal investment decisions
In Andersson (1987) GouldermiddotSummers (1987) and Pereira (1986b 1987d) this is coupled
with a improved treatment of several tax provisions like investment tax credits and depreciation
allowances In terms of the consumption taxes developments in the specification of Intertemporal
household behavior the introduction of overlapping generations and the modelling of
intergenerational links via bequest motives as well as a closer attention to demographic evolution
provided a much improved economic setting for the understanding of the several aspects of the
problem
Furthermore dynamic modelling has permitted the addressing of a set of new issues Policy
Issues ranging from dynamic tax policy analysis to the evaluation of exogenous changes in
government expenditures to the impact of different methods of financing government expenditure to
4 1
the importance of tinanciai croHrii1g out to 1he relevance of consumorS expectations in tne policy
results have been addressed
Despile all of the progress and in part due to such progress several avenues are wide open for
much neded additional research The following seem to be the most Interesting and promising
areasmiddot First the specincation of liquidity constraints to household behavior may help to obtain a
better understanding of policy chenges that affect directly the interest rates in the economy Second
major attention needs to be paid to the incorporafion of a roraign sector (with open capital markets)
Third some efforts are needed in terms of finding adequate computation techniques now that
dynamic modelling eally makes the curse of dimensionality worse than eVOrbull
The modelling of financial markets is the single most unsatisfactory speet of the dynamic
models surveyed here~ The problems associated with provlding a mo~ adequate treatment of
financial markets and in general endogenous and optimal financial deci$ions bull both at the corporate
and at the personal levelsmiddot have 10 be addressed That necessarily requires Introducing uncertainty
into the models To b adequate this must go well beyond the mere assumption of point expeetalions
with the whole array of modelling implementalion and evaluation problems thai generates
bull I i
42
REFERENCES
Abel A and O Blanchard 1983 An Intertemporal Model of Saving and Interest Econometrica
51 pp 675-92
~ Andersson K 1987 Tar-ation of Capital in Sweden - A General Equilibrium Model Ph D
Dissenation UnIversity of Lund Sweder
Auerbach A L KotlikoH 1983 National Savings Economic Welfare and the Structure of
Taxation in M Feldstein ed Behavioral Simulation Methods in Tax Policy- Analysis Chicago
University of Chicago Press
Auerbach A l Kotlikoff 1984 Social Security and the Economics of Demographic-
Transition in H Aaron and G Burtless eds Retirement and Economic Behavior Washington DC
Brookings Institution
Auerbach A L Kotlikoff 1987 Dynamic Fiscal Policy Cambridge Cambridge University
Press
Auerbach A L Kotlikof and J Skinner 1983 The Efficiency Gains from Dynamic Tax
Reform Internatiooaj Economic Reyiew 24 pp 81middot100
Ballard C 1983 Evaluation of the Consumption Tax with Dynamic General Equilibrium
Models Ph D Dissertation Stanford University
Ballard C 1987 Tax Policy and Consumer Foresight A General Equilibrium Simulation
Study Economic Inquiry 25 pp 267-284
Ballard C and L Goulder 1985 Consumption Taxes ForeSight and Welfare a Computatonal
General Equilibrium Analysis in J Piggott and J Whalley eds New Deyelooments in Apolied
General Equilibrium Analysis Cambridge Cambridge University Press
Ballard C D Fullerton J Shoven and J Whalley 1985 A General Equilibrium Model for
t 43
Tax Eoliey Eyalua~iQO NBER University or Chicago Press
Borges A (1987) bull A Survey of Compula1ional General Equilibnum Models CECD Economic
Slud~ 8
Bovenbig L 19B3 imperfect Capital Mobility in a Perfect Foresight Herberger Model
Mimea UC Br~eley
Bovenberg L 1984 Capitel Accumulation and Capital Immobility Qmiddottheory In a Dynamic General Equilibrium Framework PhD dissertation University of California at Berkeley t
Bovenberg L 19850 Dynamic General Equilibrium Tax Models with Adjustment Costs in A iManne ed bull 1985 EtQnomrc ECUjilbrium Model Formulation and Solution Mathematical
lProgramming Siudy 23 Arnsterdam NorthmiddotHoliand
Bovenberg L t 1986 Capital Income Taxation in Growing Open Eeondmies Journa of Pubtic
Eamp(omlcs 3t pp 347middot376 IBovenberg L ald W Keller 1981 Dynamics and Nonlinearlties in Applied General
Equilibrium Models Internal Report Mlmiddot81middot208 Department for Statistical Methods CBS bull
Vooiburg The Netherlands
Charnley C 1981 The Welfare Cost of C~pltal income Taxation In a Growing Economy
IJournal of Political ECQnQm~ 89 3 pp 468middot496 I
Charnley C 19810 Efficient Tax Reform in a Dynamic Model of General Equilibrium I Mimeo Yale University I Charnley C 1982 On the Optimal Taxation In Stylized Models 01 Inlartemporal General
Equilibrium Mimeo Yale University I Decaluw B and A Manens 1985 Pays en Oeveloppement at Modeles Calculables OEquilibre
General - Une Revue de al litterature empirique Monographie Sene 9 Volume 3 Centre de
Recherche et Developpemont Economlque Unrslt de Montreal
44
bull
I I
I
I
1
Erlich $ V Ginsburgh and L van der Heyden 1987 Wher~ do Real Wage Policies Lead
Belgium bull A General Equilibrium Analysis European Economic Reyiew (forthcoming)
air R and J Taylor 1983 Solution and Maximum Likelihood Estimation of Dynamic
Nonlinear Ratioant Expeclations Models Ecooometrica 51 pp 1169middot1186
Feltenstein A 1984 Money and Bonds In a Disaggregated Ec~nonYin Searl 1-1 J Shoven
Eds Aoolied ~eoeral EqJiliDCiulD tioalYsis Cambridge University Press Cambridge
Feltenstein A 19B6 A Dynamic General Equllibrlu Analysis of Financial Crowding Out
Theory with an Application to Australia Journal of Public Economics 31 No1
Fullerlon D R Gorden 1983 A Reexamination of Tax Distorllons in General Equilibrium
Models in M Feldstein Ed Behayora Simulation Methods in Tax pOlicy Analysis Chicago
University of Chipgo Press
Fvllerlon D J Shaven and J Whalley 1983 Replacing the US Income Tax with a
Progressive Consumption Tax JoulOal of Public ECQnomics 20 pp3middot23
Fullerton D Y Henderson J Shaven 1984 A COmparision of Methodologies in Empirical
General Equilibrium Models of Taxation in Searl H J Shoven Eds ~pQlied Geoaral Jguilibrium
lIoalyss Cambridge University Press Cambridge
Fullerton 0 A King J Shaven and J Whaney 1981 Corporate Tax Integration in the United
Statesmiddot a General Equiiibdufi1 Approach American Economic Review 71 pp~ 677~691
Galper H R Lucke E Tader 1985 Taxation portlor Choice and the Allocation of Capital A
General Equilibrium Approach Mimeo
Gill p W Murray M Saunders M Wright 1986 Users Guide for NPSOL (Version 40) A
Fortran Package for Nonlinear Programming T echnicsl Report SOL 86middot2 Deparlment of
Operations Research Stanford University
Glnsbur~h V and L van der Heyden 1985 General Equilibrium with Wage Rigidities An
45
AppJication to Belgium in A Manne ed t Economic EQumprit1m~ Made Formulation and $oJioo
Mathematical Programming Study 23 Amsterdam NorthmiddotHolland ~
Goulet J 1968 1 Adjustrent CO$1S in the Theory o(the Firm Reyiew of EconomiC Studies 35 bull
pp47middot55
Goulder L 1985 Tax-Financed versus BondmiddotFinancedGhanges in Governent Spending
bull Mimeo Harva(d UniversilYmiddot
Goulder l J Shaven and J Whalley 1983 Domestic Tax Policy and the Foreign Sector The
importance of altemative foreign policy formulations to results from a general equilibrium tax
analysis model In ~n methods in Ibe antico of income from caoia1 ed Martin
Feldstein Chicago University of Chicago riSS
Goulder l L Summers 1987 Tax Policy Asset Prices and 3rowtly ~
A General Equilibrium Analysis NBER Working Pper No 212B
Grandmont J 1982 Temporary General Equilibrium Theory Ch 20 in Handbook 01
Mathemalical Eccnornics Vol II NorthmiddotHoliand
Harberger A 1959 The Corporate Income Tax An empirical Appraisal In lax Revision
Comoendjum Voll House Cornmitee on Ways and Means Washington DC Govemment Printing
Office
Harberger A 1962 The Incidence of the Corporate Income Tax JOULn1 of Political
Economy 70 pp 215middot240
Harberger A 1964 Taxation Resource Aliocalrlon and Welfare fn lb Role of Dir bullbulll and
Indjresl Taxes io tbe Fede[al Reserve s~I~Ill Princeton Princeton Unlveresity Press
James J 1985 New Developments in the Applications of General Equilibrium models to
Economic History in Piggott aand Whalley Eds
Jorgenson D and K Yun 1984 Tax Policy and Capital Allocation Harvard Inslitute of
bull
I
I f
Economic Research WP 1107
Judd K 1965 ShonmiddotRun Analysis of Fiscal Policy in a Simple Perfect Foresight Model
Journal of poJiHcal 5CCJiW1l pp 298middot319 r
Upton D J Potbc J Sachsa nd L Summers 1982 Multiple Shooting in Rational
Expectations Models fuoometrlc 50 1329middot1333
Manne A ed 1985 fconQrnjc Eoullibrium Model FOrIDujaljQo and Solution Mathematical
Pogrammlng Study 23 Amsterdam NorthmiddotHolland
Manne A 1985 Or the Formulation and Solution 01 Economic Equilibrium Models In A
Manne ed ECQOQlli-Jr~ibmiddotum ~tQdel EQlJulailon and Solution Malhematical Programming
Study 23 Amsterdam NorthmiddotHolland
Mansur A and J Whalley 1984 Numerical Specification of Applied General Equilibrium
Models in Scarf H J Shoven Eds Aoplied General Eouillbrium AnaiysectIsect Cambridge Universily
Press Cambridge
Merrill 0 1972 Applications and Extensions of an Algorithm thaI Computes FixedmiddotPoints of
Cenain Upper Semimiddotcominuou Point to Set Mappings PhD Disserlation Unlvel$ity 01 Michigan
Negishi T 1960 Welfar Economics and Exislence I an Equnlbrlum for a Competitive
Economy MelroeCOOQ11ka 12 92middot97
Pereira A 1985 On the Existence and Uniqueness of Optimal Ou1put Path lor CRTS Firms in
Adjustment Costs Technologies Mimeo Stanford University
Pereira A 1985a Consumer Expectations and Ihe COSI 01 UllUty In an Intertemporat
Framework Mlmeo Stanford University
Pereira A 1985b Corporale Tax Integrallon and Government Deficits A Dynamic General
Equilibrium Anaiysls Mlmeo Stanford University
Pereir A 1987 On Ihe Computation of the Users Costs of Stocks Ecooomi (forthcoming)
47
Pereira A bull 1987b Equal Yield Alternatives and Government Oeficlts Mimeof Stanford
University
Pereira A 1987c A Dynamic Applied General ~uHlbrium Model for Tax Policy Evaluation
Comlletamp Docurnent2tion Mimeo University of California San Diego (in progress)
Pereira A 1987d Should the investment Tax Credit be Rei(ltroduced Mlmeo University
of California San Diego (ir progreSs)
Piggott J and J Whalley eds 1985 tJew QeyeQpments in Applied General EQuilibrium
bullenalysjs Cambridge Cambridge University Press
Preckel Pbull 1985 Alternative algorithms for computing economic equilibria In A Manne Ed
Radnor R 19amp3 Equllibrium under Uncertainty In K Arrow M Intrilligaor ods
I ibull
Robinson S 1 g86 Mullisectorai Models of Oeveloping CountrieS a Survey Oivision of
Agricullure and Resources UC Borkeley WP 401
Scarf H 1967 The approximation of fixed points of a continuous mapping ~MM Jouwal of
epplid Mathornallcs 15 pp 1328middot43
Scarf H with T Hansen 1973 00 the Computation of EconomiC Eguilibria New Haven Yale
University Press
Scarf H J Shoven Eds 1984 61l1llied Genera Elujlibrium 6nalyssect Cambridge University
~ Press Cambridge
Shoven J 1976 The Incidence and Efficiency Effects of Taxe on Income from Capital
Jurnal of ramie1 Economy 84 pp 1261-83
Shoven J 1983 Applied General Equilibrium Tax Modemng IMF Slaff PaoerS 30
Shoven J and LB Simon 1986 Share Repurchases and Acquisitions An Analysis of which
Firms Participate NBER Working Paper No 2243
bull
48
IShaven J and J Whalley 1972 ft General Equmbrium Calculation of lhe Effects of
Differential Taxation of Incoma from Capital in the US Jauroe of public Econgmics 1
pp281middot321 (
Shoven J bull and J Whalley 1973 General Equilibrium with Tax$s a computation procedure
and an existence pr00f ReviEhi of Economic Studies 40 pp 475middot490
Shoven J J Whalley 1977 Equal Yield Tax Alternatives A General Equilibrium
Computa~ui0nal Tfchnique J~HJrnal of PUblic EconomIcs 8 pp 211~224 bull
Shaven J J Whalley 1984 Applied GeneralmiddotEquilibrlum MOdels of Taxalion and
International Tr8de An Introduction and Survey Journal of Economic literature 23t pp
1007middot1051
Slemrod J 1980 A General Equilibrium Model of Capital Income Taxation Ph D bullDissertation Harvard Univnrsity
Slemrod J 1983 A General Equilibrium Model of Taxation with Endogenous Financial
Behavior in M Feldstein edbull Behavioral Simulation Methods In Tax Policy Analysis Chicago
UniVerSity of Chicago Press
Stone J 1985 Sequential Optimization and Complementarily Techniques for computing
Economic equilibrium in A Manne Ed
Summers L 1981 Taxation and Corporate Investment a q-Theory Approach BroQkinas
Summers L 1985 Taxation and the Size and ComposHion 01 the Capital Stock An Asset Price
Approach Mlmeo Harvard University
Whalley J 1975 A Geneal Equiloibrium Assessment of the 1973 United Kingdom tax
reform Economica 42 pp 139middot61
1JLTIllOS JIQll~)ltG PAPERS PUBLICADOS bull
I
I I
j
112 70 - LUCENA iJiCIIJo ~ IIEnvironrilent Honitoring and Orqanization Structure T (Setellibro 1987)
nI 71 - L~ENA Diogo A Note on the Representation of Information Stxueture (SeteJObro 1987)
nQ 72 - LUCENA D1CgO Envuorment 111)nitoring and Organization
Structure II (Setelilhro 1987)
ril 73 - AlmmE3 Antonio Pais e CAsPAR Yitor bull Tributacao Inoentivos e InvesUlllentos Analise Qualitativa noYellibro 19S7)
n2 74 - liAlltOSA rose Pedro Opt1lllal lIage R1gidity A SUggestedtlthodology to Teet tb Theory and an Application (Outubro 1987)
nI 75 - CABRAL L 11 B Three lIotes on SYJIIlletric Gemes witt itsYJIIlletric Equilibri (NoveJObro 1987)
n2 76 - LUCENlt D10gOTo Search or Not to Search (Fevereiro1967)
nQ 17 - SA Jorge asconcelos euroI A Theory of SynergyB (Fevereiro1987)
n2 76 - VILARES tanuei Jose Os Be IntenaMios IlIIportados ColIIO Factor de Produao (Julho 1987)
nQ 79 - SA J~rge VasCQncel03 e HoY to Copete and Co~unicate in tature Industrial Products_ (aio 1988)
n2 80 - ROIl Rafael Learning and Capacity Expansion in a New larket Under Uncertainty (Fevereiro 1988)
n2 81 PEREIRA Alfredo llarvilo Survey of Dynamic Computational General EquilibriUlll Hodel for Tax Policy Evaluation (Outubro 1987)
Qualquer in ormacao sabre o 1IIorlltlng Papers ja publlcados sera prestada pelo Secretariado de Apoio aos Docentes podendo os ~e3OSf
ser adquirido5 ne Secyo de Vlndas da Faculdade de EconollLia UNL na Travessa Esteviio Pinto Cltpolide - 1000 LISllOA
J
2S
flexible am9unt of foresight In terms of the number of years over which price movements are ~
foreseen Pereiras model (1986b 1987b)) is flexible in a somewhat different way in that it can
indude any range of foresight from myopia to perfect foresight The choice between the perfct foresight and lemporary equilibrium is ultimately to be made on
philosophic grounds It can be argued that less than perfect expectations Imply that agents are
irralional in some way (see Auerbach-Kctiikoff (1987 p 10) However the reverse argument can
be made One can question wiether agents are reaHy rationar and perfectly knowledgable about
future prices
Recent evidene of Balird (1987) Ballard-Goulder (19851 Goulder (19851 and Pereira
(1986b 1987d) suggests that the choice in modemng expectations is an important one They show
that the degree of foresight inl0 thc future (ranging from perfect foresight to-myopic expectations)
may have dramatic impacts on the pollcy conclusions of Ihe model Accordingly the best research
strategy may b to design models which are flexible enough 10 allow for different rules regarding the
formation of expectations With the exception of the articles just mentioned sensitivity analysis
bullhave previously not been performed along this dimension
In terms of implementation the two concepts of equilibrium - perfect foresight and temporary
equilibrium have different implications The dimensionality of the equ1ibrium soll1ion algorIthm
is involved Suppose we have a model with ten markets to be run for a period of 50 years Aside
from normalization a perfect foresight model implies computing prices In 500 dimensions while a
-t~mporary equilibrium model requires sorving SO equilibria each in ten dimensions Given that
computational speed often varies with the cube of the number of dimensions the lemporary
equilibrium formulation is potentially strikingly more feasible However Ballard (1987) and
Goulder-Ballard (1985) have devoloped techniques to greatly speed the computation of a perfect
foresight equilibrium
26
j
I
4 Jrnplemenlalion and Issues on policy evaluallon
41 Calibration and equilibrium comparlslon
CGE models at typically parameterized by the use of a calibration procedure Some parameters
are exogenously given However some crucial parameters are determined in such a way that the
model replicates the data for a given base year See MansurmiddotWhalleY 11984) for an eXlensive
discussion of 1~i$ issue
CalibratIon in a dynamic context is generaliy interpreted as requiring two properties First
replication of base year data is required Second the model is parameterized to simulate an
intertemporal balanced growth path when Ihe base policy is maintaineq This Is the approach
followed by Ballard (1983) BallardmiddotGoulder (1985) Goulder (1985) and SummersmiddotGoulde
(1986) It follows the practice of BallardmiddotFuliertonmiddotShovenmiddotWhalley (1985) with their
sequential cquilibrium dynamics
Other authors Auerbach-Kotlikof (1983 1984 1987) Boyenberg (1984 1985 1986)
and Pereira (198Gb 1987d) follow whal we cali a qualitalive calibration The slructural
palametars are exogenously chosen so that lh economy follows a reasonable path into Ihe fulure
ThaI has 10 do wllh the fact thaI given the recursive nature 01 the dynamic economy and aside tom
such structural parameters only iniUa stock values are needed to run these models Given initial
conditions on the stocks of say private wealth capltal and government debt agents will optimize
and Ihereby generate a lirst round of net demands and equilibrium conditions In lurn the
equilibrium prices will determine the evolution of the stock variables into the next perIod
There are several potential problems with calibration in the conlext of dynamic models The
assumption of a steadymiddotstate growth ~ath in the base case can be questioned First while
27 I Ibullbull
steadymiddotstaU is a possibity it certainly is not the only meaningfull solution to dynamic models
Even in the case of a perfect foresight equilibrium the model implies an equilibrium path which i may or may not involve colanced growth The model not the modeller should dictate the nature 01 I themiddotbase case path Second in the context of a temporarY equilibrium path a steady state solutlcn is Inol a likely model QutcOn1tl In fact unlessmiddot expectations are stattc short run behavior consistent
with a steady-state evolution will in general not be generated On the other hand if static I expectations are self~fufjili[1g we have in fact a perfect foresight model Third even the base year
replication requirement may cause problems In Ihe cOntexl of temporary equilibrium Any
calibration parameter would be conditional on expectation rules which is probably an undesirable
feature
The nalur of the two-requirement calibration strategy is very muchin the spirit of the bull
traditional design of the comparision of alternative equilibria comparis_ion between a steady-state
base case on one hand and alternative paths including a transition period and a finat sleadyyenstate on
the other hand Pereira (198Gb 1987d) compare different (not necessarily steadystate)
equilibrium paths The arguments against such a procedure are based on the idea that the impact of
policy changes can be observed most easily since all departures from the steadymiddotstate can be
attributed to the altemalive polioy On the other hand the base case so dEfined as a steadymiddotstate is
consist~nl with previous work in a less dynamtc setting and Iherefore allows a common standard
for comparing model results
42 Computation algorithms
We are still al a stage in which basically each author uses a different computational technique
Th development of dynamic models - in parlicular wilh adjustment costs andior perfect foresight
bull
28
has corresponded with the decline in the use of fixed point algorithms In -fact given the relative
arge dimensions inevitably involved such algorithms lend to be very inefficient at the best and
chen prohibitively stow See Stone (1985) and Preckel (1965) for a comparative assessment of
different computation techniques Among the models surveyed only Ball1rdFulle(ton~
-ShavenmiddotWhalley (1985) and Feltensteln 1985 use Merrills variant of the fixed point
algorithm 1echnique
AuerbachmiddotKoHikoff (1963 1984 1967) follow a three stage procedure They first compute a
base case steady-state then a revised case steady state and finally transition path for the economy
berveen these two steady-states tn all stages a Gauss-Seidel iterative procedure is used
Ballard (1982) Ballard-Goulder (1965) Goulder (HiSS) and Summers-Goulder (1986)
~
use a method developed by Ballard and Goutder which is similar in many awects to the FairmiddotTaylQr bull
(1983) algorl1hm Short-run equilibria are calculated (using Merrills algorithm) parametric on
nrice expeC~lions The model is then iterated to generate self-fulfilling intertemporaf expectations
and the corresponding perfect foresight equilibrium In a relatively similar approach Andersson
1987 uses a simulation program SIMNON developed In the University of Lund This program
can handle two-paint boundary problems In a fashion consistent with the multiple shooting
algcrlthm (see Lipton-Poterba-Sachs-Eummers (1982))
Boyenbergs computational approach (1985 1985) differs from the other models surveyed in
that he relies heavily on analytical techniques Computations are done by using a dynamic version of
Johansons linearization method Being essentially determined by the continuous time nature of the
I I
f
model thiS linearization model has the disadvantage of confining the analysls to infinltesinal changes
around the base case equilibrium (see Bovenberg (1985) p 53) Pereira (1986b 1987d) uses an optimization algorithm NPSOL - developped by
rGillMurray-Saunders-Wrlght (1986)) This algorithm is used 10 compute the sequence of
I
29
Shortmiddotrun temporary equilibrium which makes up the intertemporal equilibrium path The
equilibrium conditions are seen as nonlinear equality constraints in the minimization of an artlcial
objective function The prices are normalized to the unit simplex by an additional linear equality I I 1 bull constraint
II ErlichmiddotGinsburgh-Heyden (1987) follow a unique approach in that they use a variant of the
optimization technique introduced by Negishl (1960) The economic equilibrium can be generated i
as a solution of a mathemallcal program Ihe objective function of which is a weighted sum of the
utility functions of the various agents while tho constraints set consists of the market dearing 1
conditions Ginsburgh-Heyden (1985) have extended Negishis resulllO tho case of downward price
lrigidities I
Finally the paper by Jorgonson-Yun (1984) is also unique in tIlat it is the only I
econometrically estimated model Different blocks for the consumption and production Side of the
-
model are separately estimated to provide the necessary structural parameters i
I The diversity of computation techniques is yet another indicator of the exploratory nature of the
body of literature surveyed in this article
43 Equal yield comparisions
The link between a base case and counteriactual simulations Is usually provided by the concept of
equal yield Shaven-Whaley (1977) discuss the meaning 01 equal yield in a general equilibrium
context when government is not allowed to run defICits Equal yield is interpreted to moan constant
public utility Government base case utility is maintained in the counferfactual experiments With
balanced budgets the concepl of equal yield is unambiguous The new equUibrium prices and the
balanced budge condition will del ermine the minimal expendilure and taxes needed to maintain base
30
case publi~ utility Accordingly in general equal yield is inconsistent with equal nominal tax
revenue Some change in tax revenue Is necessary Different tax replacement schemes ate
considered to assure that enough tax revenue is collected This is the approach essentially followed
by most 01 the papers surveyed Only Feltenstein (1985) Goulder (198) and Jorgenson-Yun
(1984) chose not to follow an equal yield strategy
The question is of how to Interprete lila concept of equal yield when the government is alowed to
fUn deflcLts The optimal leve of expenditure for base ease public trliUty can now be taAinanced
bond financed or financed by a mix of bonds and taxation We have several versions of equal yield In
particular equal yield may now be consistent with equal tax revenUE Furthermore some meaSure
ltgtf financial crowding out effects of government deficns can be inferred from the comparislon of the
several equal yield alternatives These issues are extensively discussed ~Pereira (1987b)
44 Price expectations and the dynamic generalization of compensation
Indicators
The sum of equivalent and compensation variations over households is the most widely used
aggregated measure of efficiency gains or losses In the context of $teady~state eomparisions andor
when complete future markets exist andlor perfect foresighl is assumed there are no difficulties
associated with the use of the standard Hicksian indicators In fact correct future prices are known
in these cases
If expectations are not selfmiddot fulfilling andlor future markels are not available a dynamic
generalization is necessary 8allardmiddotGoulder (1985) provide some steps in that direction by
defining an indicator that accomodates periods far into the future when households do not have
perfect foresight but a steady~state prevails On he other hand this issue is more fundamental in
i
[
I
i I I I I [
I I
bullI
31
the contex~ of a Gner1 lemporary equilibrium ftamawor~ In such Circumstances Pereira
(1986a) dev~lops a cnamlc generalization which is obtained as the present discounted value of a i i
sequence of short-fUn optimal expenditure functions consistent with a base case expected future I stream of IlWities I
32
5 Empirical evidenee from selected policy Issues
Dynamic tax models have generated several important results which escaped Sialic modellers
The following is a seleCpoundd set of issuos whIch stress the marginal benefits of dynam~ modelHng of
economic behavior in innovltlttive areas of analysis The discussion of a consumption tax emphasizes
the benefits of dynamic tusehold behavior and intergenerational aspects The study of the
flliminatlon or the re~intToduction of the investsnemt tax credits is made meaningful by the dynamic
modelling of production betlavor Modelling governmenl in a dynamic conlexts allows the modeller
to study the irrrpact of government deflcits Finally a dynamic framework lets us appreCiate the I
relevance of consumer expectations in terms of the evaluation of policy alternatives
i
i
51 Consumption Tax
t
By Its very nature a consumption tax can not be adequately investigated with a static model bull
Fullerion-Shoven-Whaicy (1983) use Ihe model of the US economy s described in
Ballard-Fullerton-Shoven-Whalley (1935) to evaluate the movement from the currenl US tax Isys1em to a progressive consumption tax Since their modeJ Incorporates a Jaborleisure choice
where leisure is an untaxed commodlty their results reflect the fact that both the consumption tax
and Ihe present system are dslortionary
Concentrating on the rntertemporaf distortions they show that sheltering more saving from lhe
current US income tax could improve econornic efficiency even if marginal tax rate increases are
necessary in order to maintain government fevenue~ At first you have a revenue shortfall
However the economy moves to a higher sustatned growth path and uitimatefy lower tax rates
cangenerate the same revenue path Also wages increase in the long run with this policy The
bull
bullbull ~
33 shyt
i
I
I
present value of welfare gi[lS for a potiey of complete savin9 deduction wilh marginal tate
adjustments (consumption lax) is simulated to be around $500 billion to $600 billion 01 1973
dollars
They also investiga1e th 1ngth of ttme it takes th~ economy to adjust to these policy changes
Roughly they estimate the long run to b thirty years although this figure Is very sensitive to the
specification of the savings elasticity - a crucial parameler in this model
52 Investment T~x Credit
Goulder-$ummers (1 S87) address the impact of eliminating the lnvestmenl Tax Cradit (ITC) on
intersectoral capital formation and on economic growth Their model isYa1ticularly adequate to
addies$ this issu~ in that is postulates a forward looking investment m9we with adjustment costs
They show that the eliminating the ITC causes a reduction in the rale of investment Inveslment is
estimated to fall by about 7 in the short-run and by about 12 in Ihe new long run steady state
In lurn a previous policy anncuncement lowers Ihe overall al1ractiveness of Investment and leads 10
a downward shift of the Investrrent profile On the other hand the combined effect of a revenue
neutral simultaneous efimiraron of the ITC and reduction of the corporate tax rates is a long run
reduction ef the capital Sieck by 35 This pattern suggests that a revenue neutral increament in
both Ihe corporale tax rates and the inveSlment lax credits would be preferable 10 the formula
adopted in Ihe US Tax Reform Act of 1986
Pereira (1987d) addresses the impact of the ITe from Ihe oland point of Ihe curTenllax syslem
under the Tax Reform Act of 1986 Given Ihe concern over the potential depressiVe effects upon
savings and investment of the new tax system in conjunction with deficit reduC1ion mechanisms of
the GrammmiddotRudman type a pertinent question is should Ih lTC be e-introduced In the context
bull
34
-of his mod~1 of the US economy Pereira focuses on the trade-off between the distortion in the
intersectoral allocation of capital induced by the lTC the lower tax revenues it generates and
potential financial crowding-out effects on one hand and the positive effects of lowering the relative
price of new capital goods on the other hand Preliminary results confirm the qualitative results in
Goulder-Summers (1987) suggestting a welfare gain of about 2 of the present value of GNP in
the best scenario of absence of replacement taxes Furthermore preliminary results show an
important time pattern to welfare gains with the average benefits increasing the further you look
into the future This pattern is due to the presence of constraints to intersectoral mobility of capital
and inertia in the adjustment towards the optimal capital levels as reflected by the presence of
adjustment costs
I
I
bullAuerbach-Katlikaff (1gB) in the context of their intergenerational model of the US economy
consider the impact on savings capital formation and interest rates of deficit policies and balanced
budget increases in government consumption They conclude that deficit finance and government
consumption can significa-ntly crowd out capital formation and lower the welfare of future
generations However crowding out from deficit finance is a very slow process because it results
from increased government spending over potentially long horizons Also deficit policies may
substantially influence the longmiddotrun interest rates while leaving the short-run rates essentially
unaffected Finally and in opposition to the central role adjustment costs seem to play in both
53 Financial crowding outeff~cts of government deficits
Goulder-Summers and Pereiras models Auerbach-Kotlikoff suggest that the time path of interest
rates induced by a policy of deficit financing seems to be insensitive to the presence of adjustment
costs
bull
35
54 The role 01 consumer expectatlons
The expectations analysis has uncovered one of the benefits of dynamic modelling
Ballard-Goulder (1985) find that the appeal of adopting a consumption tax depends on the level of
foresight possessed by the consumers Furthermore additional foresight may be welfare worsening
This is a second best type of result This tends 10 occur under policies that lead to capital deepening
and declining rate of return to capital over time To the extent that consumers have more foresight
they will be beller equippod to anticipate the fall in the rental price of capital Consequently if a
saving incentive is enacted people save more with myopia than with perfect foresight Given the
exi~tence of taxes on capita and the discrepancy between the private an~--sccial returns to capital
the greater savings with myopia is beller socially Ballard-Goulder find that the welfare gains
from a consumption tax are reduced by about 10 when we move from myopia to farleet foresight
Therefore the attractivenness of adopting a consumption tax seems fairly [obust across these
foresight specifications
I
bull
36
6 Tbe power and weaknesses 01 dynamic modelling the example 01 corporale
tax Integrallon
The Issue of corporate tax integration allows us to sh~w the stre-ngths and weaknesses or the
dynamic approach Empirical evidence using CGE techniques indicates that depending on the precise
scheme of integration and the 1ax replacement methods integratiOn may have substantial effects In
Ihe work of FuliertonmiddotKingmiddotShoven-Whalley (1980 1981) total integration was found to yield an bull
annual static eHielency gain of $4 to $8 billion in 1973 dollars Simulated dynamic gains may be as
large as $695 billion or abcu 14 of the present value of tulUre consumption and leisure in the
US economy These gains result primarily from interindustry reallocalions of Investment and an
improved inter~emporal allocation of consumption
Mora recent work emphasizes hOw consumers asset portfolio decisions and firms finanCial
decisions affect the efficiency gains from integration Slemrod (1980) focusing on consumers
asset ponfoio deCisions finds static efficiency gains which are about twice -as Iarge as those
repOrted iiy Fulierton-King-ShovenWhalley (1961) FuUrlon-Gurdon (1983) focus on the
firms financial decisions They roporl efficiency gains of 6 of GNP from the elimination of the
tax distor1ions favoring debt However when they eliminate the corporate tax and repiace it with
increaSed persona income taxes addlHonal dls10t1ions are created in the optimal labot~leisure
decisions These distortions tend to dominate the analysis slJ(h that the overall effects of complete
integration are very modest Galpr-LuckmiddotTodr (1986) address simultaneously consumers
asset pOrlfollo decisions and the firms financial decisions They also find very modest efficiency
gains from integralion
The above results are important but may be severely biased There are severa aspects of
economic behavior and modelling crucial for the study of Income tax Integration whiCh have no been
i 37
captured innny of the above papers One aspect is the absence of government deficitsmiddot a balanced
budget is assumed It is true Ihal resource crowding OUI is caplured In these mOdels but financial
crowding out induced by government spending is nOI Second investment is not derived from
optimtztlcn behavior Investment behavior passively accomodates endogenous saving decisions As
a consequence the differential impacts of policies in the incentives to save and to invest are not
captured Aso full capital mobility across secors is assumed with instantaneous capital
adjustmenlS towards optimal levels This assumption rules qut different costs of capital across
seclors and therefore differentiated reactions to tax policies changes Finallyhe modelling of both
gov9rnmerlt deficits and of endogenous real and financial investment decisions necessitates the
~nsideraiion of a dynamilt framework and the introduction of financiaJ assets govemment bonds and
iprivate financial assets A dynamic framework also highlights the efficiency effects of Integration on r
ithe ptimal intertemporal decisions The above models are either sIalic (Slemrods and
GalpermiddotltJckemiddotTode(sj or a dynamic sequ~nce of otherwise Slatic model bullbull I Pereir a (1986b) develops a dynamic applied general equilibrium model bull to study the tbull
efficiency effects of integration on the growth and allocation of investmenl across sectors Special
Iattention is paid to the structural affeelS of resource and financial crowding oul induced by
gove mect spending and deficits and to the real and financial Investment reactions across industries
to both tax integration and flnancial crowding out His mode departs from previous work on income
lax Integraticn and for that matler from most 01 the CGIE literalure for tax policy evaluation in
several direcllons It encompasses an endogenous sequential equilibrium struelure founded on
dynamic behavior with flexible expectations Government deficits are optlmaUy determined
Investment decisions are tward looking and the resull of optimizing behavior Several financial
assetsmiddot public and privatemiddot are considered
Simulation resulls Indicale that the welfare gains from Integration under large deficits are at
32
best lJlodest when measured in terms of the GNP The elimination of the corporate tax and its I replacemenl by increased income lax rates yields long run benefits which are never larger than
2 of the presenl yalue ct telr consumption and leisure onder the previous tax regime and 1
under the current tax regime However the short run w~lare effects of full integration tend 10 be
very low and in in some scenarios even negative This is a flew intertemporar pattern of efficiency Ieffects which reflects an adjJsiment lag in the interindustry investment decisions due to the
existence of costs of adjustment On the of her hand partial integration achieved by excluding I dividends from the corporale lax base syslematically yields negalive efficlencyeffeclS In Ihe long Inm these can be as high as 3 of present value of the future value of consumption and leisure This
is a new second best effect suggesllng Ihat less than complete integration may have perverse I efficiency effects
The dynamic slructure of PerclrBs model snowing for an improved freatment of real investment
declsions and government deficts provides a potential setting for a more accurate measure of the
costs and benefits of inlegration The simulation resullS suggest lower benefits and an intertempora Ibull
pattern of growing benefits
Simulation results also clearly suggest robust negative effects from partial integration How
reliable is this result If dJvidends were deductible from the corporate lax base (as are interest
payments) the preferenlial Ireatmant of debl over equily would be ellminaled Corporallons should
be expecled 10 react by decreasing the optimal deWeqully rallo However corporale financial
gecisions are exogenously set In Pereiras model as in aft of the other models surveyed that go as far
as dealing with the issue
Also withdrawing dividends from Ihe corporate tax base is a way of eliminating Ihe double
taxation of dividends at belh Ihe corporate and personnal Income levels How will the optimal
aliocellon of household saving accomodate that change in Ihe relalive return to corporale equity
39 I
t However hOusehold portfolio decisions are exogenouly set in Pereiras model as in all of the other
models surveyeltj go as far as dealing with the issue
IIt is sate to say that the evaluation 01 the benefits 01 partial Integration as defined above are
sevrey underestima1eO Meanwhile the distortionSoenerated by the increase in the marginal
personal tax rates designed to ralse the revenue foregone by the dividend exclusion are fully
aCCltlunted for A good measure of the costs of integration together With a poor evaluation of the
benefils may very well be the reason why partial integration is Simulated 10 yield neg alive
eHikiency gains
On a different vein as most of Ihe model surveyed here Pereiras is a closed economy model It
is legitimate to wonder how the resullS would change If international capital flows were anowed
IThis almost certainly would affect financial crowding out since a good d~al of the capital inflows
could be used to finance public debt
For these various reasons the results should be treated as preliminary but they strongly I suggosl that the dynamic structure provides valuable new insights into the corporate tax integration SSU2
40
7 Concluding remarks I I I i
What has been acccmplished Ihis far The success of Ihe CGE research in laelding the challenge
of dynamic modelling can not be denied
I iGreat progress in the modelling of household behavior has been achieved Promising
developm~nts in the modeUing of production and government behavior have also been accomplished
Interesting innovations in the concept of equilibrium and computation techniques were discussed
The policy analysis was greatly enriched by considering transitional effects together with longer
run steadymiddotstate equilibrium paths generalized equal yield strategies accomodating different
financial crowding out impacts and generalized dynamic policy evaluation indicators
An important set of issues have been addresed by the models surveyed Traditional issues
focusing on the effects of capital taxation and consumption taxes have been pursued In terms of
capital taxation for example the intertemporal nature of the issue was enhanced by allowing an
optimal evolution of capital stock in the economy and forward looking optimal investment decisions
In Andersson (1987) GouldermiddotSummers (1987) and Pereira (1986b 1987d) this is coupled
with a improved treatment of several tax provisions like investment tax credits and depreciation
allowances In terms of the consumption taxes developments in the specification of Intertemporal
household behavior the introduction of overlapping generations and the modelling of
intergenerational links via bequest motives as well as a closer attention to demographic evolution
provided a much improved economic setting for the understanding of the several aspects of the
problem
Furthermore dynamic modelling has permitted the addressing of a set of new issues Policy
Issues ranging from dynamic tax policy analysis to the evaluation of exogenous changes in
government expenditures to the impact of different methods of financing government expenditure to
4 1
the importance of tinanciai croHrii1g out to 1he relevance of consumorS expectations in tne policy
results have been addressed
Despile all of the progress and in part due to such progress several avenues are wide open for
much neded additional research The following seem to be the most Interesting and promising
areasmiddot First the specincation of liquidity constraints to household behavior may help to obtain a
better understanding of policy chenges that affect directly the interest rates in the economy Second
major attention needs to be paid to the incorporafion of a roraign sector (with open capital markets)
Third some efforts are needed in terms of finding adequate computation techniques now that
dynamic modelling eally makes the curse of dimensionality worse than eVOrbull
The modelling of financial markets is the single most unsatisfactory speet of the dynamic
models surveyed here~ The problems associated with provlding a mo~ adequate treatment of
financial markets and in general endogenous and optimal financial deci$ions bull both at the corporate
and at the personal levelsmiddot have 10 be addressed That necessarily requires Introducing uncertainty
into the models To b adequate this must go well beyond the mere assumption of point expeetalions
with the whole array of modelling implementalion and evaluation problems thai generates
bull I i
42
REFERENCES
Abel A and O Blanchard 1983 An Intertemporal Model of Saving and Interest Econometrica
51 pp 675-92
~ Andersson K 1987 Tar-ation of Capital in Sweden - A General Equilibrium Model Ph D
Dissenation UnIversity of Lund Sweder
Auerbach A L KotlikoH 1983 National Savings Economic Welfare and the Structure of
Taxation in M Feldstein ed Behavioral Simulation Methods in Tax Policy- Analysis Chicago
University of Chicago Press
Auerbach A l Kotlikoff 1984 Social Security and the Economics of Demographic-
Transition in H Aaron and G Burtless eds Retirement and Economic Behavior Washington DC
Brookings Institution
Auerbach A L Kotlikoff 1987 Dynamic Fiscal Policy Cambridge Cambridge University
Press
Auerbach A L Kotlikof and J Skinner 1983 The Efficiency Gains from Dynamic Tax
Reform Internatiooaj Economic Reyiew 24 pp 81middot100
Ballard C 1983 Evaluation of the Consumption Tax with Dynamic General Equilibrium
Models Ph D Dissertation Stanford University
Ballard C 1987 Tax Policy and Consumer Foresight A General Equilibrium Simulation
Study Economic Inquiry 25 pp 267-284
Ballard C and L Goulder 1985 Consumption Taxes ForeSight and Welfare a Computatonal
General Equilibrium Analysis in J Piggott and J Whalley eds New Deyelooments in Apolied
General Equilibrium Analysis Cambridge Cambridge University Press
Ballard C D Fullerton J Shoven and J Whalley 1985 A General Equilibrium Model for
t 43
Tax Eoliey Eyalua~iQO NBER University or Chicago Press
Borges A (1987) bull A Survey of Compula1ional General Equilibnum Models CECD Economic
Slud~ 8
Bovenbig L 19B3 imperfect Capital Mobility in a Perfect Foresight Herberger Model
Mimea UC Br~eley
Bovenberg L 1984 Capitel Accumulation and Capital Immobility Qmiddottheory In a Dynamic General Equilibrium Framework PhD dissertation University of California at Berkeley t
Bovenberg L 19850 Dynamic General Equilibrium Tax Models with Adjustment Costs in A iManne ed bull 1985 EtQnomrc ECUjilbrium Model Formulation and Solution Mathematical
lProgramming Siudy 23 Arnsterdam NorthmiddotHoliand
Bovenberg L t 1986 Capital Income Taxation in Growing Open Eeondmies Journa of Pubtic
Eamp(omlcs 3t pp 347middot376 IBovenberg L ald W Keller 1981 Dynamics and Nonlinearlties in Applied General
Equilibrium Models Internal Report Mlmiddot81middot208 Department for Statistical Methods CBS bull
Vooiburg The Netherlands
Charnley C 1981 The Welfare Cost of C~pltal income Taxation In a Growing Economy
IJournal of Political ECQnQm~ 89 3 pp 468middot496 I
Charnley C 19810 Efficient Tax Reform in a Dynamic Model of General Equilibrium I Mimeo Yale University I Charnley C 1982 On the Optimal Taxation In Stylized Models 01 Inlartemporal General
Equilibrium Mimeo Yale University I Decaluw B and A Manens 1985 Pays en Oeveloppement at Modeles Calculables OEquilibre
General - Une Revue de al litterature empirique Monographie Sene 9 Volume 3 Centre de
Recherche et Developpemont Economlque Unrslt de Montreal
44
bull
I I
I
I
1
Erlich $ V Ginsburgh and L van der Heyden 1987 Wher~ do Real Wage Policies Lead
Belgium bull A General Equilibrium Analysis European Economic Reyiew (forthcoming)
air R and J Taylor 1983 Solution and Maximum Likelihood Estimation of Dynamic
Nonlinear Ratioant Expeclations Models Ecooometrica 51 pp 1169middot1186
Feltenstein A 1984 Money and Bonds In a Disaggregated Ec~nonYin Searl 1-1 J Shoven
Eds Aoolied ~eoeral EqJiliDCiulD tioalYsis Cambridge University Press Cambridge
Feltenstein A 19B6 A Dynamic General Equllibrlu Analysis of Financial Crowding Out
Theory with an Application to Australia Journal of Public Economics 31 No1
Fullerlon D R Gorden 1983 A Reexamination of Tax Distorllons in General Equilibrium
Models in M Feldstein Ed Behayora Simulation Methods in Tax pOlicy Analysis Chicago
University of Chipgo Press
Fvllerlon D J Shaven and J Whalley 1983 Replacing the US Income Tax with a
Progressive Consumption Tax JoulOal of Public ECQnomics 20 pp3middot23
Fullerton D Y Henderson J Shaven 1984 A COmparision of Methodologies in Empirical
General Equilibrium Models of Taxation in Searl H J Shoven Eds ~pQlied Geoaral Jguilibrium
lIoalyss Cambridge University Press Cambridge
Fullerton 0 A King J Shaven and J Whaney 1981 Corporate Tax Integration in the United
Statesmiddot a General Equiiibdufi1 Approach American Economic Review 71 pp~ 677~691
Galper H R Lucke E Tader 1985 Taxation portlor Choice and the Allocation of Capital A
General Equilibrium Approach Mimeo
Gill p W Murray M Saunders M Wright 1986 Users Guide for NPSOL (Version 40) A
Fortran Package for Nonlinear Programming T echnicsl Report SOL 86middot2 Deparlment of
Operations Research Stanford University
Glnsbur~h V and L van der Heyden 1985 General Equilibrium with Wage Rigidities An
45
AppJication to Belgium in A Manne ed t Economic EQumprit1m~ Made Formulation and $oJioo
Mathematical Programming Study 23 Amsterdam NorthmiddotHolland ~
Goulet J 1968 1 Adjustrent CO$1S in the Theory o(the Firm Reyiew of EconomiC Studies 35 bull
pp47middot55
Goulder L 1985 Tax-Financed versus BondmiddotFinancedGhanges in Governent Spending
bull Mimeo Harva(d UniversilYmiddot
Goulder l J Shaven and J Whalley 1983 Domestic Tax Policy and the Foreign Sector The
importance of altemative foreign policy formulations to results from a general equilibrium tax
analysis model In ~n methods in Ibe antico of income from caoia1 ed Martin
Feldstein Chicago University of Chicago riSS
Goulder l L Summers 1987 Tax Policy Asset Prices and 3rowtly ~
A General Equilibrium Analysis NBER Working Pper No 212B
Grandmont J 1982 Temporary General Equilibrium Theory Ch 20 in Handbook 01
Mathemalical Eccnornics Vol II NorthmiddotHoliand
Harberger A 1959 The Corporate Income Tax An empirical Appraisal In lax Revision
Comoendjum Voll House Cornmitee on Ways and Means Washington DC Govemment Printing
Office
Harberger A 1962 The Incidence of the Corporate Income Tax JOULn1 of Political
Economy 70 pp 215middot240
Harberger A 1964 Taxation Resource Aliocalrlon and Welfare fn lb Role of Dir bullbulll and
Indjresl Taxes io tbe Fede[al Reserve s~I~Ill Princeton Princeton Unlveresity Press
James J 1985 New Developments in the Applications of General Equilibrium models to
Economic History in Piggott aand Whalley Eds
Jorgenson D and K Yun 1984 Tax Policy and Capital Allocation Harvard Inslitute of
bull
I
I f
Economic Research WP 1107
Judd K 1965 ShonmiddotRun Analysis of Fiscal Policy in a Simple Perfect Foresight Model
Journal of poJiHcal 5CCJiW1l pp 298middot319 r
Upton D J Potbc J Sachsa nd L Summers 1982 Multiple Shooting in Rational
Expectations Models fuoometrlc 50 1329middot1333
Manne A ed 1985 fconQrnjc Eoullibrium Model FOrIDujaljQo and Solution Mathematical
Pogrammlng Study 23 Amsterdam NorthmiddotHolland
Manne A 1985 Or the Formulation and Solution 01 Economic Equilibrium Models In A
Manne ed ECQOQlli-Jr~ibmiddotum ~tQdel EQlJulailon and Solution Malhematical Programming
Study 23 Amsterdam NorthmiddotHolland
Mansur A and J Whalley 1984 Numerical Specification of Applied General Equilibrium
Models in Scarf H J Shoven Eds Aoplied General Eouillbrium AnaiysectIsect Cambridge Universily
Press Cambridge
Merrill 0 1972 Applications and Extensions of an Algorithm thaI Computes FixedmiddotPoints of
Cenain Upper Semimiddotcominuou Point to Set Mappings PhD Disserlation Unlvel$ity 01 Michigan
Negishi T 1960 Welfar Economics and Exislence I an Equnlbrlum for a Competitive
Economy MelroeCOOQ11ka 12 92middot97
Pereira A 1985 On the Existence and Uniqueness of Optimal Ou1put Path lor CRTS Firms in
Adjustment Costs Technologies Mimeo Stanford University
Pereira A 1985a Consumer Expectations and Ihe COSI 01 UllUty In an Intertemporat
Framework Mlmeo Stanford University
Pereira A 1985b Corporale Tax Integrallon and Government Deficits A Dynamic General
Equilibrium Anaiysls Mlmeo Stanford University
Pereir A 1987 On Ihe Computation of the Users Costs of Stocks Ecooomi (forthcoming)
47
Pereira A bull 1987b Equal Yield Alternatives and Government Oeficlts Mimeof Stanford
University
Pereira A 1987c A Dynamic Applied General ~uHlbrium Model for Tax Policy Evaluation
Comlletamp Docurnent2tion Mimeo University of California San Diego (in progress)
Pereira A 1987d Should the investment Tax Credit be Rei(ltroduced Mlmeo University
of California San Diego (ir progreSs)
Piggott J and J Whalley eds 1985 tJew QeyeQpments in Applied General EQuilibrium
bullenalysjs Cambridge Cambridge University Press
Preckel Pbull 1985 Alternative algorithms for computing economic equilibria In A Manne Ed
Radnor R 19amp3 Equllibrium under Uncertainty In K Arrow M Intrilligaor ods
I ibull
Robinson S 1 g86 Mullisectorai Models of Oeveloping CountrieS a Survey Oivision of
Agricullure and Resources UC Borkeley WP 401
Scarf H 1967 The approximation of fixed points of a continuous mapping ~MM Jouwal of
epplid Mathornallcs 15 pp 1328middot43
Scarf H with T Hansen 1973 00 the Computation of EconomiC Eguilibria New Haven Yale
University Press
Scarf H J Shoven Eds 1984 61l1llied Genera Elujlibrium 6nalyssect Cambridge University
~ Press Cambridge
Shoven J 1976 The Incidence and Efficiency Effects of Taxe on Income from Capital
Jurnal of ramie1 Economy 84 pp 1261-83
Shoven J 1983 Applied General Equilibrium Tax Modemng IMF Slaff PaoerS 30
Shoven J and LB Simon 1986 Share Repurchases and Acquisitions An Analysis of which
Firms Participate NBER Working Paper No 2243
bull
48
IShaven J and J Whalley 1972 ft General Equmbrium Calculation of lhe Effects of
Differential Taxation of Incoma from Capital in the US Jauroe of public Econgmics 1
pp281middot321 (
Shoven J bull and J Whalley 1973 General Equilibrium with Tax$s a computation procedure
and an existence pr00f ReviEhi of Economic Studies 40 pp 475middot490
Shoven J J Whalley 1977 Equal Yield Tax Alternatives A General Equilibrium
Computa~ui0nal Tfchnique J~HJrnal of PUblic EconomIcs 8 pp 211~224 bull
Shaven J J Whalley 1984 Applied GeneralmiddotEquilibrlum MOdels of Taxalion and
International Tr8de An Introduction and Survey Journal of Economic literature 23t pp
1007middot1051
Slemrod J 1980 A General Equilibrium Model of Capital Income Taxation Ph D bullDissertation Harvard Univnrsity
Slemrod J 1983 A General Equilibrium Model of Taxation with Endogenous Financial
Behavior in M Feldstein edbull Behavioral Simulation Methods In Tax Policy Analysis Chicago
UniVerSity of Chicago Press
Stone J 1985 Sequential Optimization and Complementarily Techniques for computing
Economic equilibrium in A Manne Ed
Summers L 1981 Taxation and Corporate Investment a q-Theory Approach BroQkinas
Summers L 1985 Taxation and the Size and ComposHion 01 the Capital Stock An Asset Price
Approach Mlmeo Harvard University
Whalley J 1975 A Geneal Equiloibrium Assessment of the 1973 United Kingdom tax
reform Economica 42 pp 139middot61
1JLTIllOS JIQll~)ltG PAPERS PUBLICADOS bull
I
I I
j
112 70 - LUCENA iJiCIIJo ~ IIEnvironrilent Honitoring and Orqanization Structure T (Setellibro 1987)
nI 71 - L~ENA Diogo A Note on the Representation of Information Stxueture (SeteJObro 1987)
nQ 72 - LUCENA D1CgO Envuorment 111)nitoring and Organization
Structure II (Setelilhro 1987)
ril 73 - AlmmE3 Antonio Pais e CAsPAR Yitor bull Tributacao Inoentivos e InvesUlllentos Analise Qualitativa noYellibro 19S7)
n2 74 - liAlltOSA rose Pedro Opt1lllal lIage R1gidity A SUggestedtlthodology to Teet tb Theory and an Application (Outubro 1987)
nI 75 - CABRAL L 11 B Three lIotes on SYJIIlletric Gemes witt itsYJIIlletric Equilibri (NoveJObro 1987)
n2 76 - LUCENlt D10gOTo Search or Not to Search (Fevereiro1967)
nQ 17 - SA Jorge asconcelos euroI A Theory of SynergyB (Fevereiro1987)
n2 76 - VILARES tanuei Jose Os Be IntenaMios IlIIportados ColIIO Factor de Produao (Julho 1987)
nQ 79 - SA J~rge VasCQncel03 e HoY to Copete and Co~unicate in tature Industrial Products_ (aio 1988)
n2 80 - ROIl Rafael Learning and Capacity Expansion in a New larket Under Uncertainty (Fevereiro 1988)
n2 81 PEREIRA Alfredo llarvilo Survey of Dynamic Computational General EquilibriUlll Hodel for Tax Policy Evaluation (Outubro 1987)
Qualquer in ormacao sabre o 1IIorlltlng Papers ja publlcados sera prestada pelo Secretariado de Apoio aos Docentes podendo os ~e3OSf
ser adquirido5 ne Secyo de Vlndas da Faculdade de EconollLia UNL na Travessa Esteviio Pinto Cltpolide - 1000 LISllOA
26
j
I
4 Jrnplemenlalion and Issues on policy evaluallon
41 Calibration and equilibrium comparlslon
CGE models at typically parameterized by the use of a calibration procedure Some parameters
are exogenously given However some crucial parameters are determined in such a way that the
model replicates the data for a given base year See MansurmiddotWhalleY 11984) for an eXlensive
discussion of 1~i$ issue
CalibratIon in a dynamic context is generaliy interpreted as requiring two properties First
replication of base year data is required Second the model is parameterized to simulate an
intertemporal balanced growth path when Ihe base policy is maintaineq This Is the approach
followed by Ballard (1983) BallardmiddotGoulder (1985) Goulder (1985) and SummersmiddotGoulde
(1986) It follows the practice of BallardmiddotFuliertonmiddotShovenmiddotWhalley (1985) with their
sequential cquilibrium dynamics
Other authors Auerbach-Kotlikof (1983 1984 1987) Boyenberg (1984 1985 1986)
and Pereira (198Gb 1987d) follow whal we cali a qualitalive calibration The slructural
palametars are exogenously chosen so that lh economy follows a reasonable path into Ihe fulure
ThaI has 10 do wllh the fact thaI given the recursive nature 01 the dynamic economy and aside tom
such structural parameters only iniUa stock values are needed to run these models Given initial
conditions on the stocks of say private wealth capltal and government debt agents will optimize
and Ihereby generate a lirst round of net demands and equilibrium conditions In lurn the
equilibrium prices will determine the evolution of the stock variables into the next perIod
There are several potential problems with calibration in the conlext of dynamic models The
assumption of a steadymiddotstate growth ~ath in the base case can be questioned First while
27 I Ibullbull
steadymiddotstaU is a possibity it certainly is not the only meaningfull solution to dynamic models
Even in the case of a perfect foresight equilibrium the model implies an equilibrium path which i may or may not involve colanced growth The model not the modeller should dictate the nature 01 I themiddotbase case path Second in the context of a temporarY equilibrium path a steady state solutlcn is Inol a likely model QutcOn1tl In fact unlessmiddot expectations are stattc short run behavior consistent
with a steady-state evolution will in general not be generated On the other hand if static I expectations are self~fufjili[1g we have in fact a perfect foresight model Third even the base year
replication requirement may cause problems In Ihe cOntexl of temporary equilibrium Any
calibration parameter would be conditional on expectation rules which is probably an undesirable
feature
The nalur of the two-requirement calibration strategy is very muchin the spirit of the bull
traditional design of the comparision of alternative equilibria comparis_ion between a steady-state
base case on one hand and alternative paths including a transition period and a finat sleadyyenstate on
the other hand Pereira (198Gb 1987d) compare different (not necessarily steadystate)
equilibrium paths The arguments against such a procedure are based on the idea that the impact of
policy changes can be observed most easily since all departures from the steadymiddotstate can be
attributed to the altemalive polioy On the other hand the base case so dEfined as a steadymiddotstate is
consist~nl with previous work in a less dynamtc setting and Iherefore allows a common standard
for comparing model results
42 Computation algorithms
We are still al a stage in which basically each author uses a different computational technique
Th development of dynamic models - in parlicular wilh adjustment costs andior perfect foresight
bull
28
has corresponded with the decline in the use of fixed point algorithms In -fact given the relative
arge dimensions inevitably involved such algorithms lend to be very inefficient at the best and
chen prohibitively stow See Stone (1985) and Preckel (1965) for a comparative assessment of
different computation techniques Among the models surveyed only Ball1rdFulle(ton~
-ShavenmiddotWhalley (1985) and Feltensteln 1985 use Merrills variant of the fixed point
algorithm 1echnique
AuerbachmiddotKoHikoff (1963 1984 1967) follow a three stage procedure They first compute a
base case steady-state then a revised case steady state and finally transition path for the economy
berveen these two steady-states tn all stages a Gauss-Seidel iterative procedure is used
Ballard (1982) Ballard-Goulder (1965) Goulder (HiSS) and Summers-Goulder (1986)
~
use a method developed by Ballard and Goutder which is similar in many awects to the FairmiddotTaylQr bull
(1983) algorl1hm Short-run equilibria are calculated (using Merrills algorithm) parametric on
nrice expeC~lions The model is then iterated to generate self-fulfilling intertemporaf expectations
and the corresponding perfect foresight equilibrium In a relatively similar approach Andersson
1987 uses a simulation program SIMNON developed In the University of Lund This program
can handle two-paint boundary problems In a fashion consistent with the multiple shooting
algcrlthm (see Lipton-Poterba-Sachs-Eummers (1982))
Boyenbergs computational approach (1985 1985) differs from the other models surveyed in
that he relies heavily on analytical techniques Computations are done by using a dynamic version of
Johansons linearization method Being essentially determined by the continuous time nature of the
I I
f
model thiS linearization model has the disadvantage of confining the analysls to infinltesinal changes
around the base case equilibrium (see Bovenberg (1985) p 53) Pereira (1986b 1987d) uses an optimization algorithm NPSOL - developped by
rGillMurray-Saunders-Wrlght (1986)) This algorithm is used 10 compute the sequence of
I
29
Shortmiddotrun temporary equilibrium which makes up the intertemporal equilibrium path The
equilibrium conditions are seen as nonlinear equality constraints in the minimization of an artlcial
objective function The prices are normalized to the unit simplex by an additional linear equality I I 1 bull constraint
II ErlichmiddotGinsburgh-Heyden (1987) follow a unique approach in that they use a variant of the
optimization technique introduced by Negishl (1960) The economic equilibrium can be generated i
as a solution of a mathemallcal program Ihe objective function of which is a weighted sum of the
utility functions of the various agents while tho constraints set consists of the market dearing 1
conditions Ginsburgh-Heyden (1985) have extended Negishis resulllO tho case of downward price
lrigidities I
Finally the paper by Jorgonson-Yun (1984) is also unique in tIlat it is the only I
econometrically estimated model Different blocks for the consumption and production Side of the
-
model are separately estimated to provide the necessary structural parameters i
I The diversity of computation techniques is yet another indicator of the exploratory nature of the
body of literature surveyed in this article
43 Equal yield comparisions
The link between a base case and counteriactual simulations Is usually provided by the concept of
equal yield Shaven-Whaley (1977) discuss the meaning 01 equal yield in a general equilibrium
context when government is not allowed to run defICits Equal yield is interpreted to moan constant
public utility Government base case utility is maintained in the counferfactual experiments With
balanced budgets the concepl of equal yield is unambiguous The new equUibrium prices and the
balanced budge condition will del ermine the minimal expendilure and taxes needed to maintain base
30
case publi~ utility Accordingly in general equal yield is inconsistent with equal nominal tax
revenue Some change in tax revenue Is necessary Different tax replacement schemes ate
considered to assure that enough tax revenue is collected This is the approach essentially followed
by most 01 the papers surveyed Only Feltenstein (1985) Goulder (198) and Jorgenson-Yun
(1984) chose not to follow an equal yield strategy
The question is of how to Interprete lila concept of equal yield when the government is alowed to
fUn deflcLts The optimal leve of expenditure for base ease public trliUty can now be taAinanced
bond financed or financed by a mix of bonds and taxation We have several versions of equal yield In
particular equal yield may now be consistent with equal tax revenUE Furthermore some meaSure
ltgtf financial crowding out effects of government deficns can be inferred from the comparislon of the
several equal yield alternatives These issues are extensively discussed ~Pereira (1987b)
44 Price expectations and the dynamic generalization of compensation
Indicators
The sum of equivalent and compensation variations over households is the most widely used
aggregated measure of efficiency gains or losses In the context of $teady~state eomparisions andor
when complete future markets exist andlor perfect foresighl is assumed there are no difficulties
associated with the use of the standard Hicksian indicators In fact correct future prices are known
in these cases
If expectations are not selfmiddot fulfilling andlor future markels are not available a dynamic
generalization is necessary 8allardmiddotGoulder (1985) provide some steps in that direction by
defining an indicator that accomodates periods far into the future when households do not have
perfect foresight but a steady~state prevails On he other hand this issue is more fundamental in
i
[
I
i I I I I [
I I
bullI
31
the contex~ of a Gner1 lemporary equilibrium ftamawor~ In such Circumstances Pereira
(1986a) dev~lops a cnamlc generalization which is obtained as the present discounted value of a i i
sequence of short-fUn optimal expenditure functions consistent with a base case expected future I stream of IlWities I
32
5 Empirical evidenee from selected policy Issues
Dynamic tax models have generated several important results which escaped Sialic modellers
The following is a seleCpoundd set of issuos whIch stress the marginal benefits of dynam~ modelHng of
economic behavior in innovltlttive areas of analysis The discussion of a consumption tax emphasizes
the benefits of dynamic tusehold behavior and intergenerational aspects The study of the
flliminatlon or the re~intToduction of the investsnemt tax credits is made meaningful by the dynamic
modelling of production betlavor Modelling governmenl in a dynamic conlexts allows the modeller
to study the irrrpact of government deflcits Finally a dynamic framework lets us appreCiate the I
relevance of consumer expectations in terms of the evaluation of policy alternatives
i
i
51 Consumption Tax
t
By Its very nature a consumption tax can not be adequately investigated with a static model bull
Fullerion-Shoven-Whaicy (1983) use Ihe model of the US economy s described in
Ballard-Fullerton-Shoven-Whalley (1935) to evaluate the movement from the currenl US tax Isys1em to a progressive consumption tax Since their modeJ Incorporates a Jaborleisure choice
where leisure is an untaxed commodlty their results reflect the fact that both the consumption tax
and Ihe present system are dslortionary
Concentrating on the rntertemporaf distortions they show that sheltering more saving from lhe
current US income tax could improve econornic efficiency even if marginal tax rate increases are
necessary in order to maintain government fevenue~ At first you have a revenue shortfall
However the economy moves to a higher sustatned growth path and uitimatefy lower tax rates
cangenerate the same revenue path Also wages increase in the long run with this policy The
bull
bullbull ~
33 shyt
i
I
I
present value of welfare gi[lS for a potiey of complete savin9 deduction wilh marginal tate
adjustments (consumption lax) is simulated to be around $500 billion to $600 billion 01 1973
dollars
They also investiga1e th 1ngth of ttme it takes th~ economy to adjust to these policy changes
Roughly they estimate the long run to b thirty years although this figure Is very sensitive to the
specification of the savings elasticity - a crucial parameler in this model
52 Investment T~x Credit
Goulder-$ummers (1 S87) address the impact of eliminating the lnvestmenl Tax Cradit (ITC) on
intersectoral capital formation and on economic growth Their model isYa1ticularly adequate to
addies$ this issu~ in that is postulates a forward looking investment m9we with adjustment costs
They show that the eliminating the ITC causes a reduction in the rale of investment Inveslment is
estimated to fall by about 7 in the short-run and by about 12 in Ihe new long run steady state
In lurn a previous policy anncuncement lowers Ihe overall al1ractiveness of Investment and leads 10
a downward shift of the Investrrent profile On the other hand the combined effect of a revenue
neutral simultaneous efimiraron of the ITC and reduction of the corporate tax rates is a long run
reduction ef the capital Sieck by 35 This pattern suggests that a revenue neutral increament in
both Ihe corporale tax rates and the inveSlment lax credits would be preferable 10 the formula
adopted in Ihe US Tax Reform Act of 1986
Pereira (1987d) addresses the impact of the ITe from Ihe oland point of Ihe curTenllax syslem
under the Tax Reform Act of 1986 Given Ihe concern over the potential depressiVe effects upon
savings and investment of the new tax system in conjunction with deficit reduC1ion mechanisms of
the GrammmiddotRudman type a pertinent question is should Ih lTC be e-introduced In the context
bull
34
-of his mod~1 of the US economy Pereira focuses on the trade-off between the distortion in the
intersectoral allocation of capital induced by the lTC the lower tax revenues it generates and
potential financial crowding-out effects on one hand and the positive effects of lowering the relative
price of new capital goods on the other hand Preliminary results confirm the qualitative results in
Goulder-Summers (1987) suggestting a welfare gain of about 2 of the present value of GNP in
the best scenario of absence of replacement taxes Furthermore preliminary results show an
important time pattern to welfare gains with the average benefits increasing the further you look
into the future This pattern is due to the presence of constraints to intersectoral mobility of capital
and inertia in the adjustment towards the optimal capital levels as reflected by the presence of
adjustment costs
I
I
bullAuerbach-Katlikaff (1gB) in the context of their intergenerational model of the US economy
consider the impact on savings capital formation and interest rates of deficit policies and balanced
budget increases in government consumption They conclude that deficit finance and government
consumption can significa-ntly crowd out capital formation and lower the welfare of future
generations However crowding out from deficit finance is a very slow process because it results
from increased government spending over potentially long horizons Also deficit policies may
substantially influence the longmiddotrun interest rates while leaving the short-run rates essentially
unaffected Finally and in opposition to the central role adjustment costs seem to play in both
53 Financial crowding outeff~cts of government deficits
Goulder-Summers and Pereiras models Auerbach-Kotlikoff suggest that the time path of interest
rates induced by a policy of deficit financing seems to be insensitive to the presence of adjustment
costs
bull
35
54 The role 01 consumer expectatlons
The expectations analysis has uncovered one of the benefits of dynamic modelling
Ballard-Goulder (1985) find that the appeal of adopting a consumption tax depends on the level of
foresight possessed by the consumers Furthermore additional foresight may be welfare worsening
This is a second best type of result This tends 10 occur under policies that lead to capital deepening
and declining rate of return to capital over time To the extent that consumers have more foresight
they will be beller equippod to anticipate the fall in the rental price of capital Consequently if a
saving incentive is enacted people save more with myopia than with perfect foresight Given the
exi~tence of taxes on capita and the discrepancy between the private an~--sccial returns to capital
the greater savings with myopia is beller socially Ballard-Goulder find that the welfare gains
from a consumption tax are reduced by about 10 when we move from myopia to farleet foresight
Therefore the attractivenness of adopting a consumption tax seems fairly [obust across these
foresight specifications
I
bull
36
6 Tbe power and weaknesses 01 dynamic modelling the example 01 corporale
tax Integrallon
The Issue of corporate tax integration allows us to sh~w the stre-ngths and weaknesses or the
dynamic approach Empirical evidence using CGE techniques indicates that depending on the precise
scheme of integration and the 1ax replacement methods integratiOn may have substantial effects In
Ihe work of FuliertonmiddotKingmiddotShoven-Whalley (1980 1981) total integration was found to yield an bull
annual static eHielency gain of $4 to $8 billion in 1973 dollars Simulated dynamic gains may be as
large as $695 billion or abcu 14 of the present value of tulUre consumption and leisure in the
US economy These gains result primarily from interindustry reallocalions of Investment and an
improved inter~emporal allocation of consumption
Mora recent work emphasizes hOw consumers asset portfolio decisions and firms finanCial
decisions affect the efficiency gains from integration Slemrod (1980) focusing on consumers
asset ponfoio deCisions finds static efficiency gains which are about twice -as Iarge as those
repOrted iiy Fulierton-King-ShovenWhalley (1961) FuUrlon-Gurdon (1983) focus on the
firms financial decisions They roporl efficiency gains of 6 of GNP from the elimination of the
tax distor1ions favoring debt However when they eliminate the corporate tax and repiace it with
increaSed persona income taxes addlHonal dls10t1ions are created in the optimal labot~leisure
decisions These distortions tend to dominate the analysis slJ(h that the overall effects of complete
integration are very modest Galpr-LuckmiddotTodr (1986) address simultaneously consumers
asset pOrlfollo decisions and the firms financial decisions They also find very modest efficiency
gains from integralion
The above results are important but may be severely biased There are severa aspects of
economic behavior and modelling crucial for the study of Income tax Integration whiCh have no been
i 37
captured innny of the above papers One aspect is the absence of government deficitsmiddot a balanced
budget is assumed It is true Ihal resource crowding OUI is caplured In these mOdels but financial
crowding out induced by government spending is nOI Second investment is not derived from
optimtztlcn behavior Investment behavior passively accomodates endogenous saving decisions As
a consequence the differential impacts of policies in the incentives to save and to invest are not
captured Aso full capital mobility across secors is assumed with instantaneous capital
adjustmenlS towards optimal levels This assumption rules qut different costs of capital across
seclors and therefore differentiated reactions to tax policies changes Finallyhe modelling of both
gov9rnmerlt deficits and of endogenous real and financial investment decisions necessitates the
~nsideraiion of a dynamilt framework and the introduction of financiaJ assets govemment bonds and
iprivate financial assets A dynamic framework also highlights the efficiency effects of Integration on r
ithe ptimal intertemporal decisions The above models are either sIalic (Slemrods and
GalpermiddotltJckemiddotTode(sj or a dynamic sequ~nce of otherwise Slatic model bullbull I Pereir a (1986b) develops a dynamic applied general equilibrium model bull to study the tbull
efficiency effects of integration on the growth and allocation of investmenl across sectors Special
Iattention is paid to the structural affeelS of resource and financial crowding oul induced by
gove mect spending and deficits and to the real and financial Investment reactions across industries
to both tax integration and flnancial crowding out His mode departs from previous work on income
lax Integraticn and for that matler from most 01 the CGIE literalure for tax policy evaluation in
several direcllons It encompasses an endogenous sequential equilibrium struelure founded on
dynamic behavior with flexible expectations Government deficits are optlmaUy determined
Investment decisions are tward looking and the resull of optimizing behavior Several financial
assetsmiddot public and privatemiddot are considered
Simulation resulls Indicale that the welfare gains from Integration under large deficits are at
32
best lJlodest when measured in terms of the GNP The elimination of the corporate tax and its I replacemenl by increased income lax rates yields long run benefits which are never larger than
2 of the presenl yalue ct telr consumption and leisure onder the previous tax regime and 1
under the current tax regime However the short run w~lare effects of full integration tend 10 be
very low and in in some scenarios even negative This is a flew intertemporar pattern of efficiency Ieffects which reflects an adjJsiment lag in the interindustry investment decisions due to the
existence of costs of adjustment On the of her hand partial integration achieved by excluding I dividends from the corporale lax base syslematically yields negalive efficlencyeffeclS In Ihe long Inm these can be as high as 3 of present value of the future value of consumption and leisure This
is a new second best effect suggesllng Ihat less than complete integration may have perverse I efficiency effects
The dynamic slructure of PerclrBs model snowing for an improved freatment of real investment
declsions and government deficts provides a potential setting for a more accurate measure of the
costs and benefits of inlegration The simulation resullS suggest lower benefits and an intertempora Ibull
pattern of growing benefits
Simulation results also clearly suggest robust negative effects from partial integration How
reliable is this result If dJvidends were deductible from the corporate lax base (as are interest
payments) the preferenlial Ireatmant of debl over equily would be ellminaled Corporallons should
be expecled 10 react by decreasing the optimal deWeqully rallo However corporale financial
gecisions are exogenously set In Pereiras model as in aft of the other models surveyed that go as far
as dealing with the issue
Also withdrawing dividends from Ihe corporate tax base is a way of eliminating Ihe double
taxation of dividends at belh Ihe corporate and personnal Income levels How will the optimal
aliocellon of household saving accomodate that change in Ihe relalive return to corporale equity
39 I
t However hOusehold portfolio decisions are exogenouly set in Pereiras model as in all of the other
models surveyeltj go as far as dealing with the issue
IIt is sate to say that the evaluation 01 the benefits 01 partial Integration as defined above are
sevrey underestima1eO Meanwhile the distortionSoenerated by the increase in the marginal
personal tax rates designed to ralse the revenue foregone by the dividend exclusion are fully
aCCltlunted for A good measure of the costs of integration together With a poor evaluation of the
benefils may very well be the reason why partial integration is Simulated 10 yield neg alive
eHikiency gains
On a different vein as most of Ihe model surveyed here Pereiras is a closed economy model It
is legitimate to wonder how the resullS would change If international capital flows were anowed
IThis almost certainly would affect financial crowding out since a good d~al of the capital inflows
could be used to finance public debt
For these various reasons the results should be treated as preliminary but they strongly I suggosl that the dynamic structure provides valuable new insights into the corporate tax integration SSU2
40
7 Concluding remarks I I I i
What has been acccmplished Ihis far The success of Ihe CGE research in laelding the challenge
of dynamic modelling can not be denied
I iGreat progress in the modelling of household behavior has been achieved Promising
developm~nts in the modeUing of production and government behavior have also been accomplished
Interesting innovations in the concept of equilibrium and computation techniques were discussed
The policy analysis was greatly enriched by considering transitional effects together with longer
run steadymiddotstate equilibrium paths generalized equal yield strategies accomodating different
financial crowding out impacts and generalized dynamic policy evaluation indicators
An important set of issues have been addresed by the models surveyed Traditional issues
focusing on the effects of capital taxation and consumption taxes have been pursued In terms of
capital taxation for example the intertemporal nature of the issue was enhanced by allowing an
optimal evolution of capital stock in the economy and forward looking optimal investment decisions
In Andersson (1987) GouldermiddotSummers (1987) and Pereira (1986b 1987d) this is coupled
with a improved treatment of several tax provisions like investment tax credits and depreciation
allowances In terms of the consumption taxes developments in the specification of Intertemporal
household behavior the introduction of overlapping generations and the modelling of
intergenerational links via bequest motives as well as a closer attention to demographic evolution
provided a much improved economic setting for the understanding of the several aspects of the
problem
Furthermore dynamic modelling has permitted the addressing of a set of new issues Policy
Issues ranging from dynamic tax policy analysis to the evaluation of exogenous changes in
government expenditures to the impact of different methods of financing government expenditure to
4 1
the importance of tinanciai croHrii1g out to 1he relevance of consumorS expectations in tne policy
results have been addressed
Despile all of the progress and in part due to such progress several avenues are wide open for
much neded additional research The following seem to be the most Interesting and promising
areasmiddot First the specincation of liquidity constraints to household behavior may help to obtain a
better understanding of policy chenges that affect directly the interest rates in the economy Second
major attention needs to be paid to the incorporafion of a roraign sector (with open capital markets)
Third some efforts are needed in terms of finding adequate computation techniques now that
dynamic modelling eally makes the curse of dimensionality worse than eVOrbull
The modelling of financial markets is the single most unsatisfactory speet of the dynamic
models surveyed here~ The problems associated with provlding a mo~ adequate treatment of
financial markets and in general endogenous and optimal financial deci$ions bull both at the corporate
and at the personal levelsmiddot have 10 be addressed That necessarily requires Introducing uncertainty
into the models To b adequate this must go well beyond the mere assumption of point expeetalions
with the whole array of modelling implementalion and evaluation problems thai generates
bull I i
42
REFERENCES
Abel A and O Blanchard 1983 An Intertemporal Model of Saving and Interest Econometrica
51 pp 675-92
~ Andersson K 1987 Tar-ation of Capital in Sweden - A General Equilibrium Model Ph D
Dissenation UnIversity of Lund Sweder
Auerbach A L KotlikoH 1983 National Savings Economic Welfare and the Structure of
Taxation in M Feldstein ed Behavioral Simulation Methods in Tax Policy- Analysis Chicago
University of Chicago Press
Auerbach A l Kotlikoff 1984 Social Security and the Economics of Demographic-
Transition in H Aaron and G Burtless eds Retirement and Economic Behavior Washington DC
Brookings Institution
Auerbach A L Kotlikoff 1987 Dynamic Fiscal Policy Cambridge Cambridge University
Press
Auerbach A L Kotlikof and J Skinner 1983 The Efficiency Gains from Dynamic Tax
Reform Internatiooaj Economic Reyiew 24 pp 81middot100
Ballard C 1983 Evaluation of the Consumption Tax with Dynamic General Equilibrium
Models Ph D Dissertation Stanford University
Ballard C 1987 Tax Policy and Consumer Foresight A General Equilibrium Simulation
Study Economic Inquiry 25 pp 267-284
Ballard C and L Goulder 1985 Consumption Taxes ForeSight and Welfare a Computatonal
General Equilibrium Analysis in J Piggott and J Whalley eds New Deyelooments in Apolied
General Equilibrium Analysis Cambridge Cambridge University Press
Ballard C D Fullerton J Shoven and J Whalley 1985 A General Equilibrium Model for
t 43
Tax Eoliey Eyalua~iQO NBER University or Chicago Press
Borges A (1987) bull A Survey of Compula1ional General Equilibnum Models CECD Economic
Slud~ 8
Bovenbig L 19B3 imperfect Capital Mobility in a Perfect Foresight Herberger Model
Mimea UC Br~eley
Bovenberg L 1984 Capitel Accumulation and Capital Immobility Qmiddottheory In a Dynamic General Equilibrium Framework PhD dissertation University of California at Berkeley t
Bovenberg L 19850 Dynamic General Equilibrium Tax Models with Adjustment Costs in A iManne ed bull 1985 EtQnomrc ECUjilbrium Model Formulation and Solution Mathematical
lProgramming Siudy 23 Arnsterdam NorthmiddotHoliand
Bovenberg L t 1986 Capital Income Taxation in Growing Open Eeondmies Journa of Pubtic
Eamp(omlcs 3t pp 347middot376 IBovenberg L ald W Keller 1981 Dynamics and Nonlinearlties in Applied General
Equilibrium Models Internal Report Mlmiddot81middot208 Department for Statistical Methods CBS bull
Vooiburg The Netherlands
Charnley C 1981 The Welfare Cost of C~pltal income Taxation In a Growing Economy
IJournal of Political ECQnQm~ 89 3 pp 468middot496 I
Charnley C 19810 Efficient Tax Reform in a Dynamic Model of General Equilibrium I Mimeo Yale University I Charnley C 1982 On the Optimal Taxation In Stylized Models 01 Inlartemporal General
Equilibrium Mimeo Yale University I Decaluw B and A Manens 1985 Pays en Oeveloppement at Modeles Calculables OEquilibre
General - Une Revue de al litterature empirique Monographie Sene 9 Volume 3 Centre de
Recherche et Developpemont Economlque Unrslt de Montreal
44
bull
I I
I
I
1
Erlich $ V Ginsburgh and L van der Heyden 1987 Wher~ do Real Wage Policies Lead
Belgium bull A General Equilibrium Analysis European Economic Reyiew (forthcoming)
air R and J Taylor 1983 Solution and Maximum Likelihood Estimation of Dynamic
Nonlinear Ratioant Expeclations Models Ecooometrica 51 pp 1169middot1186
Feltenstein A 1984 Money and Bonds In a Disaggregated Ec~nonYin Searl 1-1 J Shoven
Eds Aoolied ~eoeral EqJiliDCiulD tioalYsis Cambridge University Press Cambridge
Feltenstein A 19B6 A Dynamic General Equllibrlu Analysis of Financial Crowding Out
Theory with an Application to Australia Journal of Public Economics 31 No1
Fullerlon D R Gorden 1983 A Reexamination of Tax Distorllons in General Equilibrium
Models in M Feldstein Ed Behayora Simulation Methods in Tax pOlicy Analysis Chicago
University of Chipgo Press
Fvllerlon D J Shaven and J Whalley 1983 Replacing the US Income Tax with a
Progressive Consumption Tax JoulOal of Public ECQnomics 20 pp3middot23
Fullerton D Y Henderson J Shaven 1984 A COmparision of Methodologies in Empirical
General Equilibrium Models of Taxation in Searl H J Shoven Eds ~pQlied Geoaral Jguilibrium
lIoalyss Cambridge University Press Cambridge
Fullerton 0 A King J Shaven and J Whaney 1981 Corporate Tax Integration in the United
Statesmiddot a General Equiiibdufi1 Approach American Economic Review 71 pp~ 677~691
Galper H R Lucke E Tader 1985 Taxation portlor Choice and the Allocation of Capital A
General Equilibrium Approach Mimeo
Gill p W Murray M Saunders M Wright 1986 Users Guide for NPSOL (Version 40) A
Fortran Package for Nonlinear Programming T echnicsl Report SOL 86middot2 Deparlment of
Operations Research Stanford University
Glnsbur~h V and L van der Heyden 1985 General Equilibrium with Wage Rigidities An
45
AppJication to Belgium in A Manne ed t Economic EQumprit1m~ Made Formulation and $oJioo
Mathematical Programming Study 23 Amsterdam NorthmiddotHolland ~
Goulet J 1968 1 Adjustrent CO$1S in the Theory o(the Firm Reyiew of EconomiC Studies 35 bull
pp47middot55
Goulder L 1985 Tax-Financed versus BondmiddotFinancedGhanges in Governent Spending
bull Mimeo Harva(d UniversilYmiddot
Goulder l J Shaven and J Whalley 1983 Domestic Tax Policy and the Foreign Sector The
importance of altemative foreign policy formulations to results from a general equilibrium tax
analysis model In ~n methods in Ibe antico of income from caoia1 ed Martin
Feldstein Chicago University of Chicago riSS
Goulder l L Summers 1987 Tax Policy Asset Prices and 3rowtly ~
A General Equilibrium Analysis NBER Working Pper No 212B
Grandmont J 1982 Temporary General Equilibrium Theory Ch 20 in Handbook 01
Mathemalical Eccnornics Vol II NorthmiddotHoliand
Harberger A 1959 The Corporate Income Tax An empirical Appraisal In lax Revision
Comoendjum Voll House Cornmitee on Ways and Means Washington DC Govemment Printing
Office
Harberger A 1962 The Incidence of the Corporate Income Tax JOULn1 of Political
Economy 70 pp 215middot240
Harberger A 1964 Taxation Resource Aliocalrlon and Welfare fn lb Role of Dir bullbulll and
Indjresl Taxes io tbe Fede[al Reserve s~I~Ill Princeton Princeton Unlveresity Press
James J 1985 New Developments in the Applications of General Equilibrium models to
Economic History in Piggott aand Whalley Eds
Jorgenson D and K Yun 1984 Tax Policy and Capital Allocation Harvard Inslitute of
bull
I
I f
Economic Research WP 1107
Judd K 1965 ShonmiddotRun Analysis of Fiscal Policy in a Simple Perfect Foresight Model
Journal of poJiHcal 5CCJiW1l pp 298middot319 r
Upton D J Potbc J Sachsa nd L Summers 1982 Multiple Shooting in Rational
Expectations Models fuoometrlc 50 1329middot1333
Manne A ed 1985 fconQrnjc Eoullibrium Model FOrIDujaljQo and Solution Mathematical
Pogrammlng Study 23 Amsterdam NorthmiddotHolland
Manne A 1985 Or the Formulation and Solution 01 Economic Equilibrium Models In A
Manne ed ECQOQlli-Jr~ibmiddotum ~tQdel EQlJulailon and Solution Malhematical Programming
Study 23 Amsterdam NorthmiddotHolland
Mansur A and J Whalley 1984 Numerical Specification of Applied General Equilibrium
Models in Scarf H J Shoven Eds Aoplied General Eouillbrium AnaiysectIsect Cambridge Universily
Press Cambridge
Merrill 0 1972 Applications and Extensions of an Algorithm thaI Computes FixedmiddotPoints of
Cenain Upper Semimiddotcominuou Point to Set Mappings PhD Disserlation Unlvel$ity 01 Michigan
Negishi T 1960 Welfar Economics and Exislence I an Equnlbrlum for a Competitive
Economy MelroeCOOQ11ka 12 92middot97
Pereira A 1985 On the Existence and Uniqueness of Optimal Ou1put Path lor CRTS Firms in
Adjustment Costs Technologies Mimeo Stanford University
Pereira A 1985a Consumer Expectations and Ihe COSI 01 UllUty In an Intertemporat
Framework Mlmeo Stanford University
Pereira A 1985b Corporale Tax Integrallon and Government Deficits A Dynamic General
Equilibrium Anaiysls Mlmeo Stanford University
Pereir A 1987 On Ihe Computation of the Users Costs of Stocks Ecooomi (forthcoming)
47
Pereira A bull 1987b Equal Yield Alternatives and Government Oeficlts Mimeof Stanford
University
Pereira A 1987c A Dynamic Applied General ~uHlbrium Model for Tax Policy Evaluation
Comlletamp Docurnent2tion Mimeo University of California San Diego (in progress)
Pereira A 1987d Should the investment Tax Credit be Rei(ltroduced Mlmeo University
of California San Diego (ir progreSs)
Piggott J and J Whalley eds 1985 tJew QeyeQpments in Applied General EQuilibrium
bullenalysjs Cambridge Cambridge University Press
Preckel Pbull 1985 Alternative algorithms for computing economic equilibria In A Manne Ed
Radnor R 19amp3 Equllibrium under Uncertainty In K Arrow M Intrilligaor ods
I ibull
Robinson S 1 g86 Mullisectorai Models of Oeveloping CountrieS a Survey Oivision of
Agricullure and Resources UC Borkeley WP 401
Scarf H 1967 The approximation of fixed points of a continuous mapping ~MM Jouwal of
epplid Mathornallcs 15 pp 1328middot43
Scarf H with T Hansen 1973 00 the Computation of EconomiC Eguilibria New Haven Yale
University Press
Scarf H J Shoven Eds 1984 61l1llied Genera Elujlibrium 6nalyssect Cambridge University
~ Press Cambridge
Shoven J 1976 The Incidence and Efficiency Effects of Taxe on Income from Capital
Jurnal of ramie1 Economy 84 pp 1261-83
Shoven J 1983 Applied General Equilibrium Tax Modemng IMF Slaff PaoerS 30
Shoven J and LB Simon 1986 Share Repurchases and Acquisitions An Analysis of which
Firms Participate NBER Working Paper No 2243
bull
48
IShaven J and J Whalley 1972 ft General Equmbrium Calculation of lhe Effects of
Differential Taxation of Incoma from Capital in the US Jauroe of public Econgmics 1
pp281middot321 (
Shoven J bull and J Whalley 1973 General Equilibrium with Tax$s a computation procedure
and an existence pr00f ReviEhi of Economic Studies 40 pp 475middot490
Shoven J J Whalley 1977 Equal Yield Tax Alternatives A General Equilibrium
Computa~ui0nal Tfchnique J~HJrnal of PUblic EconomIcs 8 pp 211~224 bull
Shaven J J Whalley 1984 Applied GeneralmiddotEquilibrlum MOdels of Taxalion and
International Tr8de An Introduction and Survey Journal of Economic literature 23t pp
1007middot1051
Slemrod J 1980 A General Equilibrium Model of Capital Income Taxation Ph D bullDissertation Harvard Univnrsity
Slemrod J 1983 A General Equilibrium Model of Taxation with Endogenous Financial
Behavior in M Feldstein edbull Behavioral Simulation Methods In Tax Policy Analysis Chicago
UniVerSity of Chicago Press
Stone J 1985 Sequential Optimization and Complementarily Techniques for computing
Economic equilibrium in A Manne Ed
Summers L 1981 Taxation and Corporate Investment a q-Theory Approach BroQkinas
Summers L 1985 Taxation and the Size and ComposHion 01 the Capital Stock An Asset Price
Approach Mlmeo Harvard University
Whalley J 1975 A Geneal Equiloibrium Assessment of the 1973 United Kingdom tax
reform Economica 42 pp 139middot61
1JLTIllOS JIQll~)ltG PAPERS PUBLICADOS bull
I
I I
j
112 70 - LUCENA iJiCIIJo ~ IIEnvironrilent Honitoring and Orqanization Structure T (Setellibro 1987)
nI 71 - L~ENA Diogo A Note on the Representation of Information Stxueture (SeteJObro 1987)
nQ 72 - LUCENA D1CgO Envuorment 111)nitoring and Organization
Structure II (Setelilhro 1987)
ril 73 - AlmmE3 Antonio Pais e CAsPAR Yitor bull Tributacao Inoentivos e InvesUlllentos Analise Qualitativa noYellibro 19S7)
n2 74 - liAlltOSA rose Pedro Opt1lllal lIage R1gidity A SUggestedtlthodology to Teet tb Theory and an Application (Outubro 1987)
nI 75 - CABRAL L 11 B Three lIotes on SYJIIlletric Gemes witt itsYJIIlletric Equilibri (NoveJObro 1987)
n2 76 - LUCENlt D10gOTo Search or Not to Search (Fevereiro1967)
nQ 17 - SA Jorge asconcelos euroI A Theory of SynergyB (Fevereiro1987)
n2 76 - VILARES tanuei Jose Os Be IntenaMios IlIIportados ColIIO Factor de Produao (Julho 1987)
nQ 79 - SA J~rge VasCQncel03 e HoY to Copete and Co~unicate in tature Industrial Products_ (aio 1988)
n2 80 - ROIl Rafael Learning and Capacity Expansion in a New larket Under Uncertainty (Fevereiro 1988)
n2 81 PEREIRA Alfredo llarvilo Survey of Dynamic Computational General EquilibriUlll Hodel for Tax Policy Evaluation (Outubro 1987)
Qualquer in ormacao sabre o 1IIorlltlng Papers ja publlcados sera prestada pelo Secretariado de Apoio aos Docentes podendo os ~e3OSf
ser adquirido5 ne Secyo de Vlndas da Faculdade de EconollLia UNL na Travessa Esteviio Pinto Cltpolide - 1000 LISllOA
27 I Ibullbull
steadymiddotstaU is a possibity it certainly is not the only meaningfull solution to dynamic models
Even in the case of a perfect foresight equilibrium the model implies an equilibrium path which i may or may not involve colanced growth The model not the modeller should dictate the nature 01 I themiddotbase case path Second in the context of a temporarY equilibrium path a steady state solutlcn is Inol a likely model QutcOn1tl In fact unlessmiddot expectations are stattc short run behavior consistent
with a steady-state evolution will in general not be generated On the other hand if static I expectations are self~fufjili[1g we have in fact a perfect foresight model Third even the base year
replication requirement may cause problems In Ihe cOntexl of temporary equilibrium Any
calibration parameter would be conditional on expectation rules which is probably an undesirable
feature
The nalur of the two-requirement calibration strategy is very muchin the spirit of the bull
traditional design of the comparision of alternative equilibria comparis_ion between a steady-state
base case on one hand and alternative paths including a transition period and a finat sleadyyenstate on
the other hand Pereira (198Gb 1987d) compare different (not necessarily steadystate)
equilibrium paths The arguments against such a procedure are based on the idea that the impact of
policy changes can be observed most easily since all departures from the steadymiddotstate can be
attributed to the altemalive polioy On the other hand the base case so dEfined as a steadymiddotstate is
consist~nl with previous work in a less dynamtc setting and Iherefore allows a common standard
for comparing model results
42 Computation algorithms
We are still al a stage in which basically each author uses a different computational technique
Th development of dynamic models - in parlicular wilh adjustment costs andior perfect foresight
bull
28
has corresponded with the decline in the use of fixed point algorithms In -fact given the relative
arge dimensions inevitably involved such algorithms lend to be very inefficient at the best and
chen prohibitively stow See Stone (1985) and Preckel (1965) for a comparative assessment of
different computation techniques Among the models surveyed only Ball1rdFulle(ton~
-ShavenmiddotWhalley (1985) and Feltensteln 1985 use Merrills variant of the fixed point
algorithm 1echnique
AuerbachmiddotKoHikoff (1963 1984 1967) follow a three stage procedure They first compute a
base case steady-state then a revised case steady state and finally transition path for the economy
berveen these two steady-states tn all stages a Gauss-Seidel iterative procedure is used
Ballard (1982) Ballard-Goulder (1965) Goulder (HiSS) and Summers-Goulder (1986)
~
use a method developed by Ballard and Goutder which is similar in many awects to the FairmiddotTaylQr bull
(1983) algorl1hm Short-run equilibria are calculated (using Merrills algorithm) parametric on
nrice expeC~lions The model is then iterated to generate self-fulfilling intertemporaf expectations
and the corresponding perfect foresight equilibrium In a relatively similar approach Andersson
1987 uses a simulation program SIMNON developed In the University of Lund This program
can handle two-paint boundary problems In a fashion consistent with the multiple shooting
algcrlthm (see Lipton-Poterba-Sachs-Eummers (1982))
Boyenbergs computational approach (1985 1985) differs from the other models surveyed in
that he relies heavily on analytical techniques Computations are done by using a dynamic version of
Johansons linearization method Being essentially determined by the continuous time nature of the
I I
f
model thiS linearization model has the disadvantage of confining the analysls to infinltesinal changes
around the base case equilibrium (see Bovenberg (1985) p 53) Pereira (1986b 1987d) uses an optimization algorithm NPSOL - developped by
rGillMurray-Saunders-Wrlght (1986)) This algorithm is used 10 compute the sequence of
I
29
Shortmiddotrun temporary equilibrium which makes up the intertemporal equilibrium path The
equilibrium conditions are seen as nonlinear equality constraints in the minimization of an artlcial
objective function The prices are normalized to the unit simplex by an additional linear equality I I 1 bull constraint
II ErlichmiddotGinsburgh-Heyden (1987) follow a unique approach in that they use a variant of the
optimization technique introduced by Negishl (1960) The economic equilibrium can be generated i
as a solution of a mathemallcal program Ihe objective function of which is a weighted sum of the
utility functions of the various agents while tho constraints set consists of the market dearing 1
conditions Ginsburgh-Heyden (1985) have extended Negishis resulllO tho case of downward price
lrigidities I
Finally the paper by Jorgonson-Yun (1984) is also unique in tIlat it is the only I
econometrically estimated model Different blocks for the consumption and production Side of the
-
model are separately estimated to provide the necessary structural parameters i
I The diversity of computation techniques is yet another indicator of the exploratory nature of the
body of literature surveyed in this article
43 Equal yield comparisions
The link between a base case and counteriactual simulations Is usually provided by the concept of
equal yield Shaven-Whaley (1977) discuss the meaning 01 equal yield in a general equilibrium
context when government is not allowed to run defICits Equal yield is interpreted to moan constant
public utility Government base case utility is maintained in the counferfactual experiments With
balanced budgets the concepl of equal yield is unambiguous The new equUibrium prices and the
balanced budge condition will del ermine the minimal expendilure and taxes needed to maintain base
30
case publi~ utility Accordingly in general equal yield is inconsistent with equal nominal tax
revenue Some change in tax revenue Is necessary Different tax replacement schemes ate
considered to assure that enough tax revenue is collected This is the approach essentially followed
by most 01 the papers surveyed Only Feltenstein (1985) Goulder (198) and Jorgenson-Yun
(1984) chose not to follow an equal yield strategy
The question is of how to Interprete lila concept of equal yield when the government is alowed to
fUn deflcLts The optimal leve of expenditure for base ease public trliUty can now be taAinanced
bond financed or financed by a mix of bonds and taxation We have several versions of equal yield In
particular equal yield may now be consistent with equal tax revenUE Furthermore some meaSure
ltgtf financial crowding out effects of government deficns can be inferred from the comparislon of the
several equal yield alternatives These issues are extensively discussed ~Pereira (1987b)
44 Price expectations and the dynamic generalization of compensation
Indicators
The sum of equivalent and compensation variations over households is the most widely used
aggregated measure of efficiency gains or losses In the context of $teady~state eomparisions andor
when complete future markets exist andlor perfect foresighl is assumed there are no difficulties
associated with the use of the standard Hicksian indicators In fact correct future prices are known
in these cases
If expectations are not selfmiddot fulfilling andlor future markels are not available a dynamic
generalization is necessary 8allardmiddotGoulder (1985) provide some steps in that direction by
defining an indicator that accomodates periods far into the future when households do not have
perfect foresight but a steady~state prevails On he other hand this issue is more fundamental in
i
[
I
i I I I I [
I I
bullI
31
the contex~ of a Gner1 lemporary equilibrium ftamawor~ In such Circumstances Pereira
(1986a) dev~lops a cnamlc generalization which is obtained as the present discounted value of a i i
sequence of short-fUn optimal expenditure functions consistent with a base case expected future I stream of IlWities I
32
5 Empirical evidenee from selected policy Issues
Dynamic tax models have generated several important results which escaped Sialic modellers
The following is a seleCpoundd set of issuos whIch stress the marginal benefits of dynam~ modelHng of
economic behavior in innovltlttive areas of analysis The discussion of a consumption tax emphasizes
the benefits of dynamic tusehold behavior and intergenerational aspects The study of the
flliminatlon or the re~intToduction of the investsnemt tax credits is made meaningful by the dynamic
modelling of production betlavor Modelling governmenl in a dynamic conlexts allows the modeller
to study the irrrpact of government deflcits Finally a dynamic framework lets us appreCiate the I
relevance of consumer expectations in terms of the evaluation of policy alternatives
i
i
51 Consumption Tax
t
By Its very nature a consumption tax can not be adequately investigated with a static model bull
Fullerion-Shoven-Whaicy (1983) use Ihe model of the US economy s described in
Ballard-Fullerton-Shoven-Whalley (1935) to evaluate the movement from the currenl US tax Isys1em to a progressive consumption tax Since their modeJ Incorporates a Jaborleisure choice
where leisure is an untaxed commodlty their results reflect the fact that both the consumption tax
and Ihe present system are dslortionary
Concentrating on the rntertemporaf distortions they show that sheltering more saving from lhe
current US income tax could improve econornic efficiency even if marginal tax rate increases are
necessary in order to maintain government fevenue~ At first you have a revenue shortfall
However the economy moves to a higher sustatned growth path and uitimatefy lower tax rates
cangenerate the same revenue path Also wages increase in the long run with this policy The
bull
bullbull ~
33 shyt
i
I
I
present value of welfare gi[lS for a potiey of complete savin9 deduction wilh marginal tate
adjustments (consumption lax) is simulated to be around $500 billion to $600 billion 01 1973
dollars
They also investiga1e th 1ngth of ttme it takes th~ economy to adjust to these policy changes
Roughly they estimate the long run to b thirty years although this figure Is very sensitive to the
specification of the savings elasticity - a crucial parameler in this model
52 Investment T~x Credit
Goulder-$ummers (1 S87) address the impact of eliminating the lnvestmenl Tax Cradit (ITC) on
intersectoral capital formation and on economic growth Their model isYa1ticularly adequate to
addies$ this issu~ in that is postulates a forward looking investment m9we with adjustment costs
They show that the eliminating the ITC causes a reduction in the rale of investment Inveslment is
estimated to fall by about 7 in the short-run and by about 12 in Ihe new long run steady state
In lurn a previous policy anncuncement lowers Ihe overall al1ractiveness of Investment and leads 10
a downward shift of the Investrrent profile On the other hand the combined effect of a revenue
neutral simultaneous efimiraron of the ITC and reduction of the corporate tax rates is a long run
reduction ef the capital Sieck by 35 This pattern suggests that a revenue neutral increament in
both Ihe corporale tax rates and the inveSlment lax credits would be preferable 10 the formula
adopted in Ihe US Tax Reform Act of 1986
Pereira (1987d) addresses the impact of the ITe from Ihe oland point of Ihe curTenllax syslem
under the Tax Reform Act of 1986 Given Ihe concern over the potential depressiVe effects upon
savings and investment of the new tax system in conjunction with deficit reduC1ion mechanisms of
the GrammmiddotRudman type a pertinent question is should Ih lTC be e-introduced In the context
bull
34
-of his mod~1 of the US economy Pereira focuses on the trade-off between the distortion in the
intersectoral allocation of capital induced by the lTC the lower tax revenues it generates and
potential financial crowding-out effects on one hand and the positive effects of lowering the relative
price of new capital goods on the other hand Preliminary results confirm the qualitative results in
Goulder-Summers (1987) suggestting a welfare gain of about 2 of the present value of GNP in
the best scenario of absence of replacement taxes Furthermore preliminary results show an
important time pattern to welfare gains with the average benefits increasing the further you look
into the future This pattern is due to the presence of constraints to intersectoral mobility of capital
and inertia in the adjustment towards the optimal capital levels as reflected by the presence of
adjustment costs
I
I
bullAuerbach-Katlikaff (1gB) in the context of their intergenerational model of the US economy
consider the impact on savings capital formation and interest rates of deficit policies and balanced
budget increases in government consumption They conclude that deficit finance and government
consumption can significa-ntly crowd out capital formation and lower the welfare of future
generations However crowding out from deficit finance is a very slow process because it results
from increased government spending over potentially long horizons Also deficit policies may
substantially influence the longmiddotrun interest rates while leaving the short-run rates essentially
unaffected Finally and in opposition to the central role adjustment costs seem to play in both
53 Financial crowding outeff~cts of government deficits
Goulder-Summers and Pereiras models Auerbach-Kotlikoff suggest that the time path of interest
rates induced by a policy of deficit financing seems to be insensitive to the presence of adjustment
costs
bull
35
54 The role 01 consumer expectatlons
The expectations analysis has uncovered one of the benefits of dynamic modelling
Ballard-Goulder (1985) find that the appeal of adopting a consumption tax depends on the level of
foresight possessed by the consumers Furthermore additional foresight may be welfare worsening
This is a second best type of result This tends 10 occur under policies that lead to capital deepening
and declining rate of return to capital over time To the extent that consumers have more foresight
they will be beller equippod to anticipate the fall in the rental price of capital Consequently if a
saving incentive is enacted people save more with myopia than with perfect foresight Given the
exi~tence of taxes on capita and the discrepancy between the private an~--sccial returns to capital
the greater savings with myopia is beller socially Ballard-Goulder find that the welfare gains
from a consumption tax are reduced by about 10 when we move from myopia to farleet foresight
Therefore the attractivenness of adopting a consumption tax seems fairly [obust across these
foresight specifications
I
bull
36
6 Tbe power and weaknesses 01 dynamic modelling the example 01 corporale
tax Integrallon
The Issue of corporate tax integration allows us to sh~w the stre-ngths and weaknesses or the
dynamic approach Empirical evidence using CGE techniques indicates that depending on the precise
scheme of integration and the 1ax replacement methods integratiOn may have substantial effects In
Ihe work of FuliertonmiddotKingmiddotShoven-Whalley (1980 1981) total integration was found to yield an bull
annual static eHielency gain of $4 to $8 billion in 1973 dollars Simulated dynamic gains may be as
large as $695 billion or abcu 14 of the present value of tulUre consumption and leisure in the
US economy These gains result primarily from interindustry reallocalions of Investment and an
improved inter~emporal allocation of consumption
Mora recent work emphasizes hOw consumers asset portfolio decisions and firms finanCial
decisions affect the efficiency gains from integration Slemrod (1980) focusing on consumers
asset ponfoio deCisions finds static efficiency gains which are about twice -as Iarge as those
repOrted iiy Fulierton-King-ShovenWhalley (1961) FuUrlon-Gurdon (1983) focus on the
firms financial decisions They roporl efficiency gains of 6 of GNP from the elimination of the
tax distor1ions favoring debt However when they eliminate the corporate tax and repiace it with
increaSed persona income taxes addlHonal dls10t1ions are created in the optimal labot~leisure
decisions These distortions tend to dominate the analysis slJ(h that the overall effects of complete
integration are very modest Galpr-LuckmiddotTodr (1986) address simultaneously consumers
asset pOrlfollo decisions and the firms financial decisions They also find very modest efficiency
gains from integralion
The above results are important but may be severely biased There are severa aspects of
economic behavior and modelling crucial for the study of Income tax Integration whiCh have no been
i 37
captured innny of the above papers One aspect is the absence of government deficitsmiddot a balanced
budget is assumed It is true Ihal resource crowding OUI is caplured In these mOdels but financial
crowding out induced by government spending is nOI Second investment is not derived from
optimtztlcn behavior Investment behavior passively accomodates endogenous saving decisions As
a consequence the differential impacts of policies in the incentives to save and to invest are not
captured Aso full capital mobility across secors is assumed with instantaneous capital
adjustmenlS towards optimal levels This assumption rules qut different costs of capital across
seclors and therefore differentiated reactions to tax policies changes Finallyhe modelling of both
gov9rnmerlt deficits and of endogenous real and financial investment decisions necessitates the
~nsideraiion of a dynamilt framework and the introduction of financiaJ assets govemment bonds and
iprivate financial assets A dynamic framework also highlights the efficiency effects of Integration on r
ithe ptimal intertemporal decisions The above models are either sIalic (Slemrods and
GalpermiddotltJckemiddotTode(sj or a dynamic sequ~nce of otherwise Slatic model bullbull I Pereir a (1986b) develops a dynamic applied general equilibrium model bull to study the tbull
efficiency effects of integration on the growth and allocation of investmenl across sectors Special
Iattention is paid to the structural affeelS of resource and financial crowding oul induced by
gove mect spending and deficits and to the real and financial Investment reactions across industries
to both tax integration and flnancial crowding out His mode departs from previous work on income
lax Integraticn and for that matler from most 01 the CGIE literalure for tax policy evaluation in
several direcllons It encompasses an endogenous sequential equilibrium struelure founded on
dynamic behavior with flexible expectations Government deficits are optlmaUy determined
Investment decisions are tward looking and the resull of optimizing behavior Several financial
assetsmiddot public and privatemiddot are considered
Simulation resulls Indicale that the welfare gains from Integration under large deficits are at
32
best lJlodest when measured in terms of the GNP The elimination of the corporate tax and its I replacemenl by increased income lax rates yields long run benefits which are never larger than
2 of the presenl yalue ct telr consumption and leisure onder the previous tax regime and 1
under the current tax regime However the short run w~lare effects of full integration tend 10 be
very low and in in some scenarios even negative This is a flew intertemporar pattern of efficiency Ieffects which reflects an adjJsiment lag in the interindustry investment decisions due to the
existence of costs of adjustment On the of her hand partial integration achieved by excluding I dividends from the corporale lax base syslematically yields negalive efficlencyeffeclS In Ihe long Inm these can be as high as 3 of present value of the future value of consumption and leisure This
is a new second best effect suggesllng Ihat less than complete integration may have perverse I efficiency effects
The dynamic slructure of PerclrBs model snowing for an improved freatment of real investment
declsions and government deficts provides a potential setting for a more accurate measure of the
costs and benefits of inlegration The simulation resullS suggest lower benefits and an intertempora Ibull
pattern of growing benefits
Simulation results also clearly suggest robust negative effects from partial integration How
reliable is this result If dJvidends were deductible from the corporate lax base (as are interest
payments) the preferenlial Ireatmant of debl over equily would be ellminaled Corporallons should
be expecled 10 react by decreasing the optimal deWeqully rallo However corporale financial
gecisions are exogenously set In Pereiras model as in aft of the other models surveyed that go as far
as dealing with the issue
Also withdrawing dividends from Ihe corporate tax base is a way of eliminating Ihe double
taxation of dividends at belh Ihe corporate and personnal Income levels How will the optimal
aliocellon of household saving accomodate that change in Ihe relalive return to corporale equity
39 I
t However hOusehold portfolio decisions are exogenouly set in Pereiras model as in all of the other
models surveyeltj go as far as dealing with the issue
IIt is sate to say that the evaluation 01 the benefits 01 partial Integration as defined above are
sevrey underestima1eO Meanwhile the distortionSoenerated by the increase in the marginal
personal tax rates designed to ralse the revenue foregone by the dividend exclusion are fully
aCCltlunted for A good measure of the costs of integration together With a poor evaluation of the
benefils may very well be the reason why partial integration is Simulated 10 yield neg alive
eHikiency gains
On a different vein as most of Ihe model surveyed here Pereiras is a closed economy model It
is legitimate to wonder how the resullS would change If international capital flows were anowed
IThis almost certainly would affect financial crowding out since a good d~al of the capital inflows
could be used to finance public debt
For these various reasons the results should be treated as preliminary but they strongly I suggosl that the dynamic structure provides valuable new insights into the corporate tax integration SSU2
40
7 Concluding remarks I I I i
What has been acccmplished Ihis far The success of Ihe CGE research in laelding the challenge
of dynamic modelling can not be denied
I iGreat progress in the modelling of household behavior has been achieved Promising
developm~nts in the modeUing of production and government behavior have also been accomplished
Interesting innovations in the concept of equilibrium and computation techniques were discussed
The policy analysis was greatly enriched by considering transitional effects together with longer
run steadymiddotstate equilibrium paths generalized equal yield strategies accomodating different
financial crowding out impacts and generalized dynamic policy evaluation indicators
An important set of issues have been addresed by the models surveyed Traditional issues
focusing on the effects of capital taxation and consumption taxes have been pursued In terms of
capital taxation for example the intertemporal nature of the issue was enhanced by allowing an
optimal evolution of capital stock in the economy and forward looking optimal investment decisions
In Andersson (1987) GouldermiddotSummers (1987) and Pereira (1986b 1987d) this is coupled
with a improved treatment of several tax provisions like investment tax credits and depreciation
allowances In terms of the consumption taxes developments in the specification of Intertemporal
household behavior the introduction of overlapping generations and the modelling of
intergenerational links via bequest motives as well as a closer attention to demographic evolution
provided a much improved economic setting for the understanding of the several aspects of the
problem
Furthermore dynamic modelling has permitted the addressing of a set of new issues Policy
Issues ranging from dynamic tax policy analysis to the evaluation of exogenous changes in
government expenditures to the impact of different methods of financing government expenditure to
4 1
the importance of tinanciai croHrii1g out to 1he relevance of consumorS expectations in tne policy
results have been addressed
Despile all of the progress and in part due to such progress several avenues are wide open for
much neded additional research The following seem to be the most Interesting and promising
areasmiddot First the specincation of liquidity constraints to household behavior may help to obtain a
better understanding of policy chenges that affect directly the interest rates in the economy Second
major attention needs to be paid to the incorporafion of a roraign sector (with open capital markets)
Third some efforts are needed in terms of finding adequate computation techniques now that
dynamic modelling eally makes the curse of dimensionality worse than eVOrbull
The modelling of financial markets is the single most unsatisfactory speet of the dynamic
models surveyed here~ The problems associated with provlding a mo~ adequate treatment of
financial markets and in general endogenous and optimal financial deci$ions bull both at the corporate
and at the personal levelsmiddot have 10 be addressed That necessarily requires Introducing uncertainty
into the models To b adequate this must go well beyond the mere assumption of point expeetalions
with the whole array of modelling implementalion and evaluation problems thai generates
bull I i
42
REFERENCES
Abel A and O Blanchard 1983 An Intertemporal Model of Saving and Interest Econometrica
51 pp 675-92
~ Andersson K 1987 Tar-ation of Capital in Sweden - A General Equilibrium Model Ph D
Dissenation UnIversity of Lund Sweder
Auerbach A L KotlikoH 1983 National Savings Economic Welfare and the Structure of
Taxation in M Feldstein ed Behavioral Simulation Methods in Tax Policy- Analysis Chicago
University of Chicago Press
Auerbach A l Kotlikoff 1984 Social Security and the Economics of Demographic-
Transition in H Aaron and G Burtless eds Retirement and Economic Behavior Washington DC
Brookings Institution
Auerbach A L Kotlikoff 1987 Dynamic Fiscal Policy Cambridge Cambridge University
Press
Auerbach A L Kotlikof and J Skinner 1983 The Efficiency Gains from Dynamic Tax
Reform Internatiooaj Economic Reyiew 24 pp 81middot100
Ballard C 1983 Evaluation of the Consumption Tax with Dynamic General Equilibrium
Models Ph D Dissertation Stanford University
Ballard C 1987 Tax Policy and Consumer Foresight A General Equilibrium Simulation
Study Economic Inquiry 25 pp 267-284
Ballard C and L Goulder 1985 Consumption Taxes ForeSight and Welfare a Computatonal
General Equilibrium Analysis in J Piggott and J Whalley eds New Deyelooments in Apolied
General Equilibrium Analysis Cambridge Cambridge University Press
Ballard C D Fullerton J Shoven and J Whalley 1985 A General Equilibrium Model for
t 43
Tax Eoliey Eyalua~iQO NBER University or Chicago Press
Borges A (1987) bull A Survey of Compula1ional General Equilibnum Models CECD Economic
Slud~ 8
Bovenbig L 19B3 imperfect Capital Mobility in a Perfect Foresight Herberger Model
Mimea UC Br~eley
Bovenberg L 1984 Capitel Accumulation and Capital Immobility Qmiddottheory In a Dynamic General Equilibrium Framework PhD dissertation University of California at Berkeley t
Bovenberg L 19850 Dynamic General Equilibrium Tax Models with Adjustment Costs in A iManne ed bull 1985 EtQnomrc ECUjilbrium Model Formulation and Solution Mathematical
lProgramming Siudy 23 Arnsterdam NorthmiddotHoliand
Bovenberg L t 1986 Capital Income Taxation in Growing Open Eeondmies Journa of Pubtic
Eamp(omlcs 3t pp 347middot376 IBovenberg L ald W Keller 1981 Dynamics and Nonlinearlties in Applied General
Equilibrium Models Internal Report Mlmiddot81middot208 Department for Statistical Methods CBS bull
Vooiburg The Netherlands
Charnley C 1981 The Welfare Cost of C~pltal income Taxation In a Growing Economy
IJournal of Political ECQnQm~ 89 3 pp 468middot496 I
Charnley C 19810 Efficient Tax Reform in a Dynamic Model of General Equilibrium I Mimeo Yale University I Charnley C 1982 On the Optimal Taxation In Stylized Models 01 Inlartemporal General
Equilibrium Mimeo Yale University I Decaluw B and A Manens 1985 Pays en Oeveloppement at Modeles Calculables OEquilibre
General - Une Revue de al litterature empirique Monographie Sene 9 Volume 3 Centre de
Recherche et Developpemont Economlque Unrslt de Montreal
44
bull
I I
I
I
1
Erlich $ V Ginsburgh and L van der Heyden 1987 Wher~ do Real Wage Policies Lead
Belgium bull A General Equilibrium Analysis European Economic Reyiew (forthcoming)
air R and J Taylor 1983 Solution and Maximum Likelihood Estimation of Dynamic
Nonlinear Ratioant Expeclations Models Ecooometrica 51 pp 1169middot1186
Feltenstein A 1984 Money and Bonds In a Disaggregated Ec~nonYin Searl 1-1 J Shoven
Eds Aoolied ~eoeral EqJiliDCiulD tioalYsis Cambridge University Press Cambridge
Feltenstein A 19B6 A Dynamic General Equllibrlu Analysis of Financial Crowding Out
Theory with an Application to Australia Journal of Public Economics 31 No1
Fullerlon D R Gorden 1983 A Reexamination of Tax Distorllons in General Equilibrium
Models in M Feldstein Ed Behayora Simulation Methods in Tax pOlicy Analysis Chicago
University of Chipgo Press
Fvllerlon D J Shaven and J Whalley 1983 Replacing the US Income Tax with a
Progressive Consumption Tax JoulOal of Public ECQnomics 20 pp3middot23
Fullerton D Y Henderson J Shaven 1984 A COmparision of Methodologies in Empirical
General Equilibrium Models of Taxation in Searl H J Shoven Eds ~pQlied Geoaral Jguilibrium
lIoalyss Cambridge University Press Cambridge
Fullerton 0 A King J Shaven and J Whaney 1981 Corporate Tax Integration in the United
Statesmiddot a General Equiiibdufi1 Approach American Economic Review 71 pp~ 677~691
Galper H R Lucke E Tader 1985 Taxation portlor Choice and the Allocation of Capital A
General Equilibrium Approach Mimeo
Gill p W Murray M Saunders M Wright 1986 Users Guide for NPSOL (Version 40) A
Fortran Package for Nonlinear Programming T echnicsl Report SOL 86middot2 Deparlment of
Operations Research Stanford University
Glnsbur~h V and L van der Heyden 1985 General Equilibrium with Wage Rigidities An
45
AppJication to Belgium in A Manne ed t Economic EQumprit1m~ Made Formulation and $oJioo
Mathematical Programming Study 23 Amsterdam NorthmiddotHolland ~
Goulet J 1968 1 Adjustrent CO$1S in the Theory o(the Firm Reyiew of EconomiC Studies 35 bull
pp47middot55
Goulder L 1985 Tax-Financed versus BondmiddotFinancedGhanges in Governent Spending
bull Mimeo Harva(d UniversilYmiddot
Goulder l J Shaven and J Whalley 1983 Domestic Tax Policy and the Foreign Sector The
importance of altemative foreign policy formulations to results from a general equilibrium tax
analysis model In ~n methods in Ibe antico of income from caoia1 ed Martin
Feldstein Chicago University of Chicago riSS
Goulder l L Summers 1987 Tax Policy Asset Prices and 3rowtly ~
A General Equilibrium Analysis NBER Working Pper No 212B
Grandmont J 1982 Temporary General Equilibrium Theory Ch 20 in Handbook 01
Mathemalical Eccnornics Vol II NorthmiddotHoliand
Harberger A 1959 The Corporate Income Tax An empirical Appraisal In lax Revision
Comoendjum Voll House Cornmitee on Ways and Means Washington DC Govemment Printing
Office
Harberger A 1962 The Incidence of the Corporate Income Tax JOULn1 of Political
Economy 70 pp 215middot240
Harberger A 1964 Taxation Resource Aliocalrlon and Welfare fn lb Role of Dir bullbulll and
Indjresl Taxes io tbe Fede[al Reserve s~I~Ill Princeton Princeton Unlveresity Press
James J 1985 New Developments in the Applications of General Equilibrium models to
Economic History in Piggott aand Whalley Eds
Jorgenson D and K Yun 1984 Tax Policy and Capital Allocation Harvard Inslitute of
bull
I
I f
Economic Research WP 1107
Judd K 1965 ShonmiddotRun Analysis of Fiscal Policy in a Simple Perfect Foresight Model
Journal of poJiHcal 5CCJiW1l pp 298middot319 r
Upton D J Potbc J Sachsa nd L Summers 1982 Multiple Shooting in Rational
Expectations Models fuoometrlc 50 1329middot1333
Manne A ed 1985 fconQrnjc Eoullibrium Model FOrIDujaljQo and Solution Mathematical
Pogrammlng Study 23 Amsterdam NorthmiddotHolland
Manne A 1985 Or the Formulation and Solution 01 Economic Equilibrium Models In A
Manne ed ECQOQlli-Jr~ibmiddotum ~tQdel EQlJulailon and Solution Malhematical Programming
Study 23 Amsterdam NorthmiddotHolland
Mansur A and J Whalley 1984 Numerical Specification of Applied General Equilibrium
Models in Scarf H J Shoven Eds Aoplied General Eouillbrium AnaiysectIsect Cambridge Universily
Press Cambridge
Merrill 0 1972 Applications and Extensions of an Algorithm thaI Computes FixedmiddotPoints of
Cenain Upper Semimiddotcominuou Point to Set Mappings PhD Disserlation Unlvel$ity 01 Michigan
Negishi T 1960 Welfar Economics and Exislence I an Equnlbrlum for a Competitive
Economy MelroeCOOQ11ka 12 92middot97
Pereira A 1985 On the Existence and Uniqueness of Optimal Ou1put Path lor CRTS Firms in
Adjustment Costs Technologies Mimeo Stanford University
Pereira A 1985a Consumer Expectations and Ihe COSI 01 UllUty In an Intertemporat
Framework Mlmeo Stanford University
Pereira A 1985b Corporale Tax Integrallon and Government Deficits A Dynamic General
Equilibrium Anaiysls Mlmeo Stanford University
Pereir A 1987 On Ihe Computation of the Users Costs of Stocks Ecooomi (forthcoming)
47
Pereira A bull 1987b Equal Yield Alternatives and Government Oeficlts Mimeof Stanford
University
Pereira A 1987c A Dynamic Applied General ~uHlbrium Model for Tax Policy Evaluation
Comlletamp Docurnent2tion Mimeo University of California San Diego (in progress)
Pereira A 1987d Should the investment Tax Credit be Rei(ltroduced Mlmeo University
of California San Diego (ir progreSs)
Piggott J and J Whalley eds 1985 tJew QeyeQpments in Applied General EQuilibrium
bullenalysjs Cambridge Cambridge University Press
Preckel Pbull 1985 Alternative algorithms for computing economic equilibria In A Manne Ed
Radnor R 19amp3 Equllibrium under Uncertainty In K Arrow M Intrilligaor ods
I ibull
Robinson S 1 g86 Mullisectorai Models of Oeveloping CountrieS a Survey Oivision of
Agricullure and Resources UC Borkeley WP 401
Scarf H 1967 The approximation of fixed points of a continuous mapping ~MM Jouwal of
epplid Mathornallcs 15 pp 1328middot43
Scarf H with T Hansen 1973 00 the Computation of EconomiC Eguilibria New Haven Yale
University Press
Scarf H J Shoven Eds 1984 61l1llied Genera Elujlibrium 6nalyssect Cambridge University
~ Press Cambridge
Shoven J 1976 The Incidence and Efficiency Effects of Taxe on Income from Capital
Jurnal of ramie1 Economy 84 pp 1261-83
Shoven J 1983 Applied General Equilibrium Tax Modemng IMF Slaff PaoerS 30
Shoven J and LB Simon 1986 Share Repurchases and Acquisitions An Analysis of which
Firms Participate NBER Working Paper No 2243
bull
48
IShaven J and J Whalley 1972 ft General Equmbrium Calculation of lhe Effects of
Differential Taxation of Incoma from Capital in the US Jauroe of public Econgmics 1
pp281middot321 (
Shoven J bull and J Whalley 1973 General Equilibrium with Tax$s a computation procedure
and an existence pr00f ReviEhi of Economic Studies 40 pp 475middot490
Shoven J J Whalley 1977 Equal Yield Tax Alternatives A General Equilibrium
Computa~ui0nal Tfchnique J~HJrnal of PUblic EconomIcs 8 pp 211~224 bull
Shaven J J Whalley 1984 Applied GeneralmiddotEquilibrlum MOdels of Taxalion and
International Tr8de An Introduction and Survey Journal of Economic literature 23t pp
1007middot1051
Slemrod J 1980 A General Equilibrium Model of Capital Income Taxation Ph D bullDissertation Harvard Univnrsity
Slemrod J 1983 A General Equilibrium Model of Taxation with Endogenous Financial
Behavior in M Feldstein edbull Behavioral Simulation Methods In Tax Policy Analysis Chicago
UniVerSity of Chicago Press
Stone J 1985 Sequential Optimization and Complementarily Techniques for computing
Economic equilibrium in A Manne Ed
Summers L 1981 Taxation and Corporate Investment a q-Theory Approach BroQkinas
Summers L 1985 Taxation and the Size and ComposHion 01 the Capital Stock An Asset Price
Approach Mlmeo Harvard University
Whalley J 1975 A Geneal Equiloibrium Assessment of the 1973 United Kingdom tax
reform Economica 42 pp 139middot61
1JLTIllOS JIQll~)ltG PAPERS PUBLICADOS bull
I
I I
j
112 70 - LUCENA iJiCIIJo ~ IIEnvironrilent Honitoring and Orqanization Structure T (Setellibro 1987)
nI 71 - L~ENA Diogo A Note on the Representation of Information Stxueture (SeteJObro 1987)
nQ 72 - LUCENA D1CgO Envuorment 111)nitoring and Organization
Structure II (Setelilhro 1987)
ril 73 - AlmmE3 Antonio Pais e CAsPAR Yitor bull Tributacao Inoentivos e InvesUlllentos Analise Qualitativa noYellibro 19S7)
n2 74 - liAlltOSA rose Pedro Opt1lllal lIage R1gidity A SUggestedtlthodology to Teet tb Theory and an Application (Outubro 1987)
nI 75 - CABRAL L 11 B Three lIotes on SYJIIlletric Gemes witt itsYJIIlletric Equilibri (NoveJObro 1987)
n2 76 - LUCENlt D10gOTo Search or Not to Search (Fevereiro1967)
nQ 17 - SA Jorge asconcelos euroI A Theory of SynergyB (Fevereiro1987)
n2 76 - VILARES tanuei Jose Os Be IntenaMios IlIIportados ColIIO Factor de Produao (Julho 1987)
nQ 79 - SA J~rge VasCQncel03 e HoY to Copete and Co~unicate in tature Industrial Products_ (aio 1988)
n2 80 - ROIl Rafael Learning and Capacity Expansion in a New larket Under Uncertainty (Fevereiro 1988)
n2 81 PEREIRA Alfredo llarvilo Survey of Dynamic Computational General EquilibriUlll Hodel for Tax Policy Evaluation (Outubro 1987)
Qualquer in ormacao sabre o 1IIorlltlng Papers ja publlcados sera prestada pelo Secretariado de Apoio aos Docentes podendo os ~e3OSf
ser adquirido5 ne Secyo de Vlndas da Faculdade de EconollLia UNL na Travessa Esteviio Pinto Cltpolide - 1000 LISllOA
bull
28
has corresponded with the decline in the use of fixed point algorithms In -fact given the relative
arge dimensions inevitably involved such algorithms lend to be very inefficient at the best and
chen prohibitively stow See Stone (1985) and Preckel (1965) for a comparative assessment of
different computation techniques Among the models surveyed only Ball1rdFulle(ton~
-ShavenmiddotWhalley (1985) and Feltensteln 1985 use Merrills variant of the fixed point
algorithm 1echnique
AuerbachmiddotKoHikoff (1963 1984 1967) follow a three stage procedure They first compute a
base case steady-state then a revised case steady state and finally transition path for the economy
berveen these two steady-states tn all stages a Gauss-Seidel iterative procedure is used
Ballard (1982) Ballard-Goulder (1965) Goulder (HiSS) and Summers-Goulder (1986)
~
use a method developed by Ballard and Goutder which is similar in many awects to the FairmiddotTaylQr bull
(1983) algorl1hm Short-run equilibria are calculated (using Merrills algorithm) parametric on
nrice expeC~lions The model is then iterated to generate self-fulfilling intertemporaf expectations
and the corresponding perfect foresight equilibrium In a relatively similar approach Andersson
1987 uses a simulation program SIMNON developed In the University of Lund This program
can handle two-paint boundary problems In a fashion consistent with the multiple shooting
algcrlthm (see Lipton-Poterba-Sachs-Eummers (1982))
Boyenbergs computational approach (1985 1985) differs from the other models surveyed in
that he relies heavily on analytical techniques Computations are done by using a dynamic version of
Johansons linearization method Being essentially determined by the continuous time nature of the
I I
f
model thiS linearization model has the disadvantage of confining the analysls to infinltesinal changes
around the base case equilibrium (see Bovenberg (1985) p 53) Pereira (1986b 1987d) uses an optimization algorithm NPSOL - developped by
rGillMurray-Saunders-Wrlght (1986)) This algorithm is used 10 compute the sequence of
I
29
Shortmiddotrun temporary equilibrium which makes up the intertemporal equilibrium path The
equilibrium conditions are seen as nonlinear equality constraints in the minimization of an artlcial
objective function The prices are normalized to the unit simplex by an additional linear equality I I 1 bull constraint
II ErlichmiddotGinsburgh-Heyden (1987) follow a unique approach in that they use a variant of the
optimization technique introduced by Negishl (1960) The economic equilibrium can be generated i
as a solution of a mathemallcal program Ihe objective function of which is a weighted sum of the
utility functions of the various agents while tho constraints set consists of the market dearing 1
conditions Ginsburgh-Heyden (1985) have extended Negishis resulllO tho case of downward price
lrigidities I
Finally the paper by Jorgonson-Yun (1984) is also unique in tIlat it is the only I
econometrically estimated model Different blocks for the consumption and production Side of the
-
model are separately estimated to provide the necessary structural parameters i
I The diversity of computation techniques is yet another indicator of the exploratory nature of the
body of literature surveyed in this article
43 Equal yield comparisions
The link between a base case and counteriactual simulations Is usually provided by the concept of
equal yield Shaven-Whaley (1977) discuss the meaning 01 equal yield in a general equilibrium
context when government is not allowed to run defICits Equal yield is interpreted to moan constant
public utility Government base case utility is maintained in the counferfactual experiments With
balanced budgets the concepl of equal yield is unambiguous The new equUibrium prices and the
balanced budge condition will del ermine the minimal expendilure and taxes needed to maintain base
30
case publi~ utility Accordingly in general equal yield is inconsistent with equal nominal tax
revenue Some change in tax revenue Is necessary Different tax replacement schemes ate
considered to assure that enough tax revenue is collected This is the approach essentially followed
by most 01 the papers surveyed Only Feltenstein (1985) Goulder (198) and Jorgenson-Yun
(1984) chose not to follow an equal yield strategy
The question is of how to Interprete lila concept of equal yield when the government is alowed to
fUn deflcLts The optimal leve of expenditure for base ease public trliUty can now be taAinanced
bond financed or financed by a mix of bonds and taxation We have several versions of equal yield In
particular equal yield may now be consistent with equal tax revenUE Furthermore some meaSure
ltgtf financial crowding out effects of government deficns can be inferred from the comparislon of the
several equal yield alternatives These issues are extensively discussed ~Pereira (1987b)
44 Price expectations and the dynamic generalization of compensation
Indicators
The sum of equivalent and compensation variations over households is the most widely used
aggregated measure of efficiency gains or losses In the context of $teady~state eomparisions andor
when complete future markets exist andlor perfect foresighl is assumed there are no difficulties
associated with the use of the standard Hicksian indicators In fact correct future prices are known
in these cases
If expectations are not selfmiddot fulfilling andlor future markels are not available a dynamic
generalization is necessary 8allardmiddotGoulder (1985) provide some steps in that direction by
defining an indicator that accomodates periods far into the future when households do not have
perfect foresight but a steady~state prevails On he other hand this issue is more fundamental in
i
[
I
i I I I I [
I I
bullI
31
the contex~ of a Gner1 lemporary equilibrium ftamawor~ In such Circumstances Pereira
(1986a) dev~lops a cnamlc generalization which is obtained as the present discounted value of a i i
sequence of short-fUn optimal expenditure functions consistent with a base case expected future I stream of IlWities I
32
5 Empirical evidenee from selected policy Issues
Dynamic tax models have generated several important results which escaped Sialic modellers
The following is a seleCpoundd set of issuos whIch stress the marginal benefits of dynam~ modelHng of
economic behavior in innovltlttive areas of analysis The discussion of a consumption tax emphasizes
the benefits of dynamic tusehold behavior and intergenerational aspects The study of the
flliminatlon or the re~intToduction of the investsnemt tax credits is made meaningful by the dynamic
modelling of production betlavor Modelling governmenl in a dynamic conlexts allows the modeller
to study the irrrpact of government deflcits Finally a dynamic framework lets us appreCiate the I
relevance of consumer expectations in terms of the evaluation of policy alternatives
i
i
51 Consumption Tax
t
By Its very nature a consumption tax can not be adequately investigated with a static model bull
Fullerion-Shoven-Whaicy (1983) use Ihe model of the US economy s described in
Ballard-Fullerton-Shoven-Whalley (1935) to evaluate the movement from the currenl US tax Isys1em to a progressive consumption tax Since their modeJ Incorporates a Jaborleisure choice
where leisure is an untaxed commodlty their results reflect the fact that both the consumption tax
and Ihe present system are dslortionary
Concentrating on the rntertemporaf distortions they show that sheltering more saving from lhe
current US income tax could improve econornic efficiency even if marginal tax rate increases are
necessary in order to maintain government fevenue~ At first you have a revenue shortfall
However the economy moves to a higher sustatned growth path and uitimatefy lower tax rates
cangenerate the same revenue path Also wages increase in the long run with this policy The
bull
bullbull ~
33 shyt
i
I
I
present value of welfare gi[lS for a potiey of complete savin9 deduction wilh marginal tate
adjustments (consumption lax) is simulated to be around $500 billion to $600 billion 01 1973
dollars
They also investiga1e th 1ngth of ttme it takes th~ economy to adjust to these policy changes
Roughly they estimate the long run to b thirty years although this figure Is very sensitive to the
specification of the savings elasticity - a crucial parameler in this model
52 Investment T~x Credit
Goulder-$ummers (1 S87) address the impact of eliminating the lnvestmenl Tax Cradit (ITC) on
intersectoral capital formation and on economic growth Their model isYa1ticularly adequate to
addies$ this issu~ in that is postulates a forward looking investment m9we with adjustment costs
They show that the eliminating the ITC causes a reduction in the rale of investment Inveslment is
estimated to fall by about 7 in the short-run and by about 12 in Ihe new long run steady state
In lurn a previous policy anncuncement lowers Ihe overall al1ractiveness of Investment and leads 10
a downward shift of the Investrrent profile On the other hand the combined effect of a revenue
neutral simultaneous efimiraron of the ITC and reduction of the corporate tax rates is a long run
reduction ef the capital Sieck by 35 This pattern suggests that a revenue neutral increament in
both Ihe corporale tax rates and the inveSlment lax credits would be preferable 10 the formula
adopted in Ihe US Tax Reform Act of 1986
Pereira (1987d) addresses the impact of the ITe from Ihe oland point of Ihe curTenllax syslem
under the Tax Reform Act of 1986 Given Ihe concern over the potential depressiVe effects upon
savings and investment of the new tax system in conjunction with deficit reduC1ion mechanisms of
the GrammmiddotRudman type a pertinent question is should Ih lTC be e-introduced In the context
bull
34
-of his mod~1 of the US economy Pereira focuses on the trade-off between the distortion in the
intersectoral allocation of capital induced by the lTC the lower tax revenues it generates and
potential financial crowding-out effects on one hand and the positive effects of lowering the relative
price of new capital goods on the other hand Preliminary results confirm the qualitative results in
Goulder-Summers (1987) suggestting a welfare gain of about 2 of the present value of GNP in
the best scenario of absence of replacement taxes Furthermore preliminary results show an
important time pattern to welfare gains with the average benefits increasing the further you look
into the future This pattern is due to the presence of constraints to intersectoral mobility of capital
and inertia in the adjustment towards the optimal capital levels as reflected by the presence of
adjustment costs
I
I
bullAuerbach-Katlikaff (1gB) in the context of their intergenerational model of the US economy
consider the impact on savings capital formation and interest rates of deficit policies and balanced
budget increases in government consumption They conclude that deficit finance and government
consumption can significa-ntly crowd out capital formation and lower the welfare of future
generations However crowding out from deficit finance is a very slow process because it results
from increased government spending over potentially long horizons Also deficit policies may
substantially influence the longmiddotrun interest rates while leaving the short-run rates essentially
unaffected Finally and in opposition to the central role adjustment costs seem to play in both
53 Financial crowding outeff~cts of government deficits
Goulder-Summers and Pereiras models Auerbach-Kotlikoff suggest that the time path of interest
rates induced by a policy of deficit financing seems to be insensitive to the presence of adjustment
costs
bull
35
54 The role 01 consumer expectatlons
The expectations analysis has uncovered one of the benefits of dynamic modelling
Ballard-Goulder (1985) find that the appeal of adopting a consumption tax depends on the level of
foresight possessed by the consumers Furthermore additional foresight may be welfare worsening
This is a second best type of result This tends 10 occur under policies that lead to capital deepening
and declining rate of return to capital over time To the extent that consumers have more foresight
they will be beller equippod to anticipate the fall in the rental price of capital Consequently if a
saving incentive is enacted people save more with myopia than with perfect foresight Given the
exi~tence of taxes on capita and the discrepancy between the private an~--sccial returns to capital
the greater savings with myopia is beller socially Ballard-Goulder find that the welfare gains
from a consumption tax are reduced by about 10 when we move from myopia to farleet foresight
Therefore the attractivenness of adopting a consumption tax seems fairly [obust across these
foresight specifications
I
bull
36
6 Tbe power and weaknesses 01 dynamic modelling the example 01 corporale
tax Integrallon
The Issue of corporate tax integration allows us to sh~w the stre-ngths and weaknesses or the
dynamic approach Empirical evidence using CGE techniques indicates that depending on the precise
scheme of integration and the 1ax replacement methods integratiOn may have substantial effects In
Ihe work of FuliertonmiddotKingmiddotShoven-Whalley (1980 1981) total integration was found to yield an bull
annual static eHielency gain of $4 to $8 billion in 1973 dollars Simulated dynamic gains may be as
large as $695 billion or abcu 14 of the present value of tulUre consumption and leisure in the
US economy These gains result primarily from interindustry reallocalions of Investment and an
improved inter~emporal allocation of consumption
Mora recent work emphasizes hOw consumers asset portfolio decisions and firms finanCial
decisions affect the efficiency gains from integration Slemrod (1980) focusing on consumers
asset ponfoio deCisions finds static efficiency gains which are about twice -as Iarge as those
repOrted iiy Fulierton-King-ShovenWhalley (1961) FuUrlon-Gurdon (1983) focus on the
firms financial decisions They roporl efficiency gains of 6 of GNP from the elimination of the
tax distor1ions favoring debt However when they eliminate the corporate tax and repiace it with
increaSed persona income taxes addlHonal dls10t1ions are created in the optimal labot~leisure
decisions These distortions tend to dominate the analysis slJ(h that the overall effects of complete
integration are very modest Galpr-LuckmiddotTodr (1986) address simultaneously consumers
asset pOrlfollo decisions and the firms financial decisions They also find very modest efficiency
gains from integralion
The above results are important but may be severely biased There are severa aspects of
economic behavior and modelling crucial for the study of Income tax Integration whiCh have no been
i 37
captured innny of the above papers One aspect is the absence of government deficitsmiddot a balanced
budget is assumed It is true Ihal resource crowding OUI is caplured In these mOdels but financial
crowding out induced by government spending is nOI Second investment is not derived from
optimtztlcn behavior Investment behavior passively accomodates endogenous saving decisions As
a consequence the differential impacts of policies in the incentives to save and to invest are not
captured Aso full capital mobility across secors is assumed with instantaneous capital
adjustmenlS towards optimal levels This assumption rules qut different costs of capital across
seclors and therefore differentiated reactions to tax policies changes Finallyhe modelling of both
gov9rnmerlt deficits and of endogenous real and financial investment decisions necessitates the
~nsideraiion of a dynamilt framework and the introduction of financiaJ assets govemment bonds and
iprivate financial assets A dynamic framework also highlights the efficiency effects of Integration on r
ithe ptimal intertemporal decisions The above models are either sIalic (Slemrods and
GalpermiddotltJckemiddotTode(sj or a dynamic sequ~nce of otherwise Slatic model bullbull I Pereir a (1986b) develops a dynamic applied general equilibrium model bull to study the tbull
efficiency effects of integration on the growth and allocation of investmenl across sectors Special
Iattention is paid to the structural affeelS of resource and financial crowding oul induced by
gove mect spending and deficits and to the real and financial Investment reactions across industries
to both tax integration and flnancial crowding out His mode departs from previous work on income
lax Integraticn and for that matler from most 01 the CGIE literalure for tax policy evaluation in
several direcllons It encompasses an endogenous sequential equilibrium struelure founded on
dynamic behavior with flexible expectations Government deficits are optlmaUy determined
Investment decisions are tward looking and the resull of optimizing behavior Several financial
assetsmiddot public and privatemiddot are considered
Simulation resulls Indicale that the welfare gains from Integration under large deficits are at
32
best lJlodest when measured in terms of the GNP The elimination of the corporate tax and its I replacemenl by increased income lax rates yields long run benefits which are never larger than
2 of the presenl yalue ct telr consumption and leisure onder the previous tax regime and 1
under the current tax regime However the short run w~lare effects of full integration tend 10 be
very low and in in some scenarios even negative This is a flew intertemporar pattern of efficiency Ieffects which reflects an adjJsiment lag in the interindustry investment decisions due to the
existence of costs of adjustment On the of her hand partial integration achieved by excluding I dividends from the corporale lax base syslematically yields negalive efficlencyeffeclS In Ihe long Inm these can be as high as 3 of present value of the future value of consumption and leisure This
is a new second best effect suggesllng Ihat less than complete integration may have perverse I efficiency effects
The dynamic slructure of PerclrBs model snowing for an improved freatment of real investment
declsions and government deficts provides a potential setting for a more accurate measure of the
costs and benefits of inlegration The simulation resullS suggest lower benefits and an intertempora Ibull
pattern of growing benefits
Simulation results also clearly suggest robust negative effects from partial integration How
reliable is this result If dJvidends were deductible from the corporate lax base (as are interest
payments) the preferenlial Ireatmant of debl over equily would be ellminaled Corporallons should
be expecled 10 react by decreasing the optimal deWeqully rallo However corporale financial
gecisions are exogenously set In Pereiras model as in aft of the other models surveyed that go as far
as dealing with the issue
Also withdrawing dividends from Ihe corporate tax base is a way of eliminating Ihe double
taxation of dividends at belh Ihe corporate and personnal Income levels How will the optimal
aliocellon of household saving accomodate that change in Ihe relalive return to corporale equity
39 I
t However hOusehold portfolio decisions are exogenouly set in Pereiras model as in all of the other
models surveyeltj go as far as dealing with the issue
IIt is sate to say that the evaluation 01 the benefits 01 partial Integration as defined above are
sevrey underestima1eO Meanwhile the distortionSoenerated by the increase in the marginal
personal tax rates designed to ralse the revenue foregone by the dividend exclusion are fully
aCCltlunted for A good measure of the costs of integration together With a poor evaluation of the
benefils may very well be the reason why partial integration is Simulated 10 yield neg alive
eHikiency gains
On a different vein as most of Ihe model surveyed here Pereiras is a closed economy model It
is legitimate to wonder how the resullS would change If international capital flows were anowed
IThis almost certainly would affect financial crowding out since a good d~al of the capital inflows
could be used to finance public debt
For these various reasons the results should be treated as preliminary but they strongly I suggosl that the dynamic structure provides valuable new insights into the corporate tax integration SSU2
40
7 Concluding remarks I I I i
What has been acccmplished Ihis far The success of Ihe CGE research in laelding the challenge
of dynamic modelling can not be denied
I iGreat progress in the modelling of household behavior has been achieved Promising
developm~nts in the modeUing of production and government behavior have also been accomplished
Interesting innovations in the concept of equilibrium and computation techniques were discussed
The policy analysis was greatly enriched by considering transitional effects together with longer
run steadymiddotstate equilibrium paths generalized equal yield strategies accomodating different
financial crowding out impacts and generalized dynamic policy evaluation indicators
An important set of issues have been addresed by the models surveyed Traditional issues
focusing on the effects of capital taxation and consumption taxes have been pursued In terms of
capital taxation for example the intertemporal nature of the issue was enhanced by allowing an
optimal evolution of capital stock in the economy and forward looking optimal investment decisions
In Andersson (1987) GouldermiddotSummers (1987) and Pereira (1986b 1987d) this is coupled
with a improved treatment of several tax provisions like investment tax credits and depreciation
allowances In terms of the consumption taxes developments in the specification of Intertemporal
household behavior the introduction of overlapping generations and the modelling of
intergenerational links via bequest motives as well as a closer attention to demographic evolution
provided a much improved economic setting for the understanding of the several aspects of the
problem
Furthermore dynamic modelling has permitted the addressing of a set of new issues Policy
Issues ranging from dynamic tax policy analysis to the evaluation of exogenous changes in
government expenditures to the impact of different methods of financing government expenditure to
4 1
the importance of tinanciai croHrii1g out to 1he relevance of consumorS expectations in tne policy
results have been addressed
Despile all of the progress and in part due to such progress several avenues are wide open for
much neded additional research The following seem to be the most Interesting and promising
areasmiddot First the specincation of liquidity constraints to household behavior may help to obtain a
better understanding of policy chenges that affect directly the interest rates in the economy Second
major attention needs to be paid to the incorporafion of a roraign sector (with open capital markets)
Third some efforts are needed in terms of finding adequate computation techniques now that
dynamic modelling eally makes the curse of dimensionality worse than eVOrbull
The modelling of financial markets is the single most unsatisfactory speet of the dynamic
models surveyed here~ The problems associated with provlding a mo~ adequate treatment of
financial markets and in general endogenous and optimal financial deci$ions bull both at the corporate
and at the personal levelsmiddot have 10 be addressed That necessarily requires Introducing uncertainty
into the models To b adequate this must go well beyond the mere assumption of point expeetalions
with the whole array of modelling implementalion and evaluation problems thai generates
bull I i
42
REFERENCES
Abel A and O Blanchard 1983 An Intertemporal Model of Saving and Interest Econometrica
51 pp 675-92
~ Andersson K 1987 Tar-ation of Capital in Sweden - A General Equilibrium Model Ph D
Dissenation UnIversity of Lund Sweder
Auerbach A L KotlikoH 1983 National Savings Economic Welfare and the Structure of
Taxation in M Feldstein ed Behavioral Simulation Methods in Tax Policy- Analysis Chicago
University of Chicago Press
Auerbach A l Kotlikoff 1984 Social Security and the Economics of Demographic-
Transition in H Aaron and G Burtless eds Retirement and Economic Behavior Washington DC
Brookings Institution
Auerbach A L Kotlikoff 1987 Dynamic Fiscal Policy Cambridge Cambridge University
Press
Auerbach A L Kotlikof and J Skinner 1983 The Efficiency Gains from Dynamic Tax
Reform Internatiooaj Economic Reyiew 24 pp 81middot100
Ballard C 1983 Evaluation of the Consumption Tax with Dynamic General Equilibrium
Models Ph D Dissertation Stanford University
Ballard C 1987 Tax Policy and Consumer Foresight A General Equilibrium Simulation
Study Economic Inquiry 25 pp 267-284
Ballard C and L Goulder 1985 Consumption Taxes ForeSight and Welfare a Computatonal
General Equilibrium Analysis in J Piggott and J Whalley eds New Deyelooments in Apolied
General Equilibrium Analysis Cambridge Cambridge University Press
Ballard C D Fullerton J Shoven and J Whalley 1985 A General Equilibrium Model for
t 43
Tax Eoliey Eyalua~iQO NBER University or Chicago Press
Borges A (1987) bull A Survey of Compula1ional General Equilibnum Models CECD Economic
Slud~ 8
Bovenbig L 19B3 imperfect Capital Mobility in a Perfect Foresight Herberger Model
Mimea UC Br~eley
Bovenberg L 1984 Capitel Accumulation and Capital Immobility Qmiddottheory In a Dynamic General Equilibrium Framework PhD dissertation University of California at Berkeley t
Bovenberg L 19850 Dynamic General Equilibrium Tax Models with Adjustment Costs in A iManne ed bull 1985 EtQnomrc ECUjilbrium Model Formulation and Solution Mathematical
lProgramming Siudy 23 Arnsterdam NorthmiddotHoliand
Bovenberg L t 1986 Capital Income Taxation in Growing Open Eeondmies Journa of Pubtic
Eamp(omlcs 3t pp 347middot376 IBovenberg L ald W Keller 1981 Dynamics and Nonlinearlties in Applied General
Equilibrium Models Internal Report Mlmiddot81middot208 Department for Statistical Methods CBS bull
Vooiburg The Netherlands
Charnley C 1981 The Welfare Cost of C~pltal income Taxation In a Growing Economy
IJournal of Political ECQnQm~ 89 3 pp 468middot496 I
Charnley C 19810 Efficient Tax Reform in a Dynamic Model of General Equilibrium I Mimeo Yale University I Charnley C 1982 On the Optimal Taxation In Stylized Models 01 Inlartemporal General
Equilibrium Mimeo Yale University I Decaluw B and A Manens 1985 Pays en Oeveloppement at Modeles Calculables OEquilibre
General - Une Revue de al litterature empirique Monographie Sene 9 Volume 3 Centre de
Recherche et Developpemont Economlque Unrslt de Montreal
44
bull
I I
I
I
1
Erlich $ V Ginsburgh and L van der Heyden 1987 Wher~ do Real Wage Policies Lead
Belgium bull A General Equilibrium Analysis European Economic Reyiew (forthcoming)
air R and J Taylor 1983 Solution and Maximum Likelihood Estimation of Dynamic
Nonlinear Ratioant Expeclations Models Ecooometrica 51 pp 1169middot1186
Feltenstein A 1984 Money and Bonds In a Disaggregated Ec~nonYin Searl 1-1 J Shoven
Eds Aoolied ~eoeral EqJiliDCiulD tioalYsis Cambridge University Press Cambridge
Feltenstein A 19B6 A Dynamic General Equllibrlu Analysis of Financial Crowding Out
Theory with an Application to Australia Journal of Public Economics 31 No1
Fullerlon D R Gorden 1983 A Reexamination of Tax Distorllons in General Equilibrium
Models in M Feldstein Ed Behayora Simulation Methods in Tax pOlicy Analysis Chicago
University of Chipgo Press
Fvllerlon D J Shaven and J Whalley 1983 Replacing the US Income Tax with a
Progressive Consumption Tax JoulOal of Public ECQnomics 20 pp3middot23
Fullerton D Y Henderson J Shaven 1984 A COmparision of Methodologies in Empirical
General Equilibrium Models of Taxation in Searl H J Shoven Eds ~pQlied Geoaral Jguilibrium
lIoalyss Cambridge University Press Cambridge
Fullerton 0 A King J Shaven and J Whaney 1981 Corporate Tax Integration in the United
Statesmiddot a General Equiiibdufi1 Approach American Economic Review 71 pp~ 677~691
Galper H R Lucke E Tader 1985 Taxation portlor Choice and the Allocation of Capital A
General Equilibrium Approach Mimeo
Gill p W Murray M Saunders M Wright 1986 Users Guide for NPSOL (Version 40) A
Fortran Package for Nonlinear Programming T echnicsl Report SOL 86middot2 Deparlment of
Operations Research Stanford University
Glnsbur~h V and L van der Heyden 1985 General Equilibrium with Wage Rigidities An
45
AppJication to Belgium in A Manne ed t Economic EQumprit1m~ Made Formulation and $oJioo
Mathematical Programming Study 23 Amsterdam NorthmiddotHolland ~
Goulet J 1968 1 Adjustrent CO$1S in the Theory o(the Firm Reyiew of EconomiC Studies 35 bull
pp47middot55
Goulder L 1985 Tax-Financed versus BondmiddotFinancedGhanges in Governent Spending
bull Mimeo Harva(d UniversilYmiddot
Goulder l J Shaven and J Whalley 1983 Domestic Tax Policy and the Foreign Sector The
importance of altemative foreign policy formulations to results from a general equilibrium tax
analysis model In ~n methods in Ibe antico of income from caoia1 ed Martin
Feldstein Chicago University of Chicago riSS
Goulder l L Summers 1987 Tax Policy Asset Prices and 3rowtly ~
A General Equilibrium Analysis NBER Working Pper No 212B
Grandmont J 1982 Temporary General Equilibrium Theory Ch 20 in Handbook 01
Mathemalical Eccnornics Vol II NorthmiddotHoliand
Harberger A 1959 The Corporate Income Tax An empirical Appraisal In lax Revision
Comoendjum Voll House Cornmitee on Ways and Means Washington DC Govemment Printing
Office
Harberger A 1962 The Incidence of the Corporate Income Tax JOULn1 of Political
Economy 70 pp 215middot240
Harberger A 1964 Taxation Resource Aliocalrlon and Welfare fn lb Role of Dir bullbulll and
Indjresl Taxes io tbe Fede[al Reserve s~I~Ill Princeton Princeton Unlveresity Press
James J 1985 New Developments in the Applications of General Equilibrium models to
Economic History in Piggott aand Whalley Eds
Jorgenson D and K Yun 1984 Tax Policy and Capital Allocation Harvard Inslitute of
bull
I
I f
Economic Research WP 1107
Judd K 1965 ShonmiddotRun Analysis of Fiscal Policy in a Simple Perfect Foresight Model
Journal of poJiHcal 5CCJiW1l pp 298middot319 r
Upton D J Potbc J Sachsa nd L Summers 1982 Multiple Shooting in Rational
Expectations Models fuoometrlc 50 1329middot1333
Manne A ed 1985 fconQrnjc Eoullibrium Model FOrIDujaljQo and Solution Mathematical
Pogrammlng Study 23 Amsterdam NorthmiddotHolland
Manne A 1985 Or the Formulation and Solution 01 Economic Equilibrium Models In A
Manne ed ECQOQlli-Jr~ibmiddotum ~tQdel EQlJulailon and Solution Malhematical Programming
Study 23 Amsterdam NorthmiddotHolland
Mansur A and J Whalley 1984 Numerical Specification of Applied General Equilibrium
Models in Scarf H J Shoven Eds Aoplied General Eouillbrium AnaiysectIsect Cambridge Universily
Press Cambridge
Merrill 0 1972 Applications and Extensions of an Algorithm thaI Computes FixedmiddotPoints of
Cenain Upper Semimiddotcominuou Point to Set Mappings PhD Disserlation Unlvel$ity 01 Michigan
Negishi T 1960 Welfar Economics and Exislence I an Equnlbrlum for a Competitive
Economy MelroeCOOQ11ka 12 92middot97
Pereira A 1985 On the Existence and Uniqueness of Optimal Ou1put Path lor CRTS Firms in
Adjustment Costs Technologies Mimeo Stanford University
Pereira A 1985a Consumer Expectations and Ihe COSI 01 UllUty In an Intertemporat
Framework Mlmeo Stanford University
Pereira A 1985b Corporale Tax Integrallon and Government Deficits A Dynamic General
Equilibrium Anaiysls Mlmeo Stanford University
Pereir A 1987 On Ihe Computation of the Users Costs of Stocks Ecooomi (forthcoming)
47
Pereira A bull 1987b Equal Yield Alternatives and Government Oeficlts Mimeof Stanford
University
Pereira A 1987c A Dynamic Applied General ~uHlbrium Model for Tax Policy Evaluation
Comlletamp Docurnent2tion Mimeo University of California San Diego (in progress)
Pereira A 1987d Should the investment Tax Credit be Rei(ltroduced Mlmeo University
of California San Diego (ir progreSs)
Piggott J and J Whalley eds 1985 tJew QeyeQpments in Applied General EQuilibrium
bullenalysjs Cambridge Cambridge University Press
Preckel Pbull 1985 Alternative algorithms for computing economic equilibria In A Manne Ed
Radnor R 19amp3 Equllibrium under Uncertainty In K Arrow M Intrilligaor ods
I ibull
Robinson S 1 g86 Mullisectorai Models of Oeveloping CountrieS a Survey Oivision of
Agricullure and Resources UC Borkeley WP 401
Scarf H 1967 The approximation of fixed points of a continuous mapping ~MM Jouwal of
epplid Mathornallcs 15 pp 1328middot43
Scarf H with T Hansen 1973 00 the Computation of EconomiC Eguilibria New Haven Yale
University Press
Scarf H J Shoven Eds 1984 61l1llied Genera Elujlibrium 6nalyssect Cambridge University
~ Press Cambridge
Shoven J 1976 The Incidence and Efficiency Effects of Taxe on Income from Capital
Jurnal of ramie1 Economy 84 pp 1261-83
Shoven J 1983 Applied General Equilibrium Tax Modemng IMF Slaff PaoerS 30
Shoven J and LB Simon 1986 Share Repurchases and Acquisitions An Analysis of which
Firms Participate NBER Working Paper No 2243
bull
48
IShaven J and J Whalley 1972 ft General Equmbrium Calculation of lhe Effects of
Differential Taxation of Incoma from Capital in the US Jauroe of public Econgmics 1
pp281middot321 (
Shoven J bull and J Whalley 1973 General Equilibrium with Tax$s a computation procedure
and an existence pr00f ReviEhi of Economic Studies 40 pp 475middot490
Shoven J J Whalley 1977 Equal Yield Tax Alternatives A General Equilibrium
Computa~ui0nal Tfchnique J~HJrnal of PUblic EconomIcs 8 pp 211~224 bull
Shaven J J Whalley 1984 Applied GeneralmiddotEquilibrlum MOdels of Taxalion and
International Tr8de An Introduction and Survey Journal of Economic literature 23t pp
1007middot1051
Slemrod J 1980 A General Equilibrium Model of Capital Income Taxation Ph D bullDissertation Harvard Univnrsity
Slemrod J 1983 A General Equilibrium Model of Taxation with Endogenous Financial
Behavior in M Feldstein edbull Behavioral Simulation Methods In Tax Policy Analysis Chicago
UniVerSity of Chicago Press
Stone J 1985 Sequential Optimization and Complementarily Techniques for computing
Economic equilibrium in A Manne Ed
Summers L 1981 Taxation and Corporate Investment a q-Theory Approach BroQkinas
Summers L 1985 Taxation and the Size and ComposHion 01 the Capital Stock An Asset Price
Approach Mlmeo Harvard University
Whalley J 1975 A Geneal Equiloibrium Assessment of the 1973 United Kingdom tax
reform Economica 42 pp 139middot61
1JLTIllOS JIQll~)ltG PAPERS PUBLICADOS bull
I
I I
j
112 70 - LUCENA iJiCIIJo ~ IIEnvironrilent Honitoring and Orqanization Structure T (Setellibro 1987)
nI 71 - L~ENA Diogo A Note on the Representation of Information Stxueture (SeteJObro 1987)
nQ 72 - LUCENA D1CgO Envuorment 111)nitoring and Organization
Structure II (Setelilhro 1987)
ril 73 - AlmmE3 Antonio Pais e CAsPAR Yitor bull Tributacao Inoentivos e InvesUlllentos Analise Qualitativa noYellibro 19S7)
n2 74 - liAlltOSA rose Pedro Opt1lllal lIage R1gidity A SUggestedtlthodology to Teet tb Theory and an Application (Outubro 1987)
nI 75 - CABRAL L 11 B Three lIotes on SYJIIlletric Gemes witt itsYJIIlletric Equilibri (NoveJObro 1987)
n2 76 - LUCENlt D10gOTo Search or Not to Search (Fevereiro1967)
nQ 17 - SA Jorge asconcelos euroI A Theory of SynergyB (Fevereiro1987)
n2 76 - VILARES tanuei Jose Os Be IntenaMios IlIIportados ColIIO Factor de Produao (Julho 1987)
nQ 79 - SA J~rge VasCQncel03 e HoY to Copete and Co~unicate in tature Industrial Products_ (aio 1988)
n2 80 - ROIl Rafael Learning and Capacity Expansion in a New larket Under Uncertainty (Fevereiro 1988)
n2 81 PEREIRA Alfredo llarvilo Survey of Dynamic Computational General EquilibriUlll Hodel for Tax Policy Evaluation (Outubro 1987)
Qualquer in ormacao sabre o 1IIorlltlng Papers ja publlcados sera prestada pelo Secretariado de Apoio aos Docentes podendo os ~e3OSf
ser adquirido5 ne Secyo de Vlndas da Faculdade de EconollLia UNL na Travessa Esteviio Pinto Cltpolide - 1000 LISllOA
29
Shortmiddotrun temporary equilibrium which makes up the intertemporal equilibrium path The
equilibrium conditions are seen as nonlinear equality constraints in the minimization of an artlcial
objective function The prices are normalized to the unit simplex by an additional linear equality I I 1 bull constraint
II ErlichmiddotGinsburgh-Heyden (1987) follow a unique approach in that they use a variant of the
optimization technique introduced by Negishl (1960) The economic equilibrium can be generated i
as a solution of a mathemallcal program Ihe objective function of which is a weighted sum of the
utility functions of the various agents while tho constraints set consists of the market dearing 1
conditions Ginsburgh-Heyden (1985) have extended Negishis resulllO tho case of downward price
lrigidities I
Finally the paper by Jorgonson-Yun (1984) is also unique in tIlat it is the only I
econometrically estimated model Different blocks for the consumption and production Side of the
-
model are separately estimated to provide the necessary structural parameters i
I The diversity of computation techniques is yet another indicator of the exploratory nature of the
body of literature surveyed in this article
43 Equal yield comparisions
The link between a base case and counteriactual simulations Is usually provided by the concept of
equal yield Shaven-Whaley (1977) discuss the meaning 01 equal yield in a general equilibrium
context when government is not allowed to run defICits Equal yield is interpreted to moan constant
public utility Government base case utility is maintained in the counferfactual experiments With
balanced budgets the concepl of equal yield is unambiguous The new equUibrium prices and the
balanced budge condition will del ermine the minimal expendilure and taxes needed to maintain base
30
case publi~ utility Accordingly in general equal yield is inconsistent with equal nominal tax
revenue Some change in tax revenue Is necessary Different tax replacement schemes ate
considered to assure that enough tax revenue is collected This is the approach essentially followed
by most 01 the papers surveyed Only Feltenstein (1985) Goulder (198) and Jorgenson-Yun
(1984) chose not to follow an equal yield strategy
The question is of how to Interprete lila concept of equal yield when the government is alowed to
fUn deflcLts The optimal leve of expenditure for base ease public trliUty can now be taAinanced
bond financed or financed by a mix of bonds and taxation We have several versions of equal yield In
particular equal yield may now be consistent with equal tax revenUE Furthermore some meaSure
ltgtf financial crowding out effects of government deficns can be inferred from the comparislon of the
several equal yield alternatives These issues are extensively discussed ~Pereira (1987b)
44 Price expectations and the dynamic generalization of compensation
Indicators
The sum of equivalent and compensation variations over households is the most widely used
aggregated measure of efficiency gains or losses In the context of $teady~state eomparisions andor
when complete future markets exist andlor perfect foresighl is assumed there are no difficulties
associated with the use of the standard Hicksian indicators In fact correct future prices are known
in these cases
If expectations are not selfmiddot fulfilling andlor future markels are not available a dynamic
generalization is necessary 8allardmiddotGoulder (1985) provide some steps in that direction by
defining an indicator that accomodates periods far into the future when households do not have
perfect foresight but a steady~state prevails On he other hand this issue is more fundamental in
i
[
I
i I I I I [
I I
bullI
31
the contex~ of a Gner1 lemporary equilibrium ftamawor~ In such Circumstances Pereira
(1986a) dev~lops a cnamlc generalization which is obtained as the present discounted value of a i i
sequence of short-fUn optimal expenditure functions consistent with a base case expected future I stream of IlWities I
32
5 Empirical evidenee from selected policy Issues
Dynamic tax models have generated several important results which escaped Sialic modellers
The following is a seleCpoundd set of issuos whIch stress the marginal benefits of dynam~ modelHng of
economic behavior in innovltlttive areas of analysis The discussion of a consumption tax emphasizes
the benefits of dynamic tusehold behavior and intergenerational aspects The study of the
flliminatlon or the re~intToduction of the investsnemt tax credits is made meaningful by the dynamic
modelling of production betlavor Modelling governmenl in a dynamic conlexts allows the modeller
to study the irrrpact of government deflcits Finally a dynamic framework lets us appreCiate the I
relevance of consumer expectations in terms of the evaluation of policy alternatives
i
i
51 Consumption Tax
t
By Its very nature a consumption tax can not be adequately investigated with a static model bull
Fullerion-Shoven-Whaicy (1983) use Ihe model of the US economy s described in
Ballard-Fullerton-Shoven-Whalley (1935) to evaluate the movement from the currenl US tax Isys1em to a progressive consumption tax Since their modeJ Incorporates a Jaborleisure choice
where leisure is an untaxed commodlty their results reflect the fact that both the consumption tax
and Ihe present system are dslortionary
Concentrating on the rntertemporaf distortions they show that sheltering more saving from lhe
current US income tax could improve econornic efficiency even if marginal tax rate increases are
necessary in order to maintain government fevenue~ At first you have a revenue shortfall
However the economy moves to a higher sustatned growth path and uitimatefy lower tax rates
cangenerate the same revenue path Also wages increase in the long run with this policy The
bull
bullbull ~
33 shyt
i
I
I
present value of welfare gi[lS for a potiey of complete savin9 deduction wilh marginal tate
adjustments (consumption lax) is simulated to be around $500 billion to $600 billion 01 1973
dollars
They also investiga1e th 1ngth of ttme it takes th~ economy to adjust to these policy changes
Roughly they estimate the long run to b thirty years although this figure Is very sensitive to the
specification of the savings elasticity - a crucial parameler in this model
52 Investment T~x Credit
Goulder-$ummers (1 S87) address the impact of eliminating the lnvestmenl Tax Cradit (ITC) on
intersectoral capital formation and on economic growth Their model isYa1ticularly adequate to
addies$ this issu~ in that is postulates a forward looking investment m9we with adjustment costs
They show that the eliminating the ITC causes a reduction in the rale of investment Inveslment is
estimated to fall by about 7 in the short-run and by about 12 in Ihe new long run steady state
In lurn a previous policy anncuncement lowers Ihe overall al1ractiveness of Investment and leads 10
a downward shift of the Investrrent profile On the other hand the combined effect of a revenue
neutral simultaneous efimiraron of the ITC and reduction of the corporate tax rates is a long run
reduction ef the capital Sieck by 35 This pattern suggests that a revenue neutral increament in
both Ihe corporale tax rates and the inveSlment lax credits would be preferable 10 the formula
adopted in Ihe US Tax Reform Act of 1986
Pereira (1987d) addresses the impact of the ITe from Ihe oland point of Ihe curTenllax syslem
under the Tax Reform Act of 1986 Given Ihe concern over the potential depressiVe effects upon
savings and investment of the new tax system in conjunction with deficit reduC1ion mechanisms of
the GrammmiddotRudman type a pertinent question is should Ih lTC be e-introduced In the context
bull
34
-of his mod~1 of the US economy Pereira focuses on the trade-off between the distortion in the
intersectoral allocation of capital induced by the lTC the lower tax revenues it generates and
potential financial crowding-out effects on one hand and the positive effects of lowering the relative
price of new capital goods on the other hand Preliminary results confirm the qualitative results in
Goulder-Summers (1987) suggestting a welfare gain of about 2 of the present value of GNP in
the best scenario of absence of replacement taxes Furthermore preliminary results show an
important time pattern to welfare gains with the average benefits increasing the further you look
into the future This pattern is due to the presence of constraints to intersectoral mobility of capital
and inertia in the adjustment towards the optimal capital levels as reflected by the presence of
adjustment costs
I
I
bullAuerbach-Katlikaff (1gB) in the context of their intergenerational model of the US economy
consider the impact on savings capital formation and interest rates of deficit policies and balanced
budget increases in government consumption They conclude that deficit finance and government
consumption can significa-ntly crowd out capital formation and lower the welfare of future
generations However crowding out from deficit finance is a very slow process because it results
from increased government spending over potentially long horizons Also deficit policies may
substantially influence the longmiddotrun interest rates while leaving the short-run rates essentially
unaffected Finally and in opposition to the central role adjustment costs seem to play in both
53 Financial crowding outeff~cts of government deficits
Goulder-Summers and Pereiras models Auerbach-Kotlikoff suggest that the time path of interest
rates induced by a policy of deficit financing seems to be insensitive to the presence of adjustment
costs
bull
35
54 The role 01 consumer expectatlons
The expectations analysis has uncovered one of the benefits of dynamic modelling
Ballard-Goulder (1985) find that the appeal of adopting a consumption tax depends on the level of
foresight possessed by the consumers Furthermore additional foresight may be welfare worsening
This is a second best type of result This tends 10 occur under policies that lead to capital deepening
and declining rate of return to capital over time To the extent that consumers have more foresight
they will be beller equippod to anticipate the fall in the rental price of capital Consequently if a
saving incentive is enacted people save more with myopia than with perfect foresight Given the
exi~tence of taxes on capita and the discrepancy between the private an~--sccial returns to capital
the greater savings with myopia is beller socially Ballard-Goulder find that the welfare gains
from a consumption tax are reduced by about 10 when we move from myopia to farleet foresight
Therefore the attractivenness of adopting a consumption tax seems fairly [obust across these
foresight specifications
I
bull
36
6 Tbe power and weaknesses 01 dynamic modelling the example 01 corporale
tax Integrallon
The Issue of corporate tax integration allows us to sh~w the stre-ngths and weaknesses or the
dynamic approach Empirical evidence using CGE techniques indicates that depending on the precise
scheme of integration and the 1ax replacement methods integratiOn may have substantial effects In
Ihe work of FuliertonmiddotKingmiddotShoven-Whalley (1980 1981) total integration was found to yield an bull
annual static eHielency gain of $4 to $8 billion in 1973 dollars Simulated dynamic gains may be as
large as $695 billion or abcu 14 of the present value of tulUre consumption and leisure in the
US economy These gains result primarily from interindustry reallocalions of Investment and an
improved inter~emporal allocation of consumption
Mora recent work emphasizes hOw consumers asset portfolio decisions and firms finanCial
decisions affect the efficiency gains from integration Slemrod (1980) focusing on consumers
asset ponfoio deCisions finds static efficiency gains which are about twice -as Iarge as those
repOrted iiy Fulierton-King-ShovenWhalley (1961) FuUrlon-Gurdon (1983) focus on the
firms financial decisions They roporl efficiency gains of 6 of GNP from the elimination of the
tax distor1ions favoring debt However when they eliminate the corporate tax and repiace it with
increaSed persona income taxes addlHonal dls10t1ions are created in the optimal labot~leisure
decisions These distortions tend to dominate the analysis slJ(h that the overall effects of complete
integration are very modest Galpr-LuckmiddotTodr (1986) address simultaneously consumers
asset pOrlfollo decisions and the firms financial decisions They also find very modest efficiency
gains from integralion
The above results are important but may be severely biased There are severa aspects of
economic behavior and modelling crucial for the study of Income tax Integration whiCh have no been
i 37
captured innny of the above papers One aspect is the absence of government deficitsmiddot a balanced
budget is assumed It is true Ihal resource crowding OUI is caplured In these mOdels but financial
crowding out induced by government spending is nOI Second investment is not derived from
optimtztlcn behavior Investment behavior passively accomodates endogenous saving decisions As
a consequence the differential impacts of policies in the incentives to save and to invest are not
captured Aso full capital mobility across secors is assumed with instantaneous capital
adjustmenlS towards optimal levels This assumption rules qut different costs of capital across
seclors and therefore differentiated reactions to tax policies changes Finallyhe modelling of both
gov9rnmerlt deficits and of endogenous real and financial investment decisions necessitates the
~nsideraiion of a dynamilt framework and the introduction of financiaJ assets govemment bonds and
iprivate financial assets A dynamic framework also highlights the efficiency effects of Integration on r
ithe ptimal intertemporal decisions The above models are either sIalic (Slemrods and
GalpermiddotltJckemiddotTode(sj or a dynamic sequ~nce of otherwise Slatic model bullbull I Pereir a (1986b) develops a dynamic applied general equilibrium model bull to study the tbull
efficiency effects of integration on the growth and allocation of investmenl across sectors Special
Iattention is paid to the structural affeelS of resource and financial crowding oul induced by
gove mect spending and deficits and to the real and financial Investment reactions across industries
to both tax integration and flnancial crowding out His mode departs from previous work on income
lax Integraticn and for that matler from most 01 the CGIE literalure for tax policy evaluation in
several direcllons It encompasses an endogenous sequential equilibrium struelure founded on
dynamic behavior with flexible expectations Government deficits are optlmaUy determined
Investment decisions are tward looking and the resull of optimizing behavior Several financial
assetsmiddot public and privatemiddot are considered
Simulation resulls Indicale that the welfare gains from Integration under large deficits are at
32
best lJlodest when measured in terms of the GNP The elimination of the corporate tax and its I replacemenl by increased income lax rates yields long run benefits which are never larger than
2 of the presenl yalue ct telr consumption and leisure onder the previous tax regime and 1
under the current tax regime However the short run w~lare effects of full integration tend 10 be
very low and in in some scenarios even negative This is a flew intertemporar pattern of efficiency Ieffects which reflects an adjJsiment lag in the interindustry investment decisions due to the
existence of costs of adjustment On the of her hand partial integration achieved by excluding I dividends from the corporale lax base syslematically yields negalive efficlencyeffeclS In Ihe long Inm these can be as high as 3 of present value of the future value of consumption and leisure This
is a new second best effect suggesllng Ihat less than complete integration may have perverse I efficiency effects
The dynamic slructure of PerclrBs model snowing for an improved freatment of real investment
declsions and government deficts provides a potential setting for a more accurate measure of the
costs and benefits of inlegration The simulation resullS suggest lower benefits and an intertempora Ibull
pattern of growing benefits
Simulation results also clearly suggest robust negative effects from partial integration How
reliable is this result If dJvidends were deductible from the corporate lax base (as are interest
payments) the preferenlial Ireatmant of debl over equily would be ellminaled Corporallons should
be expecled 10 react by decreasing the optimal deWeqully rallo However corporale financial
gecisions are exogenously set In Pereiras model as in aft of the other models surveyed that go as far
as dealing with the issue
Also withdrawing dividends from Ihe corporate tax base is a way of eliminating Ihe double
taxation of dividends at belh Ihe corporate and personnal Income levels How will the optimal
aliocellon of household saving accomodate that change in Ihe relalive return to corporale equity
39 I
t However hOusehold portfolio decisions are exogenouly set in Pereiras model as in all of the other
models surveyeltj go as far as dealing with the issue
IIt is sate to say that the evaluation 01 the benefits 01 partial Integration as defined above are
sevrey underestima1eO Meanwhile the distortionSoenerated by the increase in the marginal
personal tax rates designed to ralse the revenue foregone by the dividend exclusion are fully
aCCltlunted for A good measure of the costs of integration together With a poor evaluation of the
benefils may very well be the reason why partial integration is Simulated 10 yield neg alive
eHikiency gains
On a different vein as most of Ihe model surveyed here Pereiras is a closed economy model It
is legitimate to wonder how the resullS would change If international capital flows were anowed
IThis almost certainly would affect financial crowding out since a good d~al of the capital inflows
could be used to finance public debt
For these various reasons the results should be treated as preliminary but they strongly I suggosl that the dynamic structure provides valuable new insights into the corporate tax integration SSU2
40
7 Concluding remarks I I I i
What has been acccmplished Ihis far The success of Ihe CGE research in laelding the challenge
of dynamic modelling can not be denied
I iGreat progress in the modelling of household behavior has been achieved Promising
developm~nts in the modeUing of production and government behavior have also been accomplished
Interesting innovations in the concept of equilibrium and computation techniques were discussed
The policy analysis was greatly enriched by considering transitional effects together with longer
run steadymiddotstate equilibrium paths generalized equal yield strategies accomodating different
financial crowding out impacts and generalized dynamic policy evaluation indicators
An important set of issues have been addresed by the models surveyed Traditional issues
focusing on the effects of capital taxation and consumption taxes have been pursued In terms of
capital taxation for example the intertemporal nature of the issue was enhanced by allowing an
optimal evolution of capital stock in the economy and forward looking optimal investment decisions
In Andersson (1987) GouldermiddotSummers (1987) and Pereira (1986b 1987d) this is coupled
with a improved treatment of several tax provisions like investment tax credits and depreciation
allowances In terms of the consumption taxes developments in the specification of Intertemporal
household behavior the introduction of overlapping generations and the modelling of
intergenerational links via bequest motives as well as a closer attention to demographic evolution
provided a much improved economic setting for the understanding of the several aspects of the
problem
Furthermore dynamic modelling has permitted the addressing of a set of new issues Policy
Issues ranging from dynamic tax policy analysis to the evaluation of exogenous changes in
government expenditures to the impact of different methods of financing government expenditure to
4 1
the importance of tinanciai croHrii1g out to 1he relevance of consumorS expectations in tne policy
results have been addressed
Despile all of the progress and in part due to such progress several avenues are wide open for
much neded additional research The following seem to be the most Interesting and promising
areasmiddot First the specincation of liquidity constraints to household behavior may help to obtain a
better understanding of policy chenges that affect directly the interest rates in the economy Second
major attention needs to be paid to the incorporafion of a roraign sector (with open capital markets)
Third some efforts are needed in terms of finding adequate computation techniques now that
dynamic modelling eally makes the curse of dimensionality worse than eVOrbull
The modelling of financial markets is the single most unsatisfactory speet of the dynamic
models surveyed here~ The problems associated with provlding a mo~ adequate treatment of
financial markets and in general endogenous and optimal financial deci$ions bull both at the corporate
and at the personal levelsmiddot have 10 be addressed That necessarily requires Introducing uncertainty
into the models To b adequate this must go well beyond the mere assumption of point expeetalions
with the whole array of modelling implementalion and evaluation problems thai generates
bull I i
42
REFERENCES
Abel A and O Blanchard 1983 An Intertemporal Model of Saving and Interest Econometrica
51 pp 675-92
~ Andersson K 1987 Tar-ation of Capital in Sweden - A General Equilibrium Model Ph D
Dissenation UnIversity of Lund Sweder
Auerbach A L KotlikoH 1983 National Savings Economic Welfare and the Structure of
Taxation in M Feldstein ed Behavioral Simulation Methods in Tax Policy- Analysis Chicago
University of Chicago Press
Auerbach A l Kotlikoff 1984 Social Security and the Economics of Demographic-
Transition in H Aaron and G Burtless eds Retirement and Economic Behavior Washington DC
Brookings Institution
Auerbach A L Kotlikoff 1987 Dynamic Fiscal Policy Cambridge Cambridge University
Press
Auerbach A L Kotlikof and J Skinner 1983 The Efficiency Gains from Dynamic Tax
Reform Internatiooaj Economic Reyiew 24 pp 81middot100
Ballard C 1983 Evaluation of the Consumption Tax with Dynamic General Equilibrium
Models Ph D Dissertation Stanford University
Ballard C 1987 Tax Policy and Consumer Foresight A General Equilibrium Simulation
Study Economic Inquiry 25 pp 267-284
Ballard C and L Goulder 1985 Consumption Taxes ForeSight and Welfare a Computatonal
General Equilibrium Analysis in J Piggott and J Whalley eds New Deyelooments in Apolied
General Equilibrium Analysis Cambridge Cambridge University Press
Ballard C D Fullerton J Shoven and J Whalley 1985 A General Equilibrium Model for
t 43
Tax Eoliey Eyalua~iQO NBER University or Chicago Press
Borges A (1987) bull A Survey of Compula1ional General Equilibnum Models CECD Economic
Slud~ 8
Bovenbig L 19B3 imperfect Capital Mobility in a Perfect Foresight Herberger Model
Mimea UC Br~eley
Bovenberg L 1984 Capitel Accumulation and Capital Immobility Qmiddottheory In a Dynamic General Equilibrium Framework PhD dissertation University of California at Berkeley t
Bovenberg L 19850 Dynamic General Equilibrium Tax Models with Adjustment Costs in A iManne ed bull 1985 EtQnomrc ECUjilbrium Model Formulation and Solution Mathematical
lProgramming Siudy 23 Arnsterdam NorthmiddotHoliand
Bovenberg L t 1986 Capital Income Taxation in Growing Open Eeondmies Journa of Pubtic
Eamp(omlcs 3t pp 347middot376 IBovenberg L ald W Keller 1981 Dynamics and Nonlinearlties in Applied General
Equilibrium Models Internal Report Mlmiddot81middot208 Department for Statistical Methods CBS bull
Vooiburg The Netherlands
Charnley C 1981 The Welfare Cost of C~pltal income Taxation In a Growing Economy
IJournal of Political ECQnQm~ 89 3 pp 468middot496 I
Charnley C 19810 Efficient Tax Reform in a Dynamic Model of General Equilibrium I Mimeo Yale University I Charnley C 1982 On the Optimal Taxation In Stylized Models 01 Inlartemporal General
Equilibrium Mimeo Yale University I Decaluw B and A Manens 1985 Pays en Oeveloppement at Modeles Calculables OEquilibre
General - Une Revue de al litterature empirique Monographie Sene 9 Volume 3 Centre de
Recherche et Developpemont Economlque Unrslt de Montreal
44
bull
I I
I
I
1
Erlich $ V Ginsburgh and L van der Heyden 1987 Wher~ do Real Wage Policies Lead
Belgium bull A General Equilibrium Analysis European Economic Reyiew (forthcoming)
air R and J Taylor 1983 Solution and Maximum Likelihood Estimation of Dynamic
Nonlinear Ratioant Expeclations Models Ecooometrica 51 pp 1169middot1186
Feltenstein A 1984 Money and Bonds In a Disaggregated Ec~nonYin Searl 1-1 J Shoven
Eds Aoolied ~eoeral EqJiliDCiulD tioalYsis Cambridge University Press Cambridge
Feltenstein A 19B6 A Dynamic General Equllibrlu Analysis of Financial Crowding Out
Theory with an Application to Australia Journal of Public Economics 31 No1
Fullerlon D R Gorden 1983 A Reexamination of Tax Distorllons in General Equilibrium
Models in M Feldstein Ed Behayora Simulation Methods in Tax pOlicy Analysis Chicago
University of Chipgo Press
Fvllerlon D J Shaven and J Whalley 1983 Replacing the US Income Tax with a
Progressive Consumption Tax JoulOal of Public ECQnomics 20 pp3middot23
Fullerton D Y Henderson J Shaven 1984 A COmparision of Methodologies in Empirical
General Equilibrium Models of Taxation in Searl H J Shoven Eds ~pQlied Geoaral Jguilibrium
lIoalyss Cambridge University Press Cambridge
Fullerton 0 A King J Shaven and J Whaney 1981 Corporate Tax Integration in the United
Statesmiddot a General Equiiibdufi1 Approach American Economic Review 71 pp~ 677~691
Galper H R Lucke E Tader 1985 Taxation portlor Choice and the Allocation of Capital A
General Equilibrium Approach Mimeo
Gill p W Murray M Saunders M Wright 1986 Users Guide for NPSOL (Version 40) A
Fortran Package for Nonlinear Programming T echnicsl Report SOL 86middot2 Deparlment of
Operations Research Stanford University
Glnsbur~h V and L van der Heyden 1985 General Equilibrium with Wage Rigidities An
45
AppJication to Belgium in A Manne ed t Economic EQumprit1m~ Made Formulation and $oJioo
Mathematical Programming Study 23 Amsterdam NorthmiddotHolland ~
Goulet J 1968 1 Adjustrent CO$1S in the Theory o(the Firm Reyiew of EconomiC Studies 35 bull
pp47middot55
Goulder L 1985 Tax-Financed versus BondmiddotFinancedGhanges in Governent Spending
bull Mimeo Harva(d UniversilYmiddot
Goulder l J Shaven and J Whalley 1983 Domestic Tax Policy and the Foreign Sector The
importance of altemative foreign policy formulations to results from a general equilibrium tax
analysis model In ~n methods in Ibe antico of income from caoia1 ed Martin
Feldstein Chicago University of Chicago riSS
Goulder l L Summers 1987 Tax Policy Asset Prices and 3rowtly ~
A General Equilibrium Analysis NBER Working Pper No 212B
Grandmont J 1982 Temporary General Equilibrium Theory Ch 20 in Handbook 01
Mathemalical Eccnornics Vol II NorthmiddotHoliand
Harberger A 1959 The Corporate Income Tax An empirical Appraisal In lax Revision
Comoendjum Voll House Cornmitee on Ways and Means Washington DC Govemment Printing
Office
Harberger A 1962 The Incidence of the Corporate Income Tax JOULn1 of Political
Economy 70 pp 215middot240
Harberger A 1964 Taxation Resource Aliocalrlon and Welfare fn lb Role of Dir bullbulll and
Indjresl Taxes io tbe Fede[al Reserve s~I~Ill Princeton Princeton Unlveresity Press
James J 1985 New Developments in the Applications of General Equilibrium models to
Economic History in Piggott aand Whalley Eds
Jorgenson D and K Yun 1984 Tax Policy and Capital Allocation Harvard Inslitute of
bull
I
I f
Economic Research WP 1107
Judd K 1965 ShonmiddotRun Analysis of Fiscal Policy in a Simple Perfect Foresight Model
Journal of poJiHcal 5CCJiW1l pp 298middot319 r
Upton D J Potbc J Sachsa nd L Summers 1982 Multiple Shooting in Rational
Expectations Models fuoometrlc 50 1329middot1333
Manne A ed 1985 fconQrnjc Eoullibrium Model FOrIDujaljQo and Solution Mathematical
Pogrammlng Study 23 Amsterdam NorthmiddotHolland
Manne A 1985 Or the Formulation and Solution 01 Economic Equilibrium Models In A
Manne ed ECQOQlli-Jr~ibmiddotum ~tQdel EQlJulailon and Solution Malhematical Programming
Study 23 Amsterdam NorthmiddotHolland
Mansur A and J Whalley 1984 Numerical Specification of Applied General Equilibrium
Models in Scarf H J Shoven Eds Aoplied General Eouillbrium AnaiysectIsect Cambridge Universily
Press Cambridge
Merrill 0 1972 Applications and Extensions of an Algorithm thaI Computes FixedmiddotPoints of
Cenain Upper Semimiddotcominuou Point to Set Mappings PhD Disserlation Unlvel$ity 01 Michigan
Negishi T 1960 Welfar Economics and Exislence I an Equnlbrlum for a Competitive
Economy MelroeCOOQ11ka 12 92middot97
Pereira A 1985 On the Existence and Uniqueness of Optimal Ou1put Path lor CRTS Firms in
Adjustment Costs Technologies Mimeo Stanford University
Pereira A 1985a Consumer Expectations and Ihe COSI 01 UllUty In an Intertemporat
Framework Mlmeo Stanford University
Pereira A 1985b Corporale Tax Integrallon and Government Deficits A Dynamic General
Equilibrium Anaiysls Mlmeo Stanford University
Pereir A 1987 On Ihe Computation of the Users Costs of Stocks Ecooomi (forthcoming)
47
Pereira A bull 1987b Equal Yield Alternatives and Government Oeficlts Mimeof Stanford
University
Pereira A 1987c A Dynamic Applied General ~uHlbrium Model for Tax Policy Evaluation
Comlletamp Docurnent2tion Mimeo University of California San Diego (in progress)
Pereira A 1987d Should the investment Tax Credit be Rei(ltroduced Mlmeo University
of California San Diego (ir progreSs)
Piggott J and J Whalley eds 1985 tJew QeyeQpments in Applied General EQuilibrium
bullenalysjs Cambridge Cambridge University Press
Preckel Pbull 1985 Alternative algorithms for computing economic equilibria In A Manne Ed
Radnor R 19amp3 Equllibrium under Uncertainty In K Arrow M Intrilligaor ods
I ibull
Robinson S 1 g86 Mullisectorai Models of Oeveloping CountrieS a Survey Oivision of
Agricullure and Resources UC Borkeley WP 401
Scarf H 1967 The approximation of fixed points of a continuous mapping ~MM Jouwal of
epplid Mathornallcs 15 pp 1328middot43
Scarf H with T Hansen 1973 00 the Computation of EconomiC Eguilibria New Haven Yale
University Press
Scarf H J Shoven Eds 1984 61l1llied Genera Elujlibrium 6nalyssect Cambridge University
~ Press Cambridge
Shoven J 1976 The Incidence and Efficiency Effects of Taxe on Income from Capital
Jurnal of ramie1 Economy 84 pp 1261-83
Shoven J 1983 Applied General Equilibrium Tax Modemng IMF Slaff PaoerS 30
Shoven J and LB Simon 1986 Share Repurchases and Acquisitions An Analysis of which
Firms Participate NBER Working Paper No 2243
bull
48
IShaven J and J Whalley 1972 ft General Equmbrium Calculation of lhe Effects of
Differential Taxation of Incoma from Capital in the US Jauroe of public Econgmics 1
pp281middot321 (
Shoven J bull and J Whalley 1973 General Equilibrium with Tax$s a computation procedure
and an existence pr00f ReviEhi of Economic Studies 40 pp 475middot490
Shoven J J Whalley 1977 Equal Yield Tax Alternatives A General Equilibrium
Computa~ui0nal Tfchnique J~HJrnal of PUblic EconomIcs 8 pp 211~224 bull
Shaven J J Whalley 1984 Applied GeneralmiddotEquilibrlum MOdels of Taxalion and
International Tr8de An Introduction and Survey Journal of Economic literature 23t pp
1007middot1051
Slemrod J 1980 A General Equilibrium Model of Capital Income Taxation Ph D bullDissertation Harvard Univnrsity
Slemrod J 1983 A General Equilibrium Model of Taxation with Endogenous Financial
Behavior in M Feldstein edbull Behavioral Simulation Methods In Tax Policy Analysis Chicago
UniVerSity of Chicago Press
Stone J 1985 Sequential Optimization and Complementarily Techniques for computing
Economic equilibrium in A Manne Ed
Summers L 1981 Taxation and Corporate Investment a q-Theory Approach BroQkinas
Summers L 1985 Taxation and the Size and ComposHion 01 the Capital Stock An Asset Price
Approach Mlmeo Harvard University
Whalley J 1975 A Geneal Equiloibrium Assessment of the 1973 United Kingdom tax
reform Economica 42 pp 139middot61
1JLTIllOS JIQll~)ltG PAPERS PUBLICADOS bull
I
I I
j
112 70 - LUCENA iJiCIIJo ~ IIEnvironrilent Honitoring and Orqanization Structure T (Setellibro 1987)
nI 71 - L~ENA Diogo A Note on the Representation of Information Stxueture (SeteJObro 1987)
nQ 72 - LUCENA D1CgO Envuorment 111)nitoring and Organization
Structure II (Setelilhro 1987)
ril 73 - AlmmE3 Antonio Pais e CAsPAR Yitor bull Tributacao Inoentivos e InvesUlllentos Analise Qualitativa noYellibro 19S7)
n2 74 - liAlltOSA rose Pedro Opt1lllal lIage R1gidity A SUggestedtlthodology to Teet tb Theory and an Application (Outubro 1987)
nI 75 - CABRAL L 11 B Three lIotes on SYJIIlletric Gemes witt itsYJIIlletric Equilibri (NoveJObro 1987)
n2 76 - LUCENlt D10gOTo Search or Not to Search (Fevereiro1967)
nQ 17 - SA Jorge asconcelos euroI A Theory of SynergyB (Fevereiro1987)
n2 76 - VILARES tanuei Jose Os Be IntenaMios IlIIportados ColIIO Factor de Produao (Julho 1987)
nQ 79 - SA J~rge VasCQncel03 e HoY to Copete and Co~unicate in tature Industrial Products_ (aio 1988)
n2 80 - ROIl Rafael Learning and Capacity Expansion in a New larket Under Uncertainty (Fevereiro 1988)
n2 81 PEREIRA Alfredo llarvilo Survey of Dynamic Computational General EquilibriUlll Hodel for Tax Policy Evaluation (Outubro 1987)
Qualquer in ormacao sabre o 1IIorlltlng Papers ja publlcados sera prestada pelo Secretariado de Apoio aos Docentes podendo os ~e3OSf
ser adquirido5 ne Secyo de Vlndas da Faculdade de EconollLia UNL na Travessa Esteviio Pinto Cltpolide - 1000 LISllOA
30
case publi~ utility Accordingly in general equal yield is inconsistent with equal nominal tax
revenue Some change in tax revenue Is necessary Different tax replacement schemes ate
considered to assure that enough tax revenue is collected This is the approach essentially followed
by most 01 the papers surveyed Only Feltenstein (1985) Goulder (198) and Jorgenson-Yun
(1984) chose not to follow an equal yield strategy
The question is of how to Interprete lila concept of equal yield when the government is alowed to
fUn deflcLts The optimal leve of expenditure for base ease public trliUty can now be taAinanced
bond financed or financed by a mix of bonds and taxation We have several versions of equal yield In
particular equal yield may now be consistent with equal tax revenUE Furthermore some meaSure
ltgtf financial crowding out effects of government deficns can be inferred from the comparislon of the
several equal yield alternatives These issues are extensively discussed ~Pereira (1987b)
44 Price expectations and the dynamic generalization of compensation
Indicators
The sum of equivalent and compensation variations over households is the most widely used
aggregated measure of efficiency gains or losses In the context of $teady~state eomparisions andor
when complete future markets exist andlor perfect foresighl is assumed there are no difficulties
associated with the use of the standard Hicksian indicators In fact correct future prices are known
in these cases
If expectations are not selfmiddot fulfilling andlor future markels are not available a dynamic
generalization is necessary 8allardmiddotGoulder (1985) provide some steps in that direction by
defining an indicator that accomodates periods far into the future when households do not have
perfect foresight but a steady~state prevails On he other hand this issue is more fundamental in
i
[
I
i I I I I [
I I
bullI
31
the contex~ of a Gner1 lemporary equilibrium ftamawor~ In such Circumstances Pereira
(1986a) dev~lops a cnamlc generalization which is obtained as the present discounted value of a i i
sequence of short-fUn optimal expenditure functions consistent with a base case expected future I stream of IlWities I
32
5 Empirical evidenee from selected policy Issues
Dynamic tax models have generated several important results which escaped Sialic modellers
The following is a seleCpoundd set of issuos whIch stress the marginal benefits of dynam~ modelHng of
economic behavior in innovltlttive areas of analysis The discussion of a consumption tax emphasizes
the benefits of dynamic tusehold behavior and intergenerational aspects The study of the
flliminatlon or the re~intToduction of the investsnemt tax credits is made meaningful by the dynamic
modelling of production betlavor Modelling governmenl in a dynamic conlexts allows the modeller
to study the irrrpact of government deflcits Finally a dynamic framework lets us appreCiate the I
relevance of consumer expectations in terms of the evaluation of policy alternatives
i
i
51 Consumption Tax
t
By Its very nature a consumption tax can not be adequately investigated with a static model bull
Fullerion-Shoven-Whaicy (1983) use Ihe model of the US economy s described in
Ballard-Fullerton-Shoven-Whalley (1935) to evaluate the movement from the currenl US tax Isys1em to a progressive consumption tax Since their modeJ Incorporates a Jaborleisure choice
where leisure is an untaxed commodlty their results reflect the fact that both the consumption tax
and Ihe present system are dslortionary
Concentrating on the rntertemporaf distortions they show that sheltering more saving from lhe
current US income tax could improve econornic efficiency even if marginal tax rate increases are
necessary in order to maintain government fevenue~ At first you have a revenue shortfall
However the economy moves to a higher sustatned growth path and uitimatefy lower tax rates
cangenerate the same revenue path Also wages increase in the long run with this policy The
bull
bullbull ~
33 shyt
i
I
I
present value of welfare gi[lS for a potiey of complete savin9 deduction wilh marginal tate
adjustments (consumption lax) is simulated to be around $500 billion to $600 billion 01 1973
dollars
They also investiga1e th 1ngth of ttme it takes th~ economy to adjust to these policy changes
Roughly they estimate the long run to b thirty years although this figure Is very sensitive to the
specification of the savings elasticity - a crucial parameler in this model
52 Investment T~x Credit
Goulder-$ummers (1 S87) address the impact of eliminating the lnvestmenl Tax Cradit (ITC) on
intersectoral capital formation and on economic growth Their model isYa1ticularly adequate to
addies$ this issu~ in that is postulates a forward looking investment m9we with adjustment costs
They show that the eliminating the ITC causes a reduction in the rale of investment Inveslment is
estimated to fall by about 7 in the short-run and by about 12 in Ihe new long run steady state
In lurn a previous policy anncuncement lowers Ihe overall al1ractiveness of Investment and leads 10
a downward shift of the Investrrent profile On the other hand the combined effect of a revenue
neutral simultaneous efimiraron of the ITC and reduction of the corporate tax rates is a long run
reduction ef the capital Sieck by 35 This pattern suggests that a revenue neutral increament in
both Ihe corporale tax rates and the inveSlment lax credits would be preferable 10 the formula
adopted in Ihe US Tax Reform Act of 1986
Pereira (1987d) addresses the impact of the ITe from Ihe oland point of Ihe curTenllax syslem
under the Tax Reform Act of 1986 Given Ihe concern over the potential depressiVe effects upon
savings and investment of the new tax system in conjunction with deficit reduC1ion mechanisms of
the GrammmiddotRudman type a pertinent question is should Ih lTC be e-introduced In the context
bull
34
-of his mod~1 of the US economy Pereira focuses on the trade-off between the distortion in the
intersectoral allocation of capital induced by the lTC the lower tax revenues it generates and
potential financial crowding-out effects on one hand and the positive effects of lowering the relative
price of new capital goods on the other hand Preliminary results confirm the qualitative results in
Goulder-Summers (1987) suggestting a welfare gain of about 2 of the present value of GNP in
the best scenario of absence of replacement taxes Furthermore preliminary results show an
important time pattern to welfare gains with the average benefits increasing the further you look
into the future This pattern is due to the presence of constraints to intersectoral mobility of capital
and inertia in the adjustment towards the optimal capital levels as reflected by the presence of
adjustment costs
I
I
bullAuerbach-Katlikaff (1gB) in the context of their intergenerational model of the US economy
consider the impact on savings capital formation and interest rates of deficit policies and balanced
budget increases in government consumption They conclude that deficit finance and government
consumption can significa-ntly crowd out capital formation and lower the welfare of future
generations However crowding out from deficit finance is a very slow process because it results
from increased government spending over potentially long horizons Also deficit policies may
substantially influence the longmiddotrun interest rates while leaving the short-run rates essentially
unaffected Finally and in opposition to the central role adjustment costs seem to play in both
53 Financial crowding outeff~cts of government deficits
Goulder-Summers and Pereiras models Auerbach-Kotlikoff suggest that the time path of interest
rates induced by a policy of deficit financing seems to be insensitive to the presence of adjustment
costs
bull
35
54 The role 01 consumer expectatlons
The expectations analysis has uncovered one of the benefits of dynamic modelling
Ballard-Goulder (1985) find that the appeal of adopting a consumption tax depends on the level of
foresight possessed by the consumers Furthermore additional foresight may be welfare worsening
This is a second best type of result This tends 10 occur under policies that lead to capital deepening
and declining rate of return to capital over time To the extent that consumers have more foresight
they will be beller equippod to anticipate the fall in the rental price of capital Consequently if a
saving incentive is enacted people save more with myopia than with perfect foresight Given the
exi~tence of taxes on capita and the discrepancy between the private an~--sccial returns to capital
the greater savings with myopia is beller socially Ballard-Goulder find that the welfare gains
from a consumption tax are reduced by about 10 when we move from myopia to farleet foresight
Therefore the attractivenness of adopting a consumption tax seems fairly [obust across these
foresight specifications
I
bull
36
6 Tbe power and weaknesses 01 dynamic modelling the example 01 corporale
tax Integrallon
The Issue of corporate tax integration allows us to sh~w the stre-ngths and weaknesses or the
dynamic approach Empirical evidence using CGE techniques indicates that depending on the precise
scheme of integration and the 1ax replacement methods integratiOn may have substantial effects In
Ihe work of FuliertonmiddotKingmiddotShoven-Whalley (1980 1981) total integration was found to yield an bull
annual static eHielency gain of $4 to $8 billion in 1973 dollars Simulated dynamic gains may be as
large as $695 billion or abcu 14 of the present value of tulUre consumption and leisure in the
US economy These gains result primarily from interindustry reallocalions of Investment and an
improved inter~emporal allocation of consumption
Mora recent work emphasizes hOw consumers asset portfolio decisions and firms finanCial
decisions affect the efficiency gains from integration Slemrod (1980) focusing on consumers
asset ponfoio deCisions finds static efficiency gains which are about twice -as Iarge as those
repOrted iiy Fulierton-King-ShovenWhalley (1961) FuUrlon-Gurdon (1983) focus on the
firms financial decisions They roporl efficiency gains of 6 of GNP from the elimination of the
tax distor1ions favoring debt However when they eliminate the corporate tax and repiace it with
increaSed persona income taxes addlHonal dls10t1ions are created in the optimal labot~leisure
decisions These distortions tend to dominate the analysis slJ(h that the overall effects of complete
integration are very modest Galpr-LuckmiddotTodr (1986) address simultaneously consumers
asset pOrlfollo decisions and the firms financial decisions They also find very modest efficiency
gains from integralion
The above results are important but may be severely biased There are severa aspects of
economic behavior and modelling crucial for the study of Income tax Integration whiCh have no been
i 37
captured innny of the above papers One aspect is the absence of government deficitsmiddot a balanced
budget is assumed It is true Ihal resource crowding OUI is caplured In these mOdels but financial
crowding out induced by government spending is nOI Second investment is not derived from
optimtztlcn behavior Investment behavior passively accomodates endogenous saving decisions As
a consequence the differential impacts of policies in the incentives to save and to invest are not
captured Aso full capital mobility across secors is assumed with instantaneous capital
adjustmenlS towards optimal levels This assumption rules qut different costs of capital across
seclors and therefore differentiated reactions to tax policies changes Finallyhe modelling of both
gov9rnmerlt deficits and of endogenous real and financial investment decisions necessitates the
~nsideraiion of a dynamilt framework and the introduction of financiaJ assets govemment bonds and
iprivate financial assets A dynamic framework also highlights the efficiency effects of Integration on r
ithe ptimal intertemporal decisions The above models are either sIalic (Slemrods and
GalpermiddotltJckemiddotTode(sj or a dynamic sequ~nce of otherwise Slatic model bullbull I Pereir a (1986b) develops a dynamic applied general equilibrium model bull to study the tbull
efficiency effects of integration on the growth and allocation of investmenl across sectors Special
Iattention is paid to the structural affeelS of resource and financial crowding oul induced by
gove mect spending and deficits and to the real and financial Investment reactions across industries
to both tax integration and flnancial crowding out His mode departs from previous work on income
lax Integraticn and for that matler from most 01 the CGIE literalure for tax policy evaluation in
several direcllons It encompasses an endogenous sequential equilibrium struelure founded on
dynamic behavior with flexible expectations Government deficits are optlmaUy determined
Investment decisions are tward looking and the resull of optimizing behavior Several financial
assetsmiddot public and privatemiddot are considered
Simulation resulls Indicale that the welfare gains from Integration under large deficits are at
32
best lJlodest when measured in terms of the GNP The elimination of the corporate tax and its I replacemenl by increased income lax rates yields long run benefits which are never larger than
2 of the presenl yalue ct telr consumption and leisure onder the previous tax regime and 1
under the current tax regime However the short run w~lare effects of full integration tend 10 be
very low and in in some scenarios even negative This is a flew intertemporar pattern of efficiency Ieffects which reflects an adjJsiment lag in the interindustry investment decisions due to the
existence of costs of adjustment On the of her hand partial integration achieved by excluding I dividends from the corporale lax base syslematically yields negalive efficlencyeffeclS In Ihe long Inm these can be as high as 3 of present value of the future value of consumption and leisure This
is a new second best effect suggesllng Ihat less than complete integration may have perverse I efficiency effects
The dynamic slructure of PerclrBs model snowing for an improved freatment of real investment
declsions and government deficts provides a potential setting for a more accurate measure of the
costs and benefits of inlegration The simulation resullS suggest lower benefits and an intertempora Ibull
pattern of growing benefits
Simulation results also clearly suggest robust negative effects from partial integration How
reliable is this result If dJvidends were deductible from the corporate lax base (as are interest
payments) the preferenlial Ireatmant of debl over equily would be ellminaled Corporallons should
be expecled 10 react by decreasing the optimal deWeqully rallo However corporale financial
gecisions are exogenously set In Pereiras model as in aft of the other models surveyed that go as far
as dealing with the issue
Also withdrawing dividends from Ihe corporate tax base is a way of eliminating Ihe double
taxation of dividends at belh Ihe corporate and personnal Income levels How will the optimal
aliocellon of household saving accomodate that change in Ihe relalive return to corporale equity
39 I
t However hOusehold portfolio decisions are exogenouly set in Pereiras model as in all of the other
models surveyeltj go as far as dealing with the issue
IIt is sate to say that the evaluation 01 the benefits 01 partial Integration as defined above are
sevrey underestima1eO Meanwhile the distortionSoenerated by the increase in the marginal
personal tax rates designed to ralse the revenue foregone by the dividend exclusion are fully
aCCltlunted for A good measure of the costs of integration together With a poor evaluation of the
benefils may very well be the reason why partial integration is Simulated 10 yield neg alive
eHikiency gains
On a different vein as most of Ihe model surveyed here Pereiras is a closed economy model It
is legitimate to wonder how the resullS would change If international capital flows were anowed
IThis almost certainly would affect financial crowding out since a good d~al of the capital inflows
could be used to finance public debt
For these various reasons the results should be treated as preliminary but they strongly I suggosl that the dynamic structure provides valuable new insights into the corporate tax integration SSU2
40
7 Concluding remarks I I I i
What has been acccmplished Ihis far The success of Ihe CGE research in laelding the challenge
of dynamic modelling can not be denied
I iGreat progress in the modelling of household behavior has been achieved Promising
developm~nts in the modeUing of production and government behavior have also been accomplished
Interesting innovations in the concept of equilibrium and computation techniques were discussed
The policy analysis was greatly enriched by considering transitional effects together with longer
run steadymiddotstate equilibrium paths generalized equal yield strategies accomodating different
financial crowding out impacts and generalized dynamic policy evaluation indicators
An important set of issues have been addresed by the models surveyed Traditional issues
focusing on the effects of capital taxation and consumption taxes have been pursued In terms of
capital taxation for example the intertemporal nature of the issue was enhanced by allowing an
optimal evolution of capital stock in the economy and forward looking optimal investment decisions
In Andersson (1987) GouldermiddotSummers (1987) and Pereira (1986b 1987d) this is coupled
with a improved treatment of several tax provisions like investment tax credits and depreciation
allowances In terms of the consumption taxes developments in the specification of Intertemporal
household behavior the introduction of overlapping generations and the modelling of
intergenerational links via bequest motives as well as a closer attention to demographic evolution
provided a much improved economic setting for the understanding of the several aspects of the
problem
Furthermore dynamic modelling has permitted the addressing of a set of new issues Policy
Issues ranging from dynamic tax policy analysis to the evaluation of exogenous changes in
government expenditures to the impact of different methods of financing government expenditure to
4 1
the importance of tinanciai croHrii1g out to 1he relevance of consumorS expectations in tne policy
results have been addressed
Despile all of the progress and in part due to such progress several avenues are wide open for
much neded additional research The following seem to be the most Interesting and promising
areasmiddot First the specincation of liquidity constraints to household behavior may help to obtain a
better understanding of policy chenges that affect directly the interest rates in the economy Second
major attention needs to be paid to the incorporafion of a roraign sector (with open capital markets)
Third some efforts are needed in terms of finding adequate computation techniques now that
dynamic modelling eally makes the curse of dimensionality worse than eVOrbull
The modelling of financial markets is the single most unsatisfactory speet of the dynamic
models surveyed here~ The problems associated with provlding a mo~ adequate treatment of
financial markets and in general endogenous and optimal financial deci$ions bull both at the corporate
and at the personal levelsmiddot have 10 be addressed That necessarily requires Introducing uncertainty
into the models To b adequate this must go well beyond the mere assumption of point expeetalions
with the whole array of modelling implementalion and evaluation problems thai generates
bull I i
42
REFERENCES
Abel A and O Blanchard 1983 An Intertemporal Model of Saving and Interest Econometrica
51 pp 675-92
~ Andersson K 1987 Tar-ation of Capital in Sweden - A General Equilibrium Model Ph D
Dissenation UnIversity of Lund Sweder
Auerbach A L KotlikoH 1983 National Savings Economic Welfare and the Structure of
Taxation in M Feldstein ed Behavioral Simulation Methods in Tax Policy- Analysis Chicago
University of Chicago Press
Auerbach A l Kotlikoff 1984 Social Security and the Economics of Demographic-
Transition in H Aaron and G Burtless eds Retirement and Economic Behavior Washington DC
Brookings Institution
Auerbach A L Kotlikoff 1987 Dynamic Fiscal Policy Cambridge Cambridge University
Press
Auerbach A L Kotlikof and J Skinner 1983 The Efficiency Gains from Dynamic Tax
Reform Internatiooaj Economic Reyiew 24 pp 81middot100
Ballard C 1983 Evaluation of the Consumption Tax with Dynamic General Equilibrium
Models Ph D Dissertation Stanford University
Ballard C 1987 Tax Policy and Consumer Foresight A General Equilibrium Simulation
Study Economic Inquiry 25 pp 267-284
Ballard C and L Goulder 1985 Consumption Taxes ForeSight and Welfare a Computatonal
General Equilibrium Analysis in J Piggott and J Whalley eds New Deyelooments in Apolied
General Equilibrium Analysis Cambridge Cambridge University Press
Ballard C D Fullerton J Shoven and J Whalley 1985 A General Equilibrium Model for
t 43
Tax Eoliey Eyalua~iQO NBER University or Chicago Press
Borges A (1987) bull A Survey of Compula1ional General Equilibnum Models CECD Economic
Slud~ 8
Bovenbig L 19B3 imperfect Capital Mobility in a Perfect Foresight Herberger Model
Mimea UC Br~eley
Bovenberg L 1984 Capitel Accumulation and Capital Immobility Qmiddottheory In a Dynamic General Equilibrium Framework PhD dissertation University of California at Berkeley t
Bovenberg L 19850 Dynamic General Equilibrium Tax Models with Adjustment Costs in A iManne ed bull 1985 EtQnomrc ECUjilbrium Model Formulation and Solution Mathematical
lProgramming Siudy 23 Arnsterdam NorthmiddotHoliand
Bovenberg L t 1986 Capital Income Taxation in Growing Open Eeondmies Journa of Pubtic
Eamp(omlcs 3t pp 347middot376 IBovenberg L ald W Keller 1981 Dynamics and Nonlinearlties in Applied General
Equilibrium Models Internal Report Mlmiddot81middot208 Department for Statistical Methods CBS bull
Vooiburg The Netherlands
Charnley C 1981 The Welfare Cost of C~pltal income Taxation In a Growing Economy
IJournal of Political ECQnQm~ 89 3 pp 468middot496 I
Charnley C 19810 Efficient Tax Reform in a Dynamic Model of General Equilibrium I Mimeo Yale University I Charnley C 1982 On the Optimal Taxation In Stylized Models 01 Inlartemporal General
Equilibrium Mimeo Yale University I Decaluw B and A Manens 1985 Pays en Oeveloppement at Modeles Calculables OEquilibre
General - Une Revue de al litterature empirique Monographie Sene 9 Volume 3 Centre de
Recherche et Developpemont Economlque Unrslt de Montreal
44
bull
I I
I
I
1
Erlich $ V Ginsburgh and L van der Heyden 1987 Wher~ do Real Wage Policies Lead
Belgium bull A General Equilibrium Analysis European Economic Reyiew (forthcoming)
air R and J Taylor 1983 Solution and Maximum Likelihood Estimation of Dynamic
Nonlinear Ratioant Expeclations Models Ecooometrica 51 pp 1169middot1186
Feltenstein A 1984 Money and Bonds In a Disaggregated Ec~nonYin Searl 1-1 J Shoven
Eds Aoolied ~eoeral EqJiliDCiulD tioalYsis Cambridge University Press Cambridge
Feltenstein A 19B6 A Dynamic General Equllibrlu Analysis of Financial Crowding Out
Theory with an Application to Australia Journal of Public Economics 31 No1
Fullerlon D R Gorden 1983 A Reexamination of Tax Distorllons in General Equilibrium
Models in M Feldstein Ed Behayora Simulation Methods in Tax pOlicy Analysis Chicago
University of Chipgo Press
Fvllerlon D J Shaven and J Whalley 1983 Replacing the US Income Tax with a
Progressive Consumption Tax JoulOal of Public ECQnomics 20 pp3middot23
Fullerton D Y Henderson J Shaven 1984 A COmparision of Methodologies in Empirical
General Equilibrium Models of Taxation in Searl H J Shoven Eds ~pQlied Geoaral Jguilibrium
lIoalyss Cambridge University Press Cambridge
Fullerton 0 A King J Shaven and J Whaney 1981 Corporate Tax Integration in the United
Statesmiddot a General Equiiibdufi1 Approach American Economic Review 71 pp~ 677~691
Galper H R Lucke E Tader 1985 Taxation portlor Choice and the Allocation of Capital A
General Equilibrium Approach Mimeo
Gill p W Murray M Saunders M Wright 1986 Users Guide for NPSOL (Version 40) A
Fortran Package for Nonlinear Programming T echnicsl Report SOL 86middot2 Deparlment of
Operations Research Stanford University
Glnsbur~h V and L van der Heyden 1985 General Equilibrium with Wage Rigidities An
45
AppJication to Belgium in A Manne ed t Economic EQumprit1m~ Made Formulation and $oJioo
Mathematical Programming Study 23 Amsterdam NorthmiddotHolland ~
Goulet J 1968 1 Adjustrent CO$1S in the Theory o(the Firm Reyiew of EconomiC Studies 35 bull
pp47middot55
Goulder L 1985 Tax-Financed versus BondmiddotFinancedGhanges in Governent Spending
bull Mimeo Harva(d UniversilYmiddot
Goulder l J Shaven and J Whalley 1983 Domestic Tax Policy and the Foreign Sector The
importance of altemative foreign policy formulations to results from a general equilibrium tax
analysis model In ~n methods in Ibe antico of income from caoia1 ed Martin
Feldstein Chicago University of Chicago riSS
Goulder l L Summers 1987 Tax Policy Asset Prices and 3rowtly ~
A General Equilibrium Analysis NBER Working Pper No 212B
Grandmont J 1982 Temporary General Equilibrium Theory Ch 20 in Handbook 01
Mathemalical Eccnornics Vol II NorthmiddotHoliand
Harberger A 1959 The Corporate Income Tax An empirical Appraisal In lax Revision
Comoendjum Voll House Cornmitee on Ways and Means Washington DC Govemment Printing
Office
Harberger A 1962 The Incidence of the Corporate Income Tax JOULn1 of Political
Economy 70 pp 215middot240
Harberger A 1964 Taxation Resource Aliocalrlon and Welfare fn lb Role of Dir bullbulll and
Indjresl Taxes io tbe Fede[al Reserve s~I~Ill Princeton Princeton Unlveresity Press
James J 1985 New Developments in the Applications of General Equilibrium models to
Economic History in Piggott aand Whalley Eds
Jorgenson D and K Yun 1984 Tax Policy and Capital Allocation Harvard Inslitute of
bull
I
I f
Economic Research WP 1107
Judd K 1965 ShonmiddotRun Analysis of Fiscal Policy in a Simple Perfect Foresight Model
Journal of poJiHcal 5CCJiW1l pp 298middot319 r
Upton D J Potbc J Sachsa nd L Summers 1982 Multiple Shooting in Rational
Expectations Models fuoometrlc 50 1329middot1333
Manne A ed 1985 fconQrnjc Eoullibrium Model FOrIDujaljQo and Solution Mathematical
Pogrammlng Study 23 Amsterdam NorthmiddotHolland
Manne A 1985 Or the Formulation and Solution 01 Economic Equilibrium Models In A
Manne ed ECQOQlli-Jr~ibmiddotum ~tQdel EQlJulailon and Solution Malhematical Programming
Study 23 Amsterdam NorthmiddotHolland
Mansur A and J Whalley 1984 Numerical Specification of Applied General Equilibrium
Models in Scarf H J Shoven Eds Aoplied General Eouillbrium AnaiysectIsect Cambridge Universily
Press Cambridge
Merrill 0 1972 Applications and Extensions of an Algorithm thaI Computes FixedmiddotPoints of
Cenain Upper Semimiddotcominuou Point to Set Mappings PhD Disserlation Unlvel$ity 01 Michigan
Negishi T 1960 Welfar Economics and Exislence I an Equnlbrlum for a Competitive
Economy MelroeCOOQ11ka 12 92middot97
Pereira A 1985 On the Existence and Uniqueness of Optimal Ou1put Path lor CRTS Firms in
Adjustment Costs Technologies Mimeo Stanford University
Pereira A 1985a Consumer Expectations and Ihe COSI 01 UllUty In an Intertemporat
Framework Mlmeo Stanford University
Pereira A 1985b Corporale Tax Integrallon and Government Deficits A Dynamic General
Equilibrium Anaiysls Mlmeo Stanford University
Pereir A 1987 On Ihe Computation of the Users Costs of Stocks Ecooomi (forthcoming)
47
Pereira A bull 1987b Equal Yield Alternatives and Government Oeficlts Mimeof Stanford
University
Pereira A 1987c A Dynamic Applied General ~uHlbrium Model for Tax Policy Evaluation
Comlletamp Docurnent2tion Mimeo University of California San Diego (in progress)
Pereira A 1987d Should the investment Tax Credit be Rei(ltroduced Mlmeo University
of California San Diego (ir progreSs)
Piggott J and J Whalley eds 1985 tJew QeyeQpments in Applied General EQuilibrium
bullenalysjs Cambridge Cambridge University Press
Preckel Pbull 1985 Alternative algorithms for computing economic equilibria In A Manne Ed
Radnor R 19amp3 Equllibrium under Uncertainty In K Arrow M Intrilligaor ods
I ibull
Robinson S 1 g86 Mullisectorai Models of Oeveloping CountrieS a Survey Oivision of
Agricullure and Resources UC Borkeley WP 401
Scarf H 1967 The approximation of fixed points of a continuous mapping ~MM Jouwal of
epplid Mathornallcs 15 pp 1328middot43
Scarf H with T Hansen 1973 00 the Computation of EconomiC Eguilibria New Haven Yale
University Press
Scarf H J Shoven Eds 1984 61l1llied Genera Elujlibrium 6nalyssect Cambridge University
~ Press Cambridge
Shoven J 1976 The Incidence and Efficiency Effects of Taxe on Income from Capital
Jurnal of ramie1 Economy 84 pp 1261-83
Shoven J 1983 Applied General Equilibrium Tax Modemng IMF Slaff PaoerS 30
Shoven J and LB Simon 1986 Share Repurchases and Acquisitions An Analysis of which
Firms Participate NBER Working Paper No 2243
bull
48
IShaven J and J Whalley 1972 ft General Equmbrium Calculation of lhe Effects of
Differential Taxation of Incoma from Capital in the US Jauroe of public Econgmics 1
pp281middot321 (
Shoven J bull and J Whalley 1973 General Equilibrium with Tax$s a computation procedure
and an existence pr00f ReviEhi of Economic Studies 40 pp 475middot490
Shoven J J Whalley 1977 Equal Yield Tax Alternatives A General Equilibrium
Computa~ui0nal Tfchnique J~HJrnal of PUblic EconomIcs 8 pp 211~224 bull
Shaven J J Whalley 1984 Applied GeneralmiddotEquilibrlum MOdels of Taxalion and
International Tr8de An Introduction and Survey Journal of Economic literature 23t pp
1007middot1051
Slemrod J 1980 A General Equilibrium Model of Capital Income Taxation Ph D bullDissertation Harvard Univnrsity
Slemrod J 1983 A General Equilibrium Model of Taxation with Endogenous Financial
Behavior in M Feldstein edbull Behavioral Simulation Methods In Tax Policy Analysis Chicago
UniVerSity of Chicago Press
Stone J 1985 Sequential Optimization and Complementarily Techniques for computing
Economic equilibrium in A Manne Ed
Summers L 1981 Taxation and Corporate Investment a q-Theory Approach BroQkinas
Summers L 1985 Taxation and the Size and ComposHion 01 the Capital Stock An Asset Price
Approach Mlmeo Harvard University
Whalley J 1975 A Geneal Equiloibrium Assessment of the 1973 United Kingdom tax
reform Economica 42 pp 139middot61
1JLTIllOS JIQll~)ltG PAPERS PUBLICADOS bull
I
I I
j
112 70 - LUCENA iJiCIIJo ~ IIEnvironrilent Honitoring and Orqanization Structure T (Setellibro 1987)
nI 71 - L~ENA Diogo A Note on the Representation of Information Stxueture (SeteJObro 1987)
nQ 72 - LUCENA D1CgO Envuorment 111)nitoring and Organization
Structure II (Setelilhro 1987)
ril 73 - AlmmE3 Antonio Pais e CAsPAR Yitor bull Tributacao Inoentivos e InvesUlllentos Analise Qualitativa noYellibro 19S7)
n2 74 - liAlltOSA rose Pedro Opt1lllal lIage R1gidity A SUggestedtlthodology to Teet tb Theory and an Application (Outubro 1987)
nI 75 - CABRAL L 11 B Three lIotes on SYJIIlletric Gemes witt itsYJIIlletric Equilibri (NoveJObro 1987)
n2 76 - LUCENlt D10gOTo Search or Not to Search (Fevereiro1967)
nQ 17 - SA Jorge asconcelos euroI A Theory of SynergyB (Fevereiro1987)
n2 76 - VILARES tanuei Jose Os Be IntenaMios IlIIportados ColIIO Factor de Produao (Julho 1987)
nQ 79 - SA J~rge VasCQncel03 e HoY to Copete and Co~unicate in tature Industrial Products_ (aio 1988)
n2 80 - ROIl Rafael Learning and Capacity Expansion in a New larket Under Uncertainty (Fevereiro 1988)
n2 81 PEREIRA Alfredo llarvilo Survey of Dynamic Computational General EquilibriUlll Hodel for Tax Policy Evaluation (Outubro 1987)
Qualquer in ormacao sabre o 1IIorlltlng Papers ja publlcados sera prestada pelo Secretariado de Apoio aos Docentes podendo os ~e3OSf
ser adquirido5 ne Secyo de Vlndas da Faculdade de EconollLia UNL na Travessa Esteviio Pinto Cltpolide - 1000 LISllOA
31
the contex~ of a Gner1 lemporary equilibrium ftamawor~ In such Circumstances Pereira
(1986a) dev~lops a cnamlc generalization which is obtained as the present discounted value of a i i
sequence of short-fUn optimal expenditure functions consistent with a base case expected future I stream of IlWities I
32
5 Empirical evidenee from selected policy Issues
Dynamic tax models have generated several important results which escaped Sialic modellers
The following is a seleCpoundd set of issuos whIch stress the marginal benefits of dynam~ modelHng of
economic behavior in innovltlttive areas of analysis The discussion of a consumption tax emphasizes
the benefits of dynamic tusehold behavior and intergenerational aspects The study of the
flliminatlon or the re~intToduction of the investsnemt tax credits is made meaningful by the dynamic
modelling of production betlavor Modelling governmenl in a dynamic conlexts allows the modeller
to study the irrrpact of government deflcits Finally a dynamic framework lets us appreCiate the I
relevance of consumer expectations in terms of the evaluation of policy alternatives
i
i
51 Consumption Tax
t
By Its very nature a consumption tax can not be adequately investigated with a static model bull
Fullerion-Shoven-Whaicy (1983) use Ihe model of the US economy s described in
Ballard-Fullerton-Shoven-Whalley (1935) to evaluate the movement from the currenl US tax Isys1em to a progressive consumption tax Since their modeJ Incorporates a Jaborleisure choice
where leisure is an untaxed commodlty their results reflect the fact that both the consumption tax
and Ihe present system are dslortionary
Concentrating on the rntertemporaf distortions they show that sheltering more saving from lhe
current US income tax could improve econornic efficiency even if marginal tax rate increases are
necessary in order to maintain government fevenue~ At first you have a revenue shortfall
However the economy moves to a higher sustatned growth path and uitimatefy lower tax rates
cangenerate the same revenue path Also wages increase in the long run with this policy The
bull
bullbull ~
33 shyt
i
I
I
present value of welfare gi[lS for a potiey of complete savin9 deduction wilh marginal tate
adjustments (consumption lax) is simulated to be around $500 billion to $600 billion 01 1973
dollars
They also investiga1e th 1ngth of ttme it takes th~ economy to adjust to these policy changes
Roughly they estimate the long run to b thirty years although this figure Is very sensitive to the
specification of the savings elasticity - a crucial parameler in this model
52 Investment T~x Credit
Goulder-$ummers (1 S87) address the impact of eliminating the lnvestmenl Tax Cradit (ITC) on
intersectoral capital formation and on economic growth Their model isYa1ticularly adequate to
addies$ this issu~ in that is postulates a forward looking investment m9we with adjustment costs
They show that the eliminating the ITC causes a reduction in the rale of investment Inveslment is
estimated to fall by about 7 in the short-run and by about 12 in Ihe new long run steady state
In lurn a previous policy anncuncement lowers Ihe overall al1ractiveness of Investment and leads 10
a downward shift of the Investrrent profile On the other hand the combined effect of a revenue
neutral simultaneous efimiraron of the ITC and reduction of the corporate tax rates is a long run
reduction ef the capital Sieck by 35 This pattern suggests that a revenue neutral increament in
both Ihe corporale tax rates and the inveSlment lax credits would be preferable 10 the formula
adopted in Ihe US Tax Reform Act of 1986
Pereira (1987d) addresses the impact of the ITe from Ihe oland point of Ihe curTenllax syslem
under the Tax Reform Act of 1986 Given Ihe concern over the potential depressiVe effects upon
savings and investment of the new tax system in conjunction with deficit reduC1ion mechanisms of
the GrammmiddotRudman type a pertinent question is should Ih lTC be e-introduced In the context
bull
34
-of his mod~1 of the US economy Pereira focuses on the trade-off between the distortion in the
intersectoral allocation of capital induced by the lTC the lower tax revenues it generates and
potential financial crowding-out effects on one hand and the positive effects of lowering the relative
price of new capital goods on the other hand Preliminary results confirm the qualitative results in
Goulder-Summers (1987) suggestting a welfare gain of about 2 of the present value of GNP in
the best scenario of absence of replacement taxes Furthermore preliminary results show an
important time pattern to welfare gains with the average benefits increasing the further you look
into the future This pattern is due to the presence of constraints to intersectoral mobility of capital
and inertia in the adjustment towards the optimal capital levels as reflected by the presence of
adjustment costs
I
I
bullAuerbach-Katlikaff (1gB) in the context of their intergenerational model of the US economy
consider the impact on savings capital formation and interest rates of deficit policies and balanced
budget increases in government consumption They conclude that deficit finance and government
consumption can significa-ntly crowd out capital formation and lower the welfare of future
generations However crowding out from deficit finance is a very slow process because it results
from increased government spending over potentially long horizons Also deficit policies may
substantially influence the longmiddotrun interest rates while leaving the short-run rates essentially
unaffected Finally and in opposition to the central role adjustment costs seem to play in both
53 Financial crowding outeff~cts of government deficits
Goulder-Summers and Pereiras models Auerbach-Kotlikoff suggest that the time path of interest
rates induced by a policy of deficit financing seems to be insensitive to the presence of adjustment
costs
bull
35
54 The role 01 consumer expectatlons
The expectations analysis has uncovered one of the benefits of dynamic modelling
Ballard-Goulder (1985) find that the appeal of adopting a consumption tax depends on the level of
foresight possessed by the consumers Furthermore additional foresight may be welfare worsening
This is a second best type of result This tends 10 occur under policies that lead to capital deepening
and declining rate of return to capital over time To the extent that consumers have more foresight
they will be beller equippod to anticipate the fall in the rental price of capital Consequently if a
saving incentive is enacted people save more with myopia than with perfect foresight Given the
exi~tence of taxes on capita and the discrepancy between the private an~--sccial returns to capital
the greater savings with myopia is beller socially Ballard-Goulder find that the welfare gains
from a consumption tax are reduced by about 10 when we move from myopia to farleet foresight
Therefore the attractivenness of adopting a consumption tax seems fairly [obust across these
foresight specifications
I
bull
36
6 Tbe power and weaknesses 01 dynamic modelling the example 01 corporale
tax Integrallon
The Issue of corporate tax integration allows us to sh~w the stre-ngths and weaknesses or the
dynamic approach Empirical evidence using CGE techniques indicates that depending on the precise
scheme of integration and the 1ax replacement methods integratiOn may have substantial effects In
Ihe work of FuliertonmiddotKingmiddotShoven-Whalley (1980 1981) total integration was found to yield an bull
annual static eHielency gain of $4 to $8 billion in 1973 dollars Simulated dynamic gains may be as
large as $695 billion or abcu 14 of the present value of tulUre consumption and leisure in the
US economy These gains result primarily from interindustry reallocalions of Investment and an
improved inter~emporal allocation of consumption
Mora recent work emphasizes hOw consumers asset portfolio decisions and firms finanCial
decisions affect the efficiency gains from integration Slemrod (1980) focusing on consumers
asset ponfoio deCisions finds static efficiency gains which are about twice -as Iarge as those
repOrted iiy Fulierton-King-ShovenWhalley (1961) FuUrlon-Gurdon (1983) focus on the
firms financial decisions They roporl efficiency gains of 6 of GNP from the elimination of the
tax distor1ions favoring debt However when they eliminate the corporate tax and repiace it with
increaSed persona income taxes addlHonal dls10t1ions are created in the optimal labot~leisure
decisions These distortions tend to dominate the analysis slJ(h that the overall effects of complete
integration are very modest Galpr-LuckmiddotTodr (1986) address simultaneously consumers
asset pOrlfollo decisions and the firms financial decisions They also find very modest efficiency
gains from integralion
The above results are important but may be severely biased There are severa aspects of
economic behavior and modelling crucial for the study of Income tax Integration whiCh have no been
i 37
captured innny of the above papers One aspect is the absence of government deficitsmiddot a balanced
budget is assumed It is true Ihal resource crowding OUI is caplured In these mOdels but financial
crowding out induced by government spending is nOI Second investment is not derived from
optimtztlcn behavior Investment behavior passively accomodates endogenous saving decisions As
a consequence the differential impacts of policies in the incentives to save and to invest are not
captured Aso full capital mobility across secors is assumed with instantaneous capital
adjustmenlS towards optimal levels This assumption rules qut different costs of capital across
seclors and therefore differentiated reactions to tax policies changes Finallyhe modelling of both
gov9rnmerlt deficits and of endogenous real and financial investment decisions necessitates the
~nsideraiion of a dynamilt framework and the introduction of financiaJ assets govemment bonds and
iprivate financial assets A dynamic framework also highlights the efficiency effects of Integration on r
ithe ptimal intertemporal decisions The above models are either sIalic (Slemrods and
GalpermiddotltJckemiddotTode(sj or a dynamic sequ~nce of otherwise Slatic model bullbull I Pereir a (1986b) develops a dynamic applied general equilibrium model bull to study the tbull
efficiency effects of integration on the growth and allocation of investmenl across sectors Special
Iattention is paid to the structural affeelS of resource and financial crowding oul induced by
gove mect spending and deficits and to the real and financial Investment reactions across industries
to both tax integration and flnancial crowding out His mode departs from previous work on income
lax Integraticn and for that matler from most 01 the CGIE literalure for tax policy evaluation in
several direcllons It encompasses an endogenous sequential equilibrium struelure founded on
dynamic behavior with flexible expectations Government deficits are optlmaUy determined
Investment decisions are tward looking and the resull of optimizing behavior Several financial
assetsmiddot public and privatemiddot are considered
Simulation resulls Indicale that the welfare gains from Integration under large deficits are at
32
best lJlodest when measured in terms of the GNP The elimination of the corporate tax and its I replacemenl by increased income lax rates yields long run benefits which are never larger than
2 of the presenl yalue ct telr consumption and leisure onder the previous tax regime and 1
under the current tax regime However the short run w~lare effects of full integration tend 10 be
very low and in in some scenarios even negative This is a flew intertemporar pattern of efficiency Ieffects which reflects an adjJsiment lag in the interindustry investment decisions due to the
existence of costs of adjustment On the of her hand partial integration achieved by excluding I dividends from the corporale lax base syslematically yields negalive efficlencyeffeclS In Ihe long Inm these can be as high as 3 of present value of the future value of consumption and leisure This
is a new second best effect suggesllng Ihat less than complete integration may have perverse I efficiency effects
The dynamic slructure of PerclrBs model snowing for an improved freatment of real investment
declsions and government deficts provides a potential setting for a more accurate measure of the
costs and benefits of inlegration The simulation resullS suggest lower benefits and an intertempora Ibull
pattern of growing benefits
Simulation results also clearly suggest robust negative effects from partial integration How
reliable is this result If dJvidends were deductible from the corporate lax base (as are interest
payments) the preferenlial Ireatmant of debl over equily would be ellminaled Corporallons should
be expecled 10 react by decreasing the optimal deWeqully rallo However corporale financial
gecisions are exogenously set In Pereiras model as in aft of the other models surveyed that go as far
as dealing with the issue
Also withdrawing dividends from Ihe corporate tax base is a way of eliminating Ihe double
taxation of dividends at belh Ihe corporate and personnal Income levels How will the optimal
aliocellon of household saving accomodate that change in Ihe relalive return to corporale equity
39 I
t However hOusehold portfolio decisions are exogenouly set in Pereiras model as in all of the other
models surveyeltj go as far as dealing with the issue
IIt is sate to say that the evaluation 01 the benefits 01 partial Integration as defined above are
sevrey underestima1eO Meanwhile the distortionSoenerated by the increase in the marginal
personal tax rates designed to ralse the revenue foregone by the dividend exclusion are fully
aCCltlunted for A good measure of the costs of integration together With a poor evaluation of the
benefils may very well be the reason why partial integration is Simulated 10 yield neg alive
eHikiency gains
On a different vein as most of Ihe model surveyed here Pereiras is a closed economy model It
is legitimate to wonder how the resullS would change If international capital flows were anowed
IThis almost certainly would affect financial crowding out since a good d~al of the capital inflows
could be used to finance public debt
For these various reasons the results should be treated as preliminary but they strongly I suggosl that the dynamic structure provides valuable new insights into the corporate tax integration SSU2
40
7 Concluding remarks I I I i
What has been acccmplished Ihis far The success of Ihe CGE research in laelding the challenge
of dynamic modelling can not be denied
I iGreat progress in the modelling of household behavior has been achieved Promising
developm~nts in the modeUing of production and government behavior have also been accomplished
Interesting innovations in the concept of equilibrium and computation techniques were discussed
The policy analysis was greatly enriched by considering transitional effects together with longer
run steadymiddotstate equilibrium paths generalized equal yield strategies accomodating different
financial crowding out impacts and generalized dynamic policy evaluation indicators
An important set of issues have been addresed by the models surveyed Traditional issues
focusing on the effects of capital taxation and consumption taxes have been pursued In terms of
capital taxation for example the intertemporal nature of the issue was enhanced by allowing an
optimal evolution of capital stock in the economy and forward looking optimal investment decisions
In Andersson (1987) GouldermiddotSummers (1987) and Pereira (1986b 1987d) this is coupled
with a improved treatment of several tax provisions like investment tax credits and depreciation
allowances In terms of the consumption taxes developments in the specification of Intertemporal
household behavior the introduction of overlapping generations and the modelling of
intergenerational links via bequest motives as well as a closer attention to demographic evolution
provided a much improved economic setting for the understanding of the several aspects of the
problem
Furthermore dynamic modelling has permitted the addressing of a set of new issues Policy
Issues ranging from dynamic tax policy analysis to the evaluation of exogenous changes in
government expenditures to the impact of different methods of financing government expenditure to
4 1
the importance of tinanciai croHrii1g out to 1he relevance of consumorS expectations in tne policy
results have been addressed
Despile all of the progress and in part due to such progress several avenues are wide open for
much neded additional research The following seem to be the most Interesting and promising
areasmiddot First the specincation of liquidity constraints to household behavior may help to obtain a
better understanding of policy chenges that affect directly the interest rates in the economy Second
major attention needs to be paid to the incorporafion of a roraign sector (with open capital markets)
Third some efforts are needed in terms of finding adequate computation techniques now that
dynamic modelling eally makes the curse of dimensionality worse than eVOrbull
The modelling of financial markets is the single most unsatisfactory speet of the dynamic
models surveyed here~ The problems associated with provlding a mo~ adequate treatment of
financial markets and in general endogenous and optimal financial deci$ions bull both at the corporate
and at the personal levelsmiddot have 10 be addressed That necessarily requires Introducing uncertainty
into the models To b adequate this must go well beyond the mere assumption of point expeetalions
with the whole array of modelling implementalion and evaluation problems thai generates
bull I i
42
REFERENCES
Abel A and O Blanchard 1983 An Intertemporal Model of Saving and Interest Econometrica
51 pp 675-92
~ Andersson K 1987 Tar-ation of Capital in Sweden - A General Equilibrium Model Ph D
Dissenation UnIversity of Lund Sweder
Auerbach A L KotlikoH 1983 National Savings Economic Welfare and the Structure of
Taxation in M Feldstein ed Behavioral Simulation Methods in Tax Policy- Analysis Chicago
University of Chicago Press
Auerbach A l Kotlikoff 1984 Social Security and the Economics of Demographic-
Transition in H Aaron and G Burtless eds Retirement and Economic Behavior Washington DC
Brookings Institution
Auerbach A L Kotlikoff 1987 Dynamic Fiscal Policy Cambridge Cambridge University
Press
Auerbach A L Kotlikof and J Skinner 1983 The Efficiency Gains from Dynamic Tax
Reform Internatiooaj Economic Reyiew 24 pp 81middot100
Ballard C 1983 Evaluation of the Consumption Tax with Dynamic General Equilibrium
Models Ph D Dissertation Stanford University
Ballard C 1987 Tax Policy and Consumer Foresight A General Equilibrium Simulation
Study Economic Inquiry 25 pp 267-284
Ballard C and L Goulder 1985 Consumption Taxes ForeSight and Welfare a Computatonal
General Equilibrium Analysis in J Piggott and J Whalley eds New Deyelooments in Apolied
General Equilibrium Analysis Cambridge Cambridge University Press
Ballard C D Fullerton J Shoven and J Whalley 1985 A General Equilibrium Model for
t 43
Tax Eoliey Eyalua~iQO NBER University or Chicago Press
Borges A (1987) bull A Survey of Compula1ional General Equilibnum Models CECD Economic
Slud~ 8
Bovenbig L 19B3 imperfect Capital Mobility in a Perfect Foresight Herberger Model
Mimea UC Br~eley
Bovenberg L 1984 Capitel Accumulation and Capital Immobility Qmiddottheory In a Dynamic General Equilibrium Framework PhD dissertation University of California at Berkeley t
Bovenberg L 19850 Dynamic General Equilibrium Tax Models with Adjustment Costs in A iManne ed bull 1985 EtQnomrc ECUjilbrium Model Formulation and Solution Mathematical
lProgramming Siudy 23 Arnsterdam NorthmiddotHoliand
Bovenberg L t 1986 Capital Income Taxation in Growing Open Eeondmies Journa of Pubtic
Eamp(omlcs 3t pp 347middot376 IBovenberg L ald W Keller 1981 Dynamics and Nonlinearlties in Applied General
Equilibrium Models Internal Report Mlmiddot81middot208 Department for Statistical Methods CBS bull
Vooiburg The Netherlands
Charnley C 1981 The Welfare Cost of C~pltal income Taxation In a Growing Economy
IJournal of Political ECQnQm~ 89 3 pp 468middot496 I
Charnley C 19810 Efficient Tax Reform in a Dynamic Model of General Equilibrium I Mimeo Yale University I Charnley C 1982 On the Optimal Taxation In Stylized Models 01 Inlartemporal General
Equilibrium Mimeo Yale University I Decaluw B and A Manens 1985 Pays en Oeveloppement at Modeles Calculables OEquilibre
General - Une Revue de al litterature empirique Monographie Sene 9 Volume 3 Centre de
Recherche et Developpemont Economlque Unrslt de Montreal
44
bull
I I
I
I
1
Erlich $ V Ginsburgh and L van der Heyden 1987 Wher~ do Real Wage Policies Lead
Belgium bull A General Equilibrium Analysis European Economic Reyiew (forthcoming)
air R and J Taylor 1983 Solution and Maximum Likelihood Estimation of Dynamic
Nonlinear Ratioant Expeclations Models Ecooometrica 51 pp 1169middot1186
Feltenstein A 1984 Money and Bonds In a Disaggregated Ec~nonYin Searl 1-1 J Shoven
Eds Aoolied ~eoeral EqJiliDCiulD tioalYsis Cambridge University Press Cambridge
Feltenstein A 19B6 A Dynamic General Equllibrlu Analysis of Financial Crowding Out
Theory with an Application to Australia Journal of Public Economics 31 No1
Fullerlon D R Gorden 1983 A Reexamination of Tax Distorllons in General Equilibrium
Models in M Feldstein Ed Behayora Simulation Methods in Tax pOlicy Analysis Chicago
University of Chipgo Press
Fvllerlon D J Shaven and J Whalley 1983 Replacing the US Income Tax with a
Progressive Consumption Tax JoulOal of Public ECQnomics 20 pp3middot23
Fullerton D Y Henderson J Shaven 1984 A COmparision of Methodologies in Empirical
General Equilibrium Models of Taxation in Searl H J Shoven Eds ~pQlied Geoaral Jguilibrium
lIoalyss Cambridge University Press Cambridge
Fullerton 0 A King J Shaven and J Whaney 1981 Corporate Tax Integration in the United
Statesmiddot a General Equiiibdufi1 Approach American Economic Review 71 pp~ 677~691
Galper H R Lucke E Tader 1985 Taxation portlor Choice and the Allocation of Capital A
General Equilibrium Approach Mimeo
Gill p W Murray M Saunders M Wright 1986 Users Guide for NPSOL (Version 40) A
Fortran Package for Nonlinear Programming T echnicsl Report SOL 86middot2 Deparlment of
Operations Research Stanford University
Glnsbur~h V and L van der Heyden 1985 General Equilibrium with Wage Rigidities An
45
AppJication to Belgium in A Manne ed t Economic EQumprit1m~ Made Formulation and $oJioo
Mathematical Programming Study 23 Amsterdam NorthmiddotHolland ~
Goulet J 1968 1 Adjustrent CO$1S in the Theory o(the Firm Reyiew of EconomiC Studies 35 bull
pp47middot55
Goulder L 1985 Tax-Financed versus BondmiddotFinancedGhanges in Governent Spending
bull Mimeo Harva(d UniversilYmiddot
Goulder l J Shaven and J Whalley 1983 Domestic Tax Policy and the Foreign Sector The
importance of altemative foreign policy formulations to results from a general equilibrium tax
analysis model In ~n methods in Ibe antico of income from caoia1 ed Martin
Feldstein Chicago University of Chicago riSS
Goulder l L Summers 1987 Tax Policy Asset Prices and 3rowtly ~
A General Equilibrium Analysis NBER Working Pper No 212B
Grandmont J 1982 Temporary General Equilibrium Theory Ch 20 in Handbook 01
Mathemalical Eccnornics Vol II NorthmiddotHoliand
Harberger A 1959 The Corporate Income Tax An empirical Appraisal In lax Revision
Comoendjum Voll House Cornmitee on Ways and Means Washington DC Govemment Printing
Office
Harberger A 1962 The Incidence of the Corporate Income Tax JOULn1 of Political
Economy 70 pp 215middot240
Harberger A 1964 Taxation Resource Aliocalrlon and Welfare fn lb Role of Dir bullbulll and
Indjresl Taxes io tbe Fede[al Reserve s~I~Ill Princeton Princeton Unlveresity Press
James J 1985 New Developments in the Applications of General Equilibrium models to
Economic History in Piggott aand Whalley Eds
Jorgenson D and K Yun 1984 Tax Policy and Capital Allocation Harvard Inslitute of
bull
I
I f
Economic Research WP 1107
Judd K 1965 ShonmiddotRun Analysis of Fiscal Policy in a Simple Perfect Foresight Model
Journal of poJiHcal 5CCJiW1l pp 298middot319 r
Upton D J Potbc J Sachsa nd L Summers 1982 Multiple Shooting in Rational
Expectations Models fuoometrlc 50 1329middot1333
Manne A ed 1985 fconQrnjc Eoullibrium Model FOrIDujaljQo and Solution Mathematical
Pogrammlng Study 23 Amsterdam NorthmiddotHolland
Manne A 1985 Or the Formulation and Solution 01 Economic Equilibrium Models In A
Manne ed ECQOQlli-Jr~ibmiddotum ~tQdel EQlJulailon and Solution Malhematical Programming
Study 23 Amsterdam NorthmiddotHolland
Mansur A and J Whalley 1984 Numerical Specification of Applied General Equilibrium
Models in Scarf H J Shoven Eds Aoplied General Eouillbrium AnaiysectIsect Cambridge Universily
Press Cambridge
Merrill 0 1972 Applications and Extensions of an Algorithm thaI Computes FixedmiddotPoints of
Cenain Upper Semimiddotcominuou Point to Set Mappings PhD Disserlation Unlvel$ity 01 Michigan
Negishi T 1960 Welfar Economics and Exislence I an Equnlbrlum for a Competitive
Economy MelroeCOOQ11ka 12 92middot97
Pereira A 1985 On the Existence and Uniqueness of Optimal Ou1put Path lor CRTS Firms in
Adjustment Costs Technologies Mimeo Stanford University
Pereira A 1985a Consumer Expectations and Ihe COSI 01 UllUty In an Intertemporat
Framework Mlmeo Stanford University
Pereira A 1985b Corporale Tax Integrallon and Government Deficits A Dynamic General
Equilibrium Anaiysls Mlmeo Stanford University
Pereir A 1987 On Ihe Computation of the Users Costs of Stocks Ecooomi (forthcoming)
47
Pereira A bull 1987b Equal Yield Alternatives and Government Oeficlts Mimeof Stanford
University
Pereira A 1987c A Dynamic Applied General ~uHlbrium Model for Tax Policy Evaluation
Comlletamp Docurnent2tion Mimeo University of California San Diego (in progress)
Pereira A 1987d Should the investment Tax Credit be Rei(ltroduced Mlmeo University
of California San Diego (ir progreSs)
Piggott J and J Whalley eds 1985 tJew QeyeQpments in Applied General EQuilibrium
bullenalysjs Cambridge Cambridge University Press
Preckel Pbull 1985 Alternative algorithms for computing economic equilibria In A Manne Ed
Radnor R 19amp3 Equllibrium under Uncertainty In K Arrow M Intrilligaor ods
I ibull
Robinson S 1 g86 Mullisectorai Models of Oeveloping CountrieS a Survey Oivision of
Agricullure and Resources UC Borkeley WP 401
Scarf H 1967 The approximation of fixed points of a continuous mapping ~MM Jouwal of
epplid Mathornallcs 15 pp 1328middot43
Scarf H with T Hansen 1973 00 the Computation of EconomiC Eguilibria New Haven Yale
University Press
Scarf H J Shoven Eds 1984 61l1llied Genera Elujlibrium 6nalyssect Cambridge University
~ Press Cambridge
Shoven J 1976 The Incidence and Efficiency Effects of Taxe on Income from Capital
Jurnal of ramie1 Economy 84 pp 1261-83
Shoven J 1983 Applied General Equilibrium Tax Modemng IMF Slaff PaoerS 30
Shoven J and LB Simon 1986 Share Repurchases and Acquisitions An Analysis of which
Firms Participate NBER Working Paper No 2243
bull
48
IShaven J and J Whalley 1972 ft General Equmbrium Calculation of lhe Effects of
Differential Taxation of Incoma from Capital in the US Jauroe of public Econgmics 1
pp281middot321 (
Shoven J bull and J Whalley 1973 General Equilibrium with Tax$s a computation procedure
and an existence pr00f ReviEhi of Economic Studies 40 pp 475middot490
Shoven J J Whalley 1977 Equal Yield Tax Alternatives A General Equilibrium
Computa~ui0nal Tfchnique J~HJrnal of PUblic EconomIcs 8 pp 211~224 bull
Shaven J J Whalley 1984 Applied GeneralmiddotEquilibrlum MOdels of Taxalion and
International Tr8de An Introduction and Survey Journal of Economic literature 23t pp
1007middot1051
Slemrod J 1980 A General Equilibrium Model of Capital Income Taxation Ph D bullDissertation Harvard Univnrsity
Slemrod J 1983 A General Equilibrium Model of Taxation with Endogenous Financial
Behavior in M Feldstein edbull Behavioral Simulation Methods In Tax Policy Analysis Chicago
UniVerSity of Chicago Press
Stone J 1985 Sequential Optimization and Complementarily Techniques for computing
Economic equilibrium in A Manne Ed
Summers L 1981 Taxation and Corporate Investment a q-Theory Approach BroQkinas
Summers L 1985 Taxation and the Size and ComposHion 01 the Capital Stock An Asset Price
Approach Mlmeo Harvard University
Whalley J 1975 A Geneal Equiloibrium Assessment of the 1973 United Kingdom tax
reform Economica 42 pp 139middot61
1JLTIllOS JIQll~)ltG PAPERS PUBLICADOS bull
I
I I
j
112 70 - LUCENA iJiCIIJo ~ IIEnvironrilent Honitoring and Orqanization Structure T (Setellibro 1987)
nI 71 - L~ENA Diogo A Note on the Representation of Information Stxueture (SeteJObro 1987)
nQ 72 - LUCENA D1CgO Envuorment 111)nitoring and Organization
Structure II (Setelilhro 1987)
ril 73 - AlmmE3 Antonio Pais e CAsPAR Yitor bull Tributacao Inoentivos e InvesUlllentos Analise Qualitativa noYellibro 19S7)
n2 74 - liAlltOSA rose Pedro Opt1lllal lIage R1gidity A SUggestedtlthodology to Teet tb Theory and an Application (Outubro 1987)
nI 75 - CABRAL L 11 B Three lIotes on SYJIIlletric Gemes witt itsYJIIlletric Equilibri (NoveJObro 1987)
n2 76 - LUCENlt D10gOTo Search or Not to Search (Fevereiro1967)
nQ 17 - SA Jorge asconcelos euroI A Theory of SynergyB (Fevereiro1987)
n2 76 - VILARES tanuei Jose Os Be IntenaMios IlIIportados ColIIO Factor de Produao (Julho 1987)
nQ 79 - SA J~rge VasCQncel03 e HoY to Copete and Co~unicate in tature Industrial Products_ (aio 1988)
n2 80 - ROIl Rafael Learning and Capacity Expansion in a New larket Under Uncertainty (Fevereiro 1988)
n2 81 PEREIRA Alfredo llarvilo Survey of Dynamic Computational General EquilibriUlll Hodel for Tax Policy Evaluation (Outubro 1987)
Qualquer in ormacao sabre o 1IIorlltlng Papers ja publlcados sera prestada pelo Secretariado de Apoio aos Docentes podendo os ~e3OSf
ser adquirido5 ne Secyo de Vlndas da Faculdade de EconollLia UNL na Travessa Esteviio Pinto Cltpolide - 1000 LISllOA
32
5 Empirical evidenee from selected policy Issues
Dynamic tax models have generated several important results which escaped Sialic modellers
The following is a seleCpoundd set of issuos whIch stress the marginal benefits of dynam~ modelHng of
economic behavior in innovltlttive areas of analysis The discussion of a consumption tax emphasizes
the benefits of dynamic tusehold behavior and intergenerational aspects The study of the
flliminatlon or the re~intToduction of the investsnemt tax credits is made meaningful by the dynamic
modelling of production betlavor Modelling governmenl in a dynamic conlexts allows the modeller
to study the irrrpact of government deflcits Finally a dynamic framework lets us appreCiate the I
relevance of consumer expectations in terms of the evaluation of policy alternatives
i
i
51 Consumption Tax
t
By Its very nature a consumption tax can not be adequately investigated with a static model bull
Fullerion-Shoven-Whaicy (1983) use Ihe model of the US economy s described in
Ballard-Fullerton-Shoven-Whalley (1935) to evaluate the movement from the currenl US tax Isys1em to a progressive consumption tax Since their modeJ Incorporates a Jaborleisure choice
where leisure is an untaxed commodlty their results reflect the fact that both the consumption tax
and Ihe present system are dslortionary
Concentrating on the rntertemporaf distortions they show that sheltering more saving from lhe
current US income tax could improve econornic efficiency even if marginal tax rate increases are
necessary in order to maintain government fevenue~ At first you have a revenue shortfall
However the economy moves to a higher sustatned growth path and uitimatefy lower tax rates
cangenerate the same revenue path Also wages increase in the long run with this policy The
bull
bullbull ~
33 shyt
i
I
I
present value of welfare gi[lS for a potiey of complete savin9 deduction wilh marginal tate
adjustments (consumption lax) is simulated to be around $500 billion to $600 billion 01 1973
dollars
They also investiga1e th 1ngth of ttme it takes th~ economy to adjust to these policy changes
Roughly they estimate the long run to b thirty years although this figure Is very sensitive to the
specification of the savings elasticity - a crucial parameler in this model
52 Investment T~x Credit
Goulder-$ummers (1 S87) address the impact of eliminating the lnvestmenl Tax Cradit (ITC) on
intersectoral capital formation and on economic growth Their model isYa1ticularly adequate to
addies$ this issu~ in that is postulates a forward looking investment m9we with adjustment costs
They show that the eliminating the ITC causes a reduction in the rale of investment Inveslment is
estimated to fall by about 7 in the short-run and by about 12 in Ihe new long run steady state
In lurn a previous policy anncuncement lowers Ihe overall al1ractiveness of Investment and leads 10
a downward shift of the Investrrent profile On the other hand the combined effect of a revenue
neutral simultaneous efimiraron of the ITC and reduction of the corporate tax rates is a long run
reduction ef the capital Sieck by 35 This pattern suggests that a revenue neutral increament in
both Ihe corporale tax rates and the inveSlment lax credits would be preferable 10 the formula
adopted in Ihe US Tax Reform Act of 1986
Pereira (1987d) addresses the impact of the ITe from Ihe oland point of Ihe curTenllax syslem
under the Tax Reform Act of 1986 Given Ihe concern over the potential depressiVe effects upon
savings and investment of the new tax system in conjunction with deficit reduC1ion mechanisms of
the GrammmiddotRudman type a pertinent question is should Ih lTC be e-introduced In the context
bull
34
-of his mod~1 of the US economy Pereira focuses on the trade-off between the distortion in the
intersectoral allocation of capital induced by the lTC the lower tax revenues it generates and
potential financial crowding-out effects on one hand and the positive effects of lowering the relative
price of new capital goods on the other hand Preliminary results confirm the qualitative results in
Goulder-Summers (1987) suggestting a welfare gain of about 2 of the present value of GNP in
the best scenario of absence of replacement taxes Furthermore preliminary results show an
important time pattern to welfare gains with the average benefits increasing the further you look
into the future This pattern is due to the presence of constraints to intersectoral mobility of capital
and inertia in the adjustment towards the optimal capital levels as reflected by the presence of
adjustment costs
I
I
bullAuerbach-Katlikaff (1gB) in the context of their intergenerational model of the US economy
consider the impact on savings capital formation and interest rates of deficit policies and balanced
budget increases in government consumption They conclude that deficit finance and government
consumption can significa-ntly crowd out capital formation and lower the welfare of future
generations However crowding out from deficit finance is a very slow process because it results
from increased government spending over potentially long horizons Also deficit policies may
substantially influence the longmiddotrun interest rates while leaving the short-run rates essentially
unaffected Finally and in opposition to the central role adjustment costs seem to play in both
53 Financial crowding outeff~cts of government deficits
Goulder-Summers and Pereiras models Auerbach-Kotlikoff suggest that the time path of interest
rates induced by a policy of deficit financing seems to be insensitive to the presence of adjustment
costs
bull
35
54 The role 01 consumer expectatlons
The expectations analysis has uncovered one of the benefits of dynamic modelling
Ballard-Goulder (1985) find that the appeal of adopting a consumption tax depends on the level of
foresight possessed by the consumers Furthermore additional foresight may be welfare worsening
This is a second best type of result This tends 10 occur under policies that lead to capital deepening
and declining rate of return to capital over time To the extent that consumers have more foresight
they will be beller equippod to anticipate the fall in the rental price of capital Consequently if a
saving incentive is enacted people save more with myopia than with perfect foresight Given the
exi~tence of taxes on capita and the discrepancy between the private an~--sccial returns to capital
the greater savings with myopia is beller socially Ballard-Goulder find that the welfare gains
from a consumption tax are reduced by about 10 when we move from myopia to farleet foresight
Therefore the attractivenness of adopting a consumption tax seems fairly [obust across these
foresight specifications
I
bull
36
6 Tbe power and weaknesses 01 dynamic modelling the example 01 corporale
tax Integrallon
The Issue of corporate tax integration allows us to sh~w the stre-ngths and weaknesses or the
dynamic approach Empirical evidence using CGE techniques indicates that depending on the precise
scheme of integration and the 1ax replacement methods integratiOn may have substantial effects In
Ihe work of FuliertonmiddotKingmiddotShoven-Whalley (1980 1981) total integration was found to yield an bull
annual static eHielency gain of $4 to $8 billion in 1973 dollars Simulated dynamic gains may be as
large as $695 billion or abcu 14 of the present value of tulUre consumption and leisure in the
US economy These gains result primarily from interindustry reallocalions of Investment and an
improved inter~emporal allocation of consumption
Mora recent work emphasizes hOw consumers asset portfolio decisions and firms finanCial
decisions affect the efficiency gains from integration Slemrod (1980) focusing on consumers
asset ponfoio deCisions finds static efficiency gains which are about twice -as Iarge as those
repOrted iiy Fulierton-King-ShovenWhalley (1961) FuUrlon-Gurdon (1983) focus on the
firms financial decisions They roporl efficiency gains of 6 of GNP from the elimination of the
tax distor1ions favoring debt However when they eliminate the corporate tax and repiace it with
increaSed persona income taxes addlHonal dls10t1ions are created in the optimal labot~leisure
decisions These distortions tend to dominate the analysis slJ(h that the overall effects of complete
integration are very modest Galpr-LuckmiddotTodr (1986) address simultaneously consumers
asset pOrlfollo decisions and the firms financial decisions They also find very modest efficiency
gains from integralion
The above results are important but may be severely biased There are severa aspects of
economic behavior and modelling crucial for the study of Income tax Integration whiCh have no been
i 37
captured innny of the above papers One aspect is the absence of government deficitsmiddot a balanced
budget is assumed It is true Ihal resource crowding OUI is caplured In these mOdels but financial
crowding out induced by government spending is nOI Second investment is not derived from
optimtztlcn behavior Investment behavior passively accomodates endogenous saving decisions As
a consequence the differential impacts of policies in the incentives to save and to invest are not
captured Aso full capital mobility across secors is assumed with instantaneous capital
adjustmenlS towards optimal levels This assumption rules qut different costs of capital across
seclors and therefore differentiated reactions to tax policies changes Finallyhe modelling of both
gov9rnmerlt deficits and of endogenous real and financial investment decisions necessitates the
~nsideraiion of a dynamilt framework and the introduction of financiaJ assets govemment bonds and
iprivate financial assets A dynamic framework also highlights the efficiency effects of Integration on r
ithe ptimal intertemporal decisions The above models are either sIalic (Slemrods and
GalpermiddotltJckemiddotTode(sj or a dynamic sequ~nce of otherwise Slatic model bullbull I Pereir a (1986b) develops a dynamic applied general equilibrium model bull to study the tbull
efficiency effects of integration on the growth and allocation of investmenl across sectors Special
Iattention is paid to the structural affeelS of resource and financial crowding oul induced by
gove mect spending and deficits and to the real and financial Investment reactions across industries
to both tax integration and flnancial crowding out His mode departs from previous work on income
lax Integraticn and for that matler from most 01 the CGIE literalure for tax policy evaluation in
several direcllons It encompasses an endogenous sequential equilibrium struelure founded on
dynamic behavior with flexible expectations Government deficits are optlmaUy determined
Investment decisions are tward looking and the resull of optimizing behavior Several financial
assetsmiddot public and privatemiddot are considered
Simulation resulls Indicale that the welfare gains from Integration under large deficits are at
32
best lJlodest when measured in terms of the GNP The elimination of the corporate tax and its I replacemenl by increased income lax rates yields long run benefits which are never larger than
2 of the presenl yalue ct telr consumption and leisure onder the previous tax regime and 1
under the current tax regime However the short run w~lare effects of full integration tend 10 be
very low and in in some scenarios even negative This is a flew intertemporar pattern of efficiency Ieffects which reflects an adjJsiment lag in the interindustry investment decisions due to the
existence of costs of adjustment On the of her hand partial integration achieved by excluding I dividends from the corporale lax base syslematically yields negalive efficlencyeffeclS In Ihe long Inm these can be as high as 3 of present value of the future value of consumption and leisure This
is a new second best effect suggesllng Ihat less than complete integration may have perverse I efficiency effects
The dynamic slructure of PerclrBs model snowing for an improved freatment of real investment
declsions and government deficts provides a potential setting for a more accurate measure of the
costs and benefits of inlegration The simulation resullS suggest lower benefits and an intertempora Ibull
pattern of growing benefits
Simulation results also clearly suggest robust negative effects from partial integration How
reliable is this result If dJvidends were deductible from the corporate lax base (as are interest
payments) the preferenlial Ireatmant of debl over equily would be ellminaled Corporallons should
be expecled 10 react by decreasing the optimal deWeqully rallo However corporale financial
gecisions are exogenously set In Pereiras model as in aft of the other models surveyed that go as far
as dealing with the issue
Also withdrawing dividends from Ihe corporate tax base is a way of eliminating Ihe double
taxation of dividends at belh Ihe corporate and personnal Income levels How will the optimal
aliocellon of household saving accomodate that change in Ihe relalive return to corporale equity
39 I
t However hOusehold portfolio decisions are exogenouly set in Pereiras model as in all of the other
models surveyeltj go as far as dealing with the issue
IIt is sate to say that the evaluation 01 the benefits 01 partial Integration as defined above are
sevrey underestima1eO Meanwhile the distortionSoenerated by the increase in the marginal
personal tax rates designed to ralse the revenue foregone by the dividend exclusion are fully
aCCltlunted for A good measure of the costs of integration together With a poor evaluation of the
benefils may very well be the reason why partial integration is Simulated 10 yield neg alive
eHikiency gains
On a different vein as most of Ihe model surveyed here Pereiras is a closed economy model It
is legitimate to wonder how the resullS would change If international capital flows were anowed
IThis almost certainly would affect financial crowding out since a good d~al of the capital inflows
could be used to finance public debt
For these various reasons the results should be treated as preliminary but they strongly I suggosl that the dynamic structure provides valuable new insights into the corporate tax integration SSU2
40
7 Concluding remarks I I I i
What has been acccmplished Ihis far The success of Ihe CGE research in laelding the challenge
of dynamic modelling can not be denied
I iGreat progress in the modelling of household behavior has been achieved Promising
developm~nts in the modeUing of production and government behavior have also been accomplished
Interesting innovations in the concept of equilibrium and computation techniques were discussed
The policy analysis was greatly enriched by considering transitional effects together with longer
run steadymiddotstate equilibrium paths generalized equal yield strategies accomodating different
financial crowding out impacts and generalized dynamic policy evaluation indicators
An important set of issues have been addresed by the models surveyed Traditional issues
focusing on the effects of capital taxation and consumption taxes have been pursued In terms of
capital taxation for example the intertemporal nature of the issue was enhanced by allowing an
optimal evolution of capital stock in the economy and forward looking optimal investment decisions
In Andersson (1987) GouldermiddotSummers (1987) and Pereira (1986b 1987d) this is coupled
with a improved treatment of several tax provisions like investment tax credits and depreciation
allowances In terms of the consumption taxes developments in the specification of Intertemporal
household behavior the introduction of overlapping generations and the modelling of
intergenerational links via bequest motives as well as a closer attention to demographic evolution
provided a much improved economic setting for the understanding of the several aspects of the
problem
Furthermore dynamic modelling has permitted the addressing of a set of new issues Policy
Issues ranging from dynamic tax policy analysis to the evaluation of exogenous changes in
government expenditures to the impact of different methods of financing government expenditure to
4 1
the importance of tinanciai croHrii1g out to 1he relevance of consumorS expectations in tne policy
results have been addressed
Despile all of the progress and in part due to such progress several avenues are wide open for
much neded additional research The following seem to be the most Interesting and promising
areasmiddot First the specincation of liquidity constraints to household behavior may help to obtain a
better understanding of policy chenges that affect directly the interest rates in the economy Second
major attention needs to be paid to the incorporafion of a roraign sector (with open capital markets)
Third some efforts are needed in terms of finding adequate computation techniques now that
dynamic modelling eally makes the curse of dimensionality worse than eVOrbull
The modelling of financial markets is the single most unsatisfactory speet of the dynamic
models surveyed here~ The problems associated with provlding a mo~ adequate treatment of
financial markets and in general endogenous and optimal financial deci$ions bull both at the corporate
and at the personal levelsmiddot have 10 be addressed That necessarily requires Introducing uncertainty
into the models To b adequate this must go well beyond the mere assumption of point expeetalions
with the whole array of modelling implementalion and evaluation problems thai generates
bull I i
42
REFERENCES
Abel A and O Blanchard 1983 An Intertemporal Model of Saving and Interest Econometrica
51 pp 675-92
~ Andersson K 1987 Tar-ation of Capital in Sweden - A General Equilibrium Model Ph D
Dissenation UnIversity of Lund Sweder
Auerbach A L KotlikoH 1983 National Savings Economic Welfare and the Structure of
Taxation in M Feldstein ed Behavioral Simulation Methods in Tax Policy- Analysis Chicago
University of Chicago Press
Auerbach A l Kotlikoff 1984 Social Security and the Economics of Demographic-
Transition in H Aaron and G Burtless eds Retirement and Economic Behavior Washington DC
Brookings Institution
Auerbach A L Kotlikoff 1987 Dynamic Fiscal Policy Cambridge Cambridge University
Press
Auerbach A L Kotlikof and J Skinner 1983 The Efficiency Gains from Dynamic Tax
Reform Internatiooaj Economic Reyiew 24 pp 81middot100
Ballard C 1983 Evaluation of the Consumption Tax with Dynamic General Equilibrium
Models Ph D Dissertation Stanford University
Ballard C 1987 Tax Policy and Consumer Foresight A General Equilibrium Simulation
Study Economic Inquiry 25 pp 267-284
Ballard C and L Goulder 1985 Consumption Taxes ForeSight and Welfare a Computatonal
General Equilibrium Analysis in J Piggott and J Whalley eds New Deyelooments in Apolied
General Equilibrium Analysis Cambridge Cambridge University Press
Ballard C D Fullerton J Shoven and J Whalley 1985 A General Equilibrium Model for
t 43
Tax Eoliey Eyalua~iQO NBER University or Chicago Press
Borges A (1987) bull A Survey of Compula1ional General Equilibnum Models CECD Economic
Slud~ 8
Bovenbig L 19B3 imperfect Capital Mobility in a Perfect Foresight Herberger Model
Mimea UC Br~eley
Bovenberg L 1984 Capitel Accumulation and Capital Immobility Qmiddottheory In a Dynamic General Equilibrium Framework PhD dissertation University of California at Berkeley t
Bovenberg L 19850 Dynamic General Equilibrium Tax Models with Adjustment Costs in A iManne ed bull 1985 EtQnomrc ECUjilbrium Model Formulation and Solution Mathematical
lProgramming Siudy 23 Arnsterdam NorthmiddotHoliand
Bovenberg L t 1986 Capital Income Taxation in Growing Open Eeondmies Journa of Pubtic
Eamp(omlcs 3t pp 347middot376 IBovenberg L ald W Keller 1981 Dynamics and Nonlinearlties in Applied General
Equilibrium Models Internal Report Mlmiddot81middot208 Department for Statistical Methods CBS bull
Vooiburg The Netherlands
Charnley C 1981 The Welfare Cost of C~pltal income Taxation In a Growing Economy
IJournal of Political ECQnQm~ 89 3 pp 468middot496 I
Charnley C 19810 Efficient Tax Reform in a Dynamic Model of General Equilibrium I Mimeo Yale University I Charnley C 1982 On the Optimal Taxation In Stylized Models 01 Inlartemporal General
Equilibrium Mimeo Yale University I Decaluw B and A Manens 1985 Pays en Oeveloppement at Modeles Calculables OEquilibre
General - Une Revue de al litterature empirique Monographie Sene 9 Volume 3 Centre de
Recherche et Developpemont Economlque Unrslt de Montreal
44
bull
I I
I
I
1
Erlich $ V Ginsburgh and L van der Heyden 1987 Wher~ do Real Wage Policies Lead
Belgium bull A General Equilibrium Analysis European Economic Reyiew (forthcoming)
air R and J Taylor 1983 Solution and Maximum Likelihood Estimation of Dynamic
Nonlinear Ratioant Expeclations Models Ecooometrica 51 pp 1169middot1186
Feltenstein A 1984 Money and Bonds In a Disaggregated Ec~nonYin Searl 1-1 J Shoven
Eds Aoolied ~eoeral EqJiliDCiulD tioalYsis Cambridge University Press Cambridge
Feltenstein A 19B6 A Dynamic General Equllibrlu Analysis of Financial Crowding Out
Theory with an Application to Australia Journal of Public Economics 31 No1
Fullerlon D R Gorden 1983 A Reexamination of Tax Distorllons in General Equilibrium
Models in M Feldstein Ed Behayora Simulation Methods in Tax pOlicy Analysis Chicago
University of Chipgo Press
Fvllerlon D J Shaven and J Whalley 1983 Replacing the US Income Tax with a
Progressive Consumption Tax JoulOal of Public ECQnomics 20 pp3middot23
Fullerton D Y Henderson J Shaven 1984 A COmparision of Methodologies in Empirical
General Equilibrium Models of Taxation in Searl H J Shoven Eds ~pQlied Geoaral Jguilibrium
lIoalyss Cambridge University Press Cambridge
Fullerton 0 A King J Shaven and J Whaney 1981 Corporate Tax Integration in the United
Statesmiddot a General Equiiibdufi1 Approach American Economic Review 71 pp~ 677~691
Galper H R Lucke E Tader 1985 Taxation portlor Choice and the Allocation of Capital A
General Equilibrium Approach Mimeo
Gill p W Murray M Saunders M Wright 1986 Users Guide for NPSOL (Version 40) A
Fortran Package for Nonlinear Programming T echnicsl Report SOL 86middot2 Deparlment of
Operations Research Stanford University
Glnsbur~h V and L van der Heyden 1985 General Equilibrium with Wage Rigidities An
45
AppJication to Belgium in A Manne ed t Economic EQumprit1m~ Made Formulation and $oJioo
Mathematical Programming Study 23 Amsterdam NorthmiddotHolland ~
Goulet J 1968 1 Adjustrent CO$1S in the Theory o(the Firm Reyiew of EconomiC Studies 35 bull
pp47middot55
Goulder L 1985 Tax-Financed versus BondmiddotFinancedGhanges in Governent Spending
bull Mimeo Harva(d UniversilYmiddot
Goulder l J Shaven and J Whalley 1983 Domestic Tax Policy and the Foreign Sector The
importance of altemative foreign policy formulations to results from a general equilibrium tax
analysis model In ~n methods in Ibe antico of income from caoia1 ed Martin
Feldstein Chicago University of Chicago riSS
Goulder l L Summers 1987 Tax Policy Asset Prices and 3rowtly ~
A General Equilibrium Analysis NBER Working Pper No 212B
Grandmont J 1982 Temporary General Equilibrium Theory Ch 20 in Handbook 01
Mathemalical Eccnornics Vol II NorthmiddotHoliand
Harberger A 1959 The Corporate Income Tax An empirical Appraisal In lax Revision
Comoendjum Voll House Cornmitee on Ways and Means Washington DC Govemment Printing
Office
Harberger A 1962 The Incidence of the Corporate Income Tax JOULn1 of Political
Economy 70 pp 215middot240
Harberger A 1964 Taxation Resource Aliocalrlon and Welfare fn lb Role of Dir bullbulll and
Indjresl Taxes io tbe Fede[al Reserve s~I~Ill Princeton Princeton Unlveresity Press
James J 1985 New Developments in the Applications of General Equilibrium models to
Economic History in Piggott aand Whalley Eds
Jorgenson D and K Yun 1984 Tax Policy and Capital Allocation Harvard Inslitute of
bull
I
I f
Economic Research WP 1107
Judd K 1965 ShonmiddotRun Analysis of Fiscal Policy in a Simple Perfect Foresight Model
Journal of poJiHcal 5CCJiW1l pp 298middot319 r
Upton D J Potbc J Sachsa nd L Summers 1982 Multiple Shooting in Rational
Expectations Models fuoometrlc 50 1329middot1333
Manne A ed 1985 fconQrnjc Eoullibrium Model FOrIDujaljQo and Solution Mathematical
Pogrammlng Study 23 Amsterdam NorthmiddotHolland
Manne A 1985 Or the Formulation and Solution 01 Economic Equilibrium Models In A
Manne ed ECQOQlli-Jr~ibmiddotum ~tQdel EQlJulailon and Solution Malhematical Programming
Study 23 Amsterdam NorthmiddotHolland
Mansur A and J Whalley 1984 Numerical Specification of Applied General Equilibrium
Models in Scarf H J Shoven Eds Aoplied General Eouillbrium AnaiysectIsect Cambridge Universily
Press Cambridge
Merrill 0 1972 Applications and Extensions of an Algorithm thaI Computes FixedmiddotPoints of
Cenain Upper Semimiddotcominuou Point to Set Mappings PhD Disserlation Unlvel$ity 01 Michigan
Negishi T 1960 Welfar Economics and Exislence I an Equnlbrlum for a Competitive
Economy MelroeCOOQ11ka 12 92middot97
Pereira A 1985 On the Existence and Uniqueness of Optimal Ou1put Path lor CRTS Firms in
Adjustment Costs Technologies Mimeo Stanford University
Pereira A 1985a Consumer Expectations and Ihe COSI 01 UllUty In an Intertemporat
Framework Mlmeo Stanford University
Pereira A 1985b Corporale Tax Integrallon and Government Deficits A Dynamic General
Equilibrium Anaiysls Mlmeo Stanford University
Pereir A 1987 On Ihe Computation of the Users Costs of Stocks Ecooomi (forthcoming)
47
Pereira A bull 1987b Equal Yield Alternatives and Government Oeficlts Mimeof Stanford
University
Pereira A 1987c A Dynamic Applied General ~uHlbrium Model for Tax Policy Evaluation
Comlletamp Docurnent2tion Mimeo University of California San Diego (in progress)
Pereira A 1987d Should the investment Tax Credit be Rei(ltroduced Mlmeo University
of California San Diego (ir progreSs)
Piggott J and J Whalley eds 1985 tJew QeyeQpments in Applied General EQuilibrium
bullenalysjs Cambridge Cambridge University Press
Preckel Pbull 1985 Alternative algorithms for computing economic equilibria In A Manne Ed
Radnor R 19amp3 Equllibrium under Uncertainty In K Arrow M Intrilligaor ods
I ibull
Robinson S 1 g86 Mullisectorai Models of Oeveloping CountrieS a Survey Oivision of
Agricullure and Resources UC Borkeley WP 401
Scarf H 1967 The approximation of fixed points of a continuous mapping ~MM Jouwal of
epplid Mathornallcs 15 pp 1328middot43
Scarf H with T Hansen 1973 00 the Computation of EconomiC Eguilibria New Haven Yale
University Press
Scarf H J Shoven Eds 1984 61l1llied Genera Elujlibrium 6nalyssect Cambridge University
~ Press Cambridge
Shoven J 1976 The Incidence and Efficiency Effects of Taxe on Income from Capital
Jurnal of ramie1 Economy 84 pp 1261-83
Shoven J 1983 Applied General Equilibrium Tax Modemng IMF Slaff PaoerS 30
Shoven J and LB Simon 1986 Share Repurchases and Acquisitions An Analysis of which
Firms Participate NBER Working Paper No 2243
bull
48
IShaven J and J Whalley 1972 ft General Equmbrium Calculation of lhe Effects of
Differential Taxation of Incoma from Capital in the US Jauroe of public Econgmics 1
pp281middot321 (
Shoven J bull and J Whalley 1973 General Equilibrium with Tax$s a computation procedure
and an existence pr00f ReviEhi of Economic Studies 40 pp 475middot490
Shoven J J Whalley 1977 Equal Yield Tax Alternatives A General Equilibrium
Computa~ui0nal Tfchnique J~HJrnal of PUblic EconomIcs 8 pp 211~224 bull
Shaven J J Whalley 1984 Applied GeneralmiddotEquilibrlum MOdels of Taxalion and
International Tr8de An Introduction and Survey Journal of Economic literature 23t pp
1007middot1051
Slemrod J 1980 A General Equilibrium Model of Capital Income Taxation Ph D bullDissertation Harvard Univnrsity
Slemrod J 1983 A General Equilibrium Model of Taxation with Endogenous Financial
Behavior in M Feldstein edbull Behavioral Simulation Methods In Tax Policy Analysis Chicago
UniVerSity of Chicago Press
Stone J 1985 Sequential Optimization and Complementarily Techniques for computing
Economic equilibrium in A Manne Ed
Summers L 1981 Taxation and Corporate Investment a q-Theory Approach BroQkinas
Summers L 1985 Taxation and the Size and ComposHion 01 the Capital Stock An Asset Price
Approach Mlmeo Harvard University
Whalley J 1975 A Geneal Equiloibrium Assessment of the 1973 United Kingdom tax
reform Economica 42 pp 139middot61
1JLTIllOS JIQll~)ltG PAPERS PUBLICADOS bull
I
I I
j
112 70 - LUCENA iJiCIIJo ~ IIEnvironrilent Honitoring and Orqanization Structure T (Setellibro 1987)
nI 71 - L~ENA Diogo A Note on the Representation of Information Stxueture (SeteJObro 1987)
nQ 72 - LUCENA D1CgO Envuorment 111)nitoring and Organization
Structure II (Setelilhro 1987)
ril 73 - AlmmE3 Antonio Pais e CAsPAR Yitor bull Tributacao Inoentivos e InvesUlllentos Analise Qualitativa noYellibro 19S7)
n2 74 - liAlltOSA rose Pedro Opt1lllal lIage R1gidity A SUggestedtlthodology to Teet tb Theory and an Application (Outubro 1987)
nI 75 - CABRAL L 11 B Three lIotes on SYJIIlletric Gemes witt itsYJIIlletric Equilibri (NoveJObro 1987)
n2 76 - LUCENlt D10gOTo Search or Not to Search (Fevereiro1967)
nQ 17 - SA Jorge asconcelos euroI A Theory of SynergyB (Fevereiro1987)
n2 76 - VILARES tanuei Jose Os Be IntenaMios IlIIportados ColIIO Factor de Produao (Julho 1987)
nQ 79 - SA J~rge VasCQncel03 e HoY to Copete and Co~unicate in tature Industrial Products_ (aio 1988)
n2 80 - ROIl Rafael Learning and Capacity Expansion in a New larket Under Uncertainty (Fevereiro 1988)
n2 81 PEREIRA Alfredo llarvilo Survey of Dynamic Computational General EquilibriUlll Hodel for Tax Policy Evaluation (Outubro 1987)
Qualquer in ormacao sabre o 1IIorlltlng Papers ja publlcados sera prestada pelo Secretariado de Apoio aos Docentes podendo os ~e3OSf
ser adquirido5 ne Secyo de Vlndas da Faculdade de EconollLia UNL na Travessa Esteviio Pinto Cltpolide - 1000 LISllOA
bullbull ~
33 shyt
i
I
I
present value of welfare gi[lS for a potiey of complete savin9 deduction wilh marginal tate
adjustments (consumption lax) is simulated to be around $500 billion to $600 billion 01 1973
dollars
They also investiga1e th 1ngth of ttme it takes th~ economy to adjust to these policy changes
Roughly they estimate the long run to b thirty years although this figure Is very sensitive to the
specification of the savings elasticity - a crucial parameler in this model
52 Investment T~x Credit
Goulder-$ummers (1 S87) address the impact of eliminating the lnvestmenl Tax Cradit (ITC) on
intersectoral capital formation and on economic growth Their model isYa1ticularly adequate to
addies$ this issu~ in that is postulates a forward looking investment m9we with adjustment costs
They show that the eliminating the ITC causes a reduction in the rale of investment Inveslment is
estimated to fall by about 7 in the short-run and by about 12 in Ihe new long run steady state
In lurn a previous policy anncuncement lowers Ihe overall al1ractiveness of Investment and leads 10
a downward shift of the Investrrent profile On the other hand the combined effect of a revenue
neutral simultaneous efimiraron of the ITC and reduction of the corporate tax rates is a long run
reduction ef the capital Sieck by 35 This pattern suggests that a revenue neutral increament in
both Ihe corporale tax rates and the inveSlment lax credits would be preferable 10 the formula
adopted in Ihe US Tax Reform Act of 1986
Pereira (1987d) addresses the impact of the ITe from Ihe oland point of Ihe curTenllax syslem
under the Tax Reform Act of 1986 Given Ihe concern over the potential depressiVe effects upon
savings and investment of the new tax system in conjunction with deficit reduC1ion mechanisms of
the GrammmiddotRudman type a pertinent question is should Ih lTC be e-introduced In the context
bull
34
-of his mod~1 of the US economy Pereira focuses on the trade-off between the distortion in the
intersectoral allocation of capital induced by the lTC the lower tax revenues it generates and
potential financial crowding-out effects on one hand and the positive effects of lowering the relative
price of new capital goods on the other hand Preliminary results confirm the qualitative results in
Goulder-Summers (1987) suggestting a welfare gain of about 2 of the present value of GNP in
the best scenario of absence of replacement taxes Furthermore preliminary results show an
important time pattern to welfare gains with the average benefits increasing the further you look
into the future This pattern is due to the presence of constraints to intersectoral mobility of capital
and inertia in the adjustment towards the optimal capital levels as reflected by the presence of
adjustment costs
I
I
bullAuerbach-Katlikaff (1gB) in the context of their intergenerational model of the US economy
consider the impact on savings capital formation and interest rates of deficit policies and balanced
budget increases in government consumption They conclude that deficit finance and government
consumption can significa-ntly crowd out capital formation and lower the welfare of future
generations However crowding out from deficit finance is a very slow process because it results
from increased government spending over potentially long horizons Also deficit policies may
substantially influence the longmiddotrun interest rates while leaving the short-run rates essentially
unaffected Finally and in opposition to the central role adjustment costs seem to play in both
53 Financial crowding outeff~cts of government deficits
Goulder-Summers and Pereiras models Auerbach-Kotlikoff suggest that the time path of interest
rates induced by a policy of deficit financing seems to be insensitive to the presence of adjustment
costs
bull
35
54 The role 01 consumer expectatlons
The expectations analysis has uncovered one of the benefits of dynamic modelling
Ballard-Goulder (1985) find that the appeal of adopting a consumption tax depends on the level of
foresight possessed by the consumers Furthermore additional foresight may be welfare worsening
This is a second best type of result This tends 10 occur under policies that lead to capital deepening
and declining rate of return to capital over time To the extent that consumers have more foresight
they will be beller equippod to anticipate the fall in the rental price of capital Consequently if a
saving incentive is enacted people save more with myopia than with perfect foresight Given the
exi~tence of taxes on capita and the discrepancy between the private an~--sccial returns to capital
the greater savings with myopia is beller socially Ballard-Goulder find that the welfare gains
from a consumption tax are reduced by about 10 when we move from myopia to farleet foresight
Therefore the attractivenness of adopting a consumption tax seems fairly [obust across these
foresight specifications
I
bull
36
6 Tbe power and weaknesses 01 dynamic modelling the example 01 corporale
tax Integrallon
The Issue of corporate tax integration allows us to sh~w the stre-ngths and weaknesses or the
dynamic approach Empirical evidence using CGE techniques indicates that depending on the precise
scheme of integration and the 1ax replacement methods integratiOn may have substantial effects In
Ihe work of FuliertonmiddotKingmiddotShoven-Whalley (1980 1981) total integration was found to yield an bull
annual static eHielency gain of $4 to $8 billion in 1973 dollars Simulated dynamic gains may be as
large as $695 billion or abcu 14 of the present value of tulUre consumption and leisure in the
US economy These gains result primarily from interindustry reallocalions of Investment and an
improved inter~emporal allocation of consumption
Mora recent work emphasizes hOw consumers asset portfolio decisions and firms finanCial
decisions affect the efficiency gains from integration Slemrod (1980) focusing on consumers
asset ponfoio deCisions finds static efficiency gains which are about twice -as Iarge as those
repOrted iiy Fulierton-King-ShovenWhalley (1961) FuUrlon-Gurdon (1983) focus on the
firms financial decisions They roporl efficiency gains of 6 of GNP from the elimination of the
tax distor1ions favoring debt However when they eliminate the corporate tax and repiace it with
increaSed persona income taxes addlHonal dls10t1ions are created in the optimal labot~leisure
decisions These distortions tend to dominate the analysis slJ(h that the overall effects of complete
integration are very modest Galpr-LuckmiddotTodr (1986) address simultaneously consumers
asset pOrlfollo decisions and the firms financial decisions They also find very modest efficiency
gains from integralion
The above results are important but may be severely biased There are severa aspects of
economic behavior and modelling crucial for the study of Income tax Integration whiCh have no been
i 37
captured innny of the above papers One aspect is the absence of government deficitsmiddot a balanced
budget is assumed It is true Ihal resource crowding OUI is caplured In these mOdels but financial
crowding out induced by government spending is nOI Second investment is not derived from
optimtztlcn behavior Investment behavior passively accomodates endogenous saving decisions As
a consequence the differential impacts of policies in the incentives to save and to invest are not
captured Aso full capital mobility across secors is assumed with instantaneous capital
adjustmenlS towards optimal levels This assumption rules qut different costs of capital across
seclors and therefore differentiated reactions to tax policies changes Finallyhe modelling of both
gov9rnmerlt deficits and of endogenous real and financial investment decisions necessitates the
~nsideraiion of a dynamilt framework and the introduction of financiaJ assets govemment bonds and
iprivate financial assets A dynamic framework also highlights the efficiency effects of Integration on r
ithe ptimal intertemporal decisions The above models are either sIalic (Slemrods and
GalpermiddotltJckemiddotTode(sj or a dynamic sequ~nce of otherwise Slatic model bullbull I Pereir a (1986b) develops a dynamic applied general equilibrium model bull to study the tbull
efficiency effects of integration on the growth and allocation of investmenl across sectors Special
Iattention is paid to the structural affeelS of resource and financial crowding oul induced by
gove mect spending and deficits and to the real and financial Investment reactions across industries
to both tax integration and flnancial crowding out His mode departs from previous work on income
lax Integraticn and for that matler from most 01 the CGIE literalure for tax policy evaluation in
several direcllons It encompasses an endogenous sequential equilibrium struelure founded on
dynamic behavior with flexible expectations Government deficits are optlmaUy determined
Investment decisions are tward looking and the resull of optimizing behavior Several financial
assetsmiddot public and privatemiddot are considered
Simulation resulls Indicale that the welfare gains from Integration under large deficits are at
32
best lJlodest when measured in terms of the GNP The elimination of the corporate tax and its I replacemenl by increased income lax rates yields long run benefits which are never larger than
2 of the presenl yalue ct telr consumption and leisure onder the previous tax regime and 1
under the current tax regime However the short run w~lare effects of full integration tend 10 be
very low and in in some scenarios even negative This is a flew intertemporar pattern of efficiency Ieffects which reflects an adjJsiment lag in the interindustry investment decisions due to the
existence of costs of adjustment On the of her hand partial integration achieved by excluding I dividends from the corporale lax base syslematically yields negalive efficlencyeffeclS In Ihe long Inm these can be as high as 3 of present value of the future value of consumption and leisure This
is a new second best effect suggesllng Ihat less than complete integration may have perverse I efficiency effects
The dynamic slructure of PerclrBs model snowing for an improved freatment of real investment
declsions and government deficts provides a potential setting for a more accurate measure of the
costs and benefits of inlegration The simulation resullS suggest lower benefits and an intertempora Ibull
pattern of growing benefits
Simulation results also clearly suggest robust negative effects from partial integration How
reliable is this result If dJvidends were deductible from the corporate lax base (as are interest
payments) the preferenlial Ireatmant of debl over equily would be ellminaled Corporallons should
be expecled 10 react by decreasing the optimal deWeqully rallo However corporale financial
gecisions are exogenously set In Pereiras model as in aft of the other models surveyed that go as far
as dealing with the issue
Also withdrawing dividends from Ihe corporate tax base is a way of eliminating Ihe double
taxation of dividends at belh Ihe corporate and personnal Income levels How will the optimal
aliocellon of household saving accomodate that change in Ihe relalive return to corporale equity
39 I
t However hOusehold portfolio decisions are exogenouly set in Pereiras model as in all of the other
models surveyeltj go as far as dealing with the issue
IIt is sate to say that the evaluation 01 the benefits 01 partial Integration as defined above are
sevrey underestima1eO Meanwhile the distortionSoenerated by the increase in the marginal
personal tax rates designed to ralse the revenue foregone by the dividend exclusion are fully
aCCltlunted for A good measure of the costs of integration together With a poor evaluation of the
benefils may very well be the reason why partial integration is Simulated 10 yield neg alive
eHikiency gains
On a different vein as most of Ihe model surveyed here Pereiras is a closed economy model It
is legitimate to wonder how the resullS would change If international capital flows were anowed
IThis almost certainly would affect financial crowding out since a good d~al of the capital inflows
could be used to finance public debt
For these various reasons the results should be treated as preliminary but they strongly I suggosl that the dynamic structure provides valuable new insights into the corporate tax integration SSU2
40
7 Concluding remarks I I I i
What has been acccmplished Ihis far The success of Ihe CGE research in laelding the challenge
of dynamic modelling can not be denied
I iGreat progress in the modelling of household behavior has been achieved Promising
developm~nts in the modeUing of production and government behavior have also been accomplished
Interesting innovations in the concept of equilibrium and computation techniques were discussed
The policy analysis was greatly enriched by considering transitional effects together with longer
run steadymiddotstate equilibrium paths generalized equal yield strategies accomodating different
financial crowding out impacts and generalized dynamic policy evaluation indicators
An important set of issues have been addresed by the models surveyed Traditional issues
focusing on the effects of capital taxation and consumption taxes have been pursued In terms of
capital taxation for example the intertemporal nature of the issue was enhanced by allowing an
optimal evolution of capital stock in the economy and forward looking optimal investment decisions
In Andersson (1987) GouldermiddotSummers (1987) and Pereira (1986b 1987d) this is coupled
with a improved treatment of several tax provisions like investment tax credits and depreciation
allowances In terms of the consumption taxes developments in the specification of Intertemporal
household behavior the introduction of overlapping generations and the modelling of
intergenerational links via bequest motives as well as a closer attention to demographic evolution
provided a much improved economic setting for the understanding of the several aspects of the
problem
Furthermore dynamic modelling has permitted the addressing of a set of new issues Policy
Issues ranging from dynamic tax policy analysis to the evaluation of exogenous changes in
government expenditures to the impact of different methods of financing government expenditure to
4 1
the importance of tinanciai croHrii1g out to 1he relevance of consumorS expectations in tne policy
results have been addressed
Despile all of the progress and in part due to such progress several avenues are wide open for
much neded additional research The following seem to be the most Interesting and promising
areasmiddot First the specincation of liquidity constraints to household behavior may help to obtain a
better understanding of policy chenges that affect directly the interest rates in the economy Second
major attention needs to be paid to the incorporafion of a roraign sector (with open capital markets)
Third some efforts are needed in terms of finding adequate computation techniques now that
dynamic modelling eally makes the curse of dimensionality worse than eVOrbull
The modelling of financial markets is the single most unsatisfactory speet of the dynamic
models surveyed here~ The problems associated with provlding a mo~ adequate treatment of
financial markets and in general endogenous and optimal financial deci$ions bull both at the corporate
and at the personal levelsmiddot have 10 be addressed That necessarily requires Introducing uncertainty
into the models To b adequate this must go well beyond the mere assumption of point expeetalions
with the whole array of modelling implementalion and evaluation problems thai generates
bull I i
42
REFERENCES
Abel A and O Blanchard 1983 An Intertemporal Model of Saving and Interest Econometrica
51 pp 675-92
~ Andersson K 1987 Tar-ation of Capital in Sweden - A General Equilibrium Model Ph D
Dissenation UnIversity of Lund Sweder
Auerbach A L KotlikoH 1983 National Savings Economic Welfare and the Structure of
Taxation in M Feldstein ed Behavioral Simulation Methods in Tax Policy- Analysis Chicago
University of Chicago Press
Auerbach A l Kotlikoff 1984 Social Security and the Economics of Demographic-
Transition in H Aaron and G Burtless eds Retirement and Economic Behavior Washington DC
Brookings Institution
Auerbach A L Kotlikoff 1987 Dynamic Fiscal Policy Cambridge Cambridge University
Press
Auerbach A L Kotlikof and J Skinner 1983 The Efficiency Gains from Dynamic Tax
Reform Internatiooaj Economic Reyiew 24 pp 81middot100
Ballard C 1983 Evaluation of the Consumption Tax with Dynamic General Equilibrium
Models Ph D Dissertation Stanford University
Ballard C 1987 Tax Policy and Consumer Foresight A General Equilibrium Simulation
Study Economic Inquiry 25 pp 267-284
Ballard C and L Goulder 1985 Consumption Taxes ForeSight and Welfare a Computatonal
General Equilibrium Analysis in J Piggott and J Whalley eds New Deyelooments in Apolied
General Equilibrium Analysis Cambridge Cambridge University Press
Ballard C D Fullerton J Shoven and J Whalley 1985 A General Equilibrium Model for
t 43
Tax Eoliey Eyalua~iQO NBER University or Chicago Press
Borges A (1987) bull A Survey of Compula1ional General Equilibnum Models CECD Economic
Slud~ 8
Bovenbig L 19B3 imperfect Capital Mobility in a Perfect Foresight Herberger Model
Mimea UC Br~eley
Bovenberg L 1984 Capitel Accumulation and Capital Immobility Qmiddottheory In a Dynamic General Equilibrium Framework PhD dissertation University of California at Berkeley t
Bovenberg L 19850 Dynamic General Equilibrium Tax Models with Adjustment Costs in A iManne ed bull 1985 EtQnomrc ECUjilbrium Model Formulation and Solution Mathematical
lProgramming Siudy 23 Arnsterdam NorthmiddotHoliand
Bovenberg L t 1986 Capital Income Taxation in Growing Open Eeondmies Journa of Pubtic
Eamp(omlcs 3t pp 347middot376 IBovenberg L ald W Keller 1981 Dynamics and Nonlinearlties in Applied General
Equilibrium Models Internal Report Mlmiddot81middot208 Department for Statistical Methods CBS bull
Vooiburg The Netherlands
Charnley C 1981 The Welfare Cost of C~pltal income Taxation In a Growing Economy
IJournal of Political ECQnQm~ 89 3 pp 468middot496 I
Charnley C 19810 Efficient Tax Reform in a Dynamic Model of General Equilibrium I Mimeo Yale University I Charnley C 1982 On the Optimal Taxation In Stylized Models 01 Inlartemporal General
Equilibrium Mimeo Yale University I Decaluw B and A Manens 1985 Pays en Oeveloppement at Modeles Calculables OEquilibre
General - Une Revue de al litterature empirique Monographie Sene 9 Volume 3 Centre de
Recherche et Developpemont Economlque Unrslt de Montreal
44
bull
I I
I
I
1
Erlich $ V Ginsburgh and L van der Heyden 1987 Wher~ do Real Wage Policies Lead
Belgium bull A General Equilibrium Analysis European Economic Reyiew (forthcoming)
air R and J Taylor 1983 Solution and Maximum Likelihood Estimation of Dynamic
Nonlinear Ratioant Expeclations Models Ecooometrica 51 pp 1169middot1186
Feltenstein A 1984 Money and Bonds In a Disaggregated Ec~nonYin Searl 1-1 J Shoven
Eds Aoolied ~eoeral EqJiliDCiulD tioalYsis Cambridge University Press Cambridge
Feltenstein A 19B6 A Dynamic General Equllibrlu Analysis of Financial Crowding Out
Theory with an Application to Australia Journal of Public Economics 31 No1
Fullerlon D R Gorden 1983 A Reexamination of Tax Distorllons in General Equilibrium
Models in M Feldstein Ed Behayora Simulation Methods in Tax pOlicy Analysis Chicago
University of Chipgo Press
Fvllerlon D J Shaven and J Whalley 1983 Replacing the US Income Tax with a
Progressive Consumption Tax JoulOal of Public ECQnomics 20 pp3middot23
Fullerton D Y Henderson J Shaven 1984 A COmparision of Methodologies in Empirical
General Equilibrium Models of Taxation in Searl H J Shoven Eds ~pQlied Geoaral Jguilibrium
lIoalyss Cambridge University Press Cambridge
Fullerton 0 A King J Shaven and J Whaney 1981 Corporate Tax Integration in the United
Statesmiddot a General Equiiibdufi1 Approach American Economic Review 71 pp~ 677~691
Galper H R Lucke E Tader 1985 Taxation portlor Choice and the Allocation of Capital A
General Equilibrium Approach Mimeo
Gill p W Murray M Saunders M Wright 1986 Users Guide for NPSOL (Version 40) A
Fortran Package for Nonlinear Programming T echnicsl Report SOL 86middot2 Deparlment of
Operations Research Stanford University
Glnsbur~h V and L van der Heyden 1985 General Equilibrium with Wage Rigidities An
45
AppJication to Belgium in A Manne ed t Economic EQumprit1m~ Made Formulation and $oJioo
Mathematical Programming Study 23 Amsterdam NorthmiddotHolland ~
Goulet J 1968 1 Adjustrent CO$1S in the Theory o(the Firm Reyiew of EconomiC Studies 35 bull
pp47middot55
Goulder L 1985 Tax-Financed versus BondmiddotFinancedGhanges in Governent Spending
bull Mimeo Harva(d UniversilYmiddot
Goulder l J Shaven and J Whalley 1983 Domestic Tax Policy and the Foreign Sector The
importance of altemative foreign policy formulations to results from a general equilibrium tax
analysis model In ~n methods in Ibe antico of income from caoia1 ed Martin
Feldstein Chicago University of Chicago riSS
Goulder l L Summers 1987 Tax Policy Asset Prices and 3rowtly ~
A General Equilibrium Analysis NBER Working Pper No 212B
Grandmont J 1982 Temporary General Equilibrium Theory Ch 20 in Handbook 01
Mathemalical Eccnornics Vol II NorthmiddotHoliand
Harberger A 1959 The Corporate Income Tax An empirical Appraisal In lax Revision
Comoendjum Voll House Cornmitee on Ways and Means Washington DC Govemment Printing
Office
Harberger A 1962 The Incidence of the Corporate Income Tax JOULn1 of Political
Economy 70 pp 215middot240
Harberger A 1964 Taxation Resource Aliocalrlon and Welfare fn lb Role of Dir bullbulll and
Indjresl Taxes io tbe Fede[al Reserve s~I~Ill Princeton Princeton Unlveresity Press
James J 1985 New Developments in the Applications of General Equilibrium models to
Economic History in Piggott aand Whalley Eds
Jorgenson D and K Yun 1984 Tax Policy and Capital Allocation Harvard Inslitute of
bull
I
I f
Economic Research WP 1107
Judd K 1965 ShonmiddotRun Analysis of Fiscal Policy in a Simple Perfect Foresight Model
Journal of poJiHcal 5CCJiW1l pp 298middot319 r
Upton D J Potbc J Sachsa nd L Summers 1982 Multiple Shooting in Rational
Expectations Models fuoometrlc 50 1329middot1333
Manne A ed 1985 fconQrnjc Eoullibrium Model FOrIDujaljQo and Solution Mathematical
Pogrammlng Study 23 Amsterdam NorthmiddotHolland
Manne A 1985 Or the Formulation and Solution 01 Economic Equilibrium Models In A
Manne ed ECQOQlli-Jr~ibmiddotum ~tQdel EQlJulailon and Solution Malhematical Programming
Study 23 Amsterdam NorthmiddotHolland
Mansur A and J Whalley 1984 Numerical Specification of Applied General Equilibrium
Models in Scarf H J Shoven Eds Aoplied General Eouillbrium AnaiysectIsect Cambridge Universily
Press Cambridge
Merrill 0 1972 Applications and Extensions of an Algorithm thaI Computes FixedmiddotPoints of
Cenain Upper Semimiddotcominuou Point to Set Mappings PhD Disserlation Unlvel$ity 01 Michigan
Negishi T 1960 Welfar Economics and Exislence I an Equnlbrlum for a Competitive
Economy MelroeCOOQ11ka 12 92middot97
Pereira A 1985 On the Existence and Uniqueness of Optimal Ou1put Path lor CRTS Firms in
Adjustment Costs Technologies Mimeo Stanford University
Pereira A 1985a Consumer Expectations and Ihe COSI 01 UllUty In an Intertemporat
Framework Mlmeo Stanford University
Pereira A 1985b Corporale Tax Integrallon and Government Deficits A Dynamic General
Equilibrium Anaiysls Mlmeo Stanford University
Pereir A 1987 On Ihe Computation of the Users Costs of Stocks Ecooomi (forthcoming)
47
Pereira A bull 1987b Equal Yield Alternatives and Government Oeficlts Mimeof Stanford
University
Pereira A 1987c A Dynamic Applied General ~uHlbrium Model for Tax Policy Evaluation
Comlletamp Docurnent2tion Mimeo University of California San Diego (in progress)
Pereira A 1987d Should the investment Tax Credit be Rei(ltroduced Mlmeo University
of California San Diego (ir progreSs)
Piggott J and J Whalley eds 1985 tJew QeyeQpments in Applied General EQuilibrium
bullenalysjs Cambridge Cambridge University Press
Preckel Pbull 1985 Alternative algorithms for computing economic equilibria In A Manne Ed
Radnor R 19amp3 Equllibrium under Uncertainty In K Arrow M Intrilligaor ods
I ibull
Robinson S 1 g86 Mullisectorai Models of Oeveloping CountrieS a Survey Oivision of
Agricullure and Resources UC Borkeley WP 401
Scarf H 1967 The approximation of fixed points of a continuous mapping ~MM Jouwal of
epplid Mathornallcs 15 pp 1328middot43
Scarf H with T Hansen 1973 00 the Computation of EconomiC Eguilibria New Haven Yale
University Press
Scarf H J Shoven Eds 1984 61l1llied Genera Elujlibrium 6nalyssect Cambridge University
~ Press Cambridge
Shoven J 1976 The Incidence and Efficiency Effects of Taxe on Income from Capital
Jurnal of ramie1 Economy 84 pp 1261-83
Shoven J 1983 Applied General Equilibrium Tax Modemng IMF Slaff PaoerS 30
Shoven J and LB Simon 1986 Share Repurchases and Acquisitions An Analysis of which
Firms Participate NBER Working Paper No 2243
bull
48
IShaven J and J Whalley 1972 ft General Equmbrium Calculation of lhe Effects of
Differential Taxation of Incoma from Capital in the US Jauroe of public Econgmics 1
pp281middot321 (
Shoven J bull and J Whalley 1973 General Equilibrium with Tax$s a computation procedure
and an existence pr00f ReviEhi of Economic Studies 40 pp 475middot490
Shoven J J Whalley 1977 Equal Yield Tax Alternatives A General Equilibrium
Computa~ui0nal Tfchnique J~HJrnal of PUblic EconomIcs 8 pp 211~224 bull
Shaven J J Whalley 1984 Applied GeneralmiddotEquilibrlum MOdels of Taxalion and
International Tr8de An Introduction and Survey Journal of Economic literature 23t pp
1007middot1051
Slemrod J 1980 A General Equilibrium Model of Capital Income Taxation Ph D bullDissertation Harvard Univnrsity
Slemrod J 1983 A General Equilibrium Model of Taxation with Endogenous Financial
Behavior in M Feldstein edbull Behavioral Simulation Methods In Tax Policy Analysis Chicago
UniVerSity of Chicago Press
Stone J 1985 Sequential Optimization and Complementarily Techniques for computing
Economic equilibrium in A Manne Ed
Summers L 1981 Taxation and Corporate Investment a q-Theory Approach BroQkinas
Summers L 1985 Taxation and the Size and ComposHion 01 the Capital Stock An Asset Price
Approach Mlmeo Harvard University
Whalley J 1975 A Geneal Equiloibrium Assessment of the 1973 United Kingdom tax
reform Economica 42 pp 139middot61
1JLTIllOS JIQll~)ltG PAPERS PUBLICADOS bull
I
I I
j
112 70 - LUCENA iJiCIIJo ~ IIEnvironrilent Honitoring and Orqanization Structure T (Setellibro 1987)
nI 71 - L~ENA Diogo A Note on the Representation of Information Stxueture (SeteJObro 1987)
nQ 72 - LUCENA D1CgO Envuorment 111)nitoring and Organization
Structure II (Setelilhro 1987)
ril 73 - AlmmE3 Antonio Pais e CAsPAR Yitor bull Tributacao Inoentivos e InvesUlllentos Analise Qualitativa noYellibro 19S7)
n2 74 - liAlltOSA rose Pedro Opt1lllal lIage R1gidity A SUggestedtlthodology to Teet tb Theory and an Application (Outubro 1987)
nI 75 - CABRAL L 11 B Three lIotes on SYJIIlletric Gemes witt itsYJIIlletric Equilibri (NoveJObro 1987)
n2 76 - LUCENlt D10gOTo Search or Not to Search (Fevereiro1967)
nQ 17 - SA Jorge asconcelos euroI A Theory of SynergyB (Fevereiro1987)
n2 76 - VILARES tanuei Jose Os Be IntenaMios IlIIportados ColIIO Factor de Produao (Julho 1987)
nQ 79 - SA J~rge VasCQncel03 e HoY to Copete and Co~unicate in tature Industrial Products_ (aio 1988)
n2 80 - ROIl Rafael Learning and Capacity Expansion in a New larket Under Uncertainty (Fevereiro 1988)
n2 81 PEREIRA Alfredo llarvilo Survey of Dynamic Computational General EquilibriUlll Hodel for Tax Policy Evaluation (Outubro 1987)
Qualquer in ormacao sabre o 1IIorlltlng Papers ja publlcados sera prestada pelo Secretariado de Apoio aos Docentes podendo os ~e3OSf
ser adquirido5 ne Secyo de Vlndas da Faculdade de EconollLia UNL na Travessa Esteviio Pinto Cltpolide - 1000 LISllOA
bull
34
-of his mod~1 of the US economy Pereira focuses on the trade-off between the distortion in the
intersectoral allocation of capital induced by the lTC the lower tax revenues it generates and
potential financial crowding-out effects on one hand and the positive effects of lowering the relative
price of new capital goods on the other hand Preliminary results confirm the qualitative results in
Goulder-Summers (1987) suggestting a welfare gain of about 2 of the present value of GNP in
the best scenario of absence of replacement taxes Furthermore preliminary results show an
important time pattern to welfare gains with the average benefits increasing the further you look
into the future This pattern is due to the presence of constraints to intersectoral mobility of capital
and inertia in the adjustment towards the optimal capital levels as reflected by the presence of
adjustment costs
I
I
bullAuerbach-Katlikaff (1gB) in the context of their intergenerational model of the US economy
consider the impact on savings capital formation and interest rates of deficit policies and balanced
budget increases in government consumption They conclude that deficit finance and government
consumption can significa-ntly crowd out capital formation and lower the welfare of future
generations However crowding out from deficit finance is a very slow process because it results
from increased government spending over potentially long horizons Also deficit policies may
substantially influence the longmiddotrun interest rates while leaving the short-run rates essentially
unaffected Finally and in opposition to the central role adjustment costs seem to play in both
53 Financial crowding outeff~cts of government deficits
Goulder-Summers and Pereiras models Auerbach-Kotlikoff suggest that the time path of interest
rates induced by a policy of deficit financing seems to be insensitive to the presence of adjustment
costs
bull
35
54 The role 01 consumer expectatlons
The expectations analysis has uncovered one of the benefits of dynamic modelling
Ballard-Goulder (1985) find that the appeal of adopting a consumption tax depends on the level of
foresight possessed by the consumers Furthermore additional foresight may be welfare worsening
This is a second best type of result This tends 10 occur under policies that lead to capital deepening
and declining rate of return to capital over time To the extent that consumers have more foresight
they will be beller equippod to anticipate the fall in the rental price of capital Consequently if a
saving incentive is enacted people save more with myopia than with perfect foresight Given the
exi~tence of taxes on capita and the discrepancy between the private an~--sccial returns to capital
the greater savings with myopia is beller socially Ballard-Goulder find that the welfare gains
from a consumption tax are reduced by about 10 when we move from myopia to farleet foresight
Therefore the attractivenness of adopting a consumption tax seems fairly [obust across these
foresight specifications
I
bull
36
6 Tbe power and weaknesses 01 dynamic modelling the example 01 corporale
tax Integrallon
The Issue of corporate tax integration allows us to sh~w the stre-ngths and weaknesses or the
dynamic approach Empirical evidence using CGE techniques indicates that depending on the precise
scheme of integration and the 1ax replacement methods integratiOn may have substantial effects In
Ihe work of FuliertonmiddotKingmiddotShoven-Whalley (1980 1981) total integration was found to yield an bull
annual static eHielency gain of $4 to $8 billion in 1973 dollars Simulated dynamic gains may be as
large as $695 billion or abcu 14 of the present value of tulUre consumption and leisure in the
US economy These gains result primarily from interindustry reallocalions of Investment and an
improved inter~emporal allocation of consumption
Mora recent work emphasizes hOw consumers asset portfolio decisions and firms finanCial
decisions affect the efficiency gains from integration Slemrod (1980) focusing on consumers
asset ponfoio deCisions finds static efficiency gains which are about twice -as Iarge as those
repOrted iiy Fulierton-King-ShovenWhalley (1961) FuUrlon-Gurdon (1983) focus on the
firms financial decisions They roporl efficiency gains of 6 of GNP from the elimination of the
tax distor1ions favoring debt However when they eliminate the corporate tax and repiace it with
increaSed persona income taxes addlHonal dls10t1ions are created in the optimal labot~leisure
decisions These distortions tend to dominate the analysis slJ(h that the overall effects of complete
integration are very modest Galpr-LuckmiddotTodr (1986) address simultaneously consumers
asset pOrlfollo decisions and the firms financial decisions They also find very modest efficiency
gains from integralion
The above results are important but may be severely biased There are severa aspects of
economic behavior and modelling crucial for the study of Income tax Integration whiCh have no been
i 37
captured innny of the above papers One aspect is the absence of government deficitsmiddot a balanced
budget is assumed It is true Ihal resource crowding OUI is caplured In these mOdels but financial
crowding out induced by government spending is nOI Second investment is not derived from
optimtztlcn behavior Investment behavior passively accomodates endogenous saving decisions As
a consequence the differential impacts of policies in the incentives to save and to invest are not
captured Aso full capital mobility across secors is assumed with instantaneous capital
adjustmenlS towards optimal levels This assumption rules qut different costs of capital across
seclors and therefore differentiated reactions to tax policies changes Finallyhe modelling of both
gov9rnmerlt deficits and of endogenous real and financial investment decisions necessitates the
~nsideraiion of a dynamilt framework and the introduction of financiaJ assets govemment bonds and
iprivate financial assets A dynamic framework also highlights the efficiency effects of Integration on r
ithe ptimal intertemporal decisions The above models are either sIalic (Slemrods and
GalpermiddotltJckemiddotTode(sj or a dynamic sequ~nce of otherwise Slatic model bullbull I Pereir a (1986b) develops a dynamic applied general equilibrium model bull to study the tbull
efficiency effects of integration on the growth and allocation of investmenl across sectors Special
Iattention is paid to the structural affeelS of resource and financial crowding oul induced by
gove mect spending and deficits and to the real and financial Investment reactions across industries
to both tax integration and flnancial crowding out His mode departs from previous work on income
lax Integraticn and for that matler from most 01 the CGIE literalure for tax policy evaluation in
several direcllons It encompasses an endogenous sequential equilibrium struelure founded on
dynamic behavior with flexible expectations Government deficits are optlmaUy determined
Investment decisions are tward looking and the resull of optimizing behavior Several financial
assetsmiddot public and privatemiddot are considered
Simulation resulls Indicale that the welfare gains from Integration under large deficits are at
32
best lJlodest when measured in terms of the GNP The elimination of the corporate tax and its I replacemenl by increased income lax rates yields long run benefits which are never larger than
2 of the presenl yalue ct telr consumption and leisure onder the previous tax regime and 1
under the current tax regime However the short run w~lare effects of full integration tend 10 be
very low and in in some scenarios even negative This is a flew intertemporar pattern of efficiency Ieffects which reflects an adjJsiment lag in the interindustry investment decisions due to the
existence of costs of adjustment On the of her hand partial integration achieved by excluding I dividends from the corporale lax base syslematically yields negalive efficlencyeffeclS In Ihe long Inm these can be as high as 3 of present value of the future value of consumption and leisure This
is a new second best effect suggesllng Ihat less than complete integration may have perverse I efficiency effects
The dynamic slructure of PerclrBs model snowing for an improved freatment of real investment
declsions and government deficts provides a potential setting for a more accurate measure of the
costs and benefits of inlegration The simulation resullS suggest lower benefits and an intertempora Ibull
pattern of growing benefits
Simulation results also clearly suggest robust negative effects from partial integration How
reliable is this result If dJvidends were deductible from the corporate lax base (as are interest
payments) the preferenlial Ireatmant of debl over equily would be ellminaled Corporallons should
be expecled 10 react by decreasing the optimal deWeqully rallo However corporale financial
gecisions are exogenously set In Pereiras model as in aft of the other models surveyed that go as far
as dealing with the issue
Also withdrawing dividends from Ihe corporate tax base is a way of eliminating Ihe double
taxation of dividends at belh Ihe corporate and personnal Income levels How will the optimal
aliocellon of household saving accomodate that change in Ihe relalive return to corporale equity
39 I
t However hOusehold portfolio decisions are exogenouly set in Pereiras model as in all of the other
models surveyeltj go as far as dealing with the issue
IIt is sate to say that the evaluation 01 the benefits 01 partial Integration as defined above are
sevrey underestima1eO Meanwhile the distortionSoenerated by the increase in the marginal
personal tax rates designed to ralse the revenue foregone by the dividend exclusion are fully
aCCltlunted for A good measure of the costs of integration together With a poor evaluation of the
benefils may very well be the reason why partial integration is Simulated 10 yield neg alive
eHikiency gains
On a different vein as most of Ihe model surveyed here Pereiras is a closed economy model It
is legitimate to wonder how the resullS would change If international capital flows were anowed
IThis almost certainly would affect financial crowding out since a good d~al of the capital inflows
could be used to finance public debt
For these various reasons the results should be treated as preliminary but they strongly I suggosl that the dynamic structure provides valuable new insights into the corporate tax integration SSU2
40
7 Concluding remarks I I I i
What has been acccmplished Ihis far The success of Ihe CGE research in laelding the challenge
of dynamic modelling can not be denied
I iGreat progress in the modelling of household behavior has been achieved Promising
developm~nts in the modeUing of production and government behavior have also been accomplished
Interesting innovations in the concept of equilibrium and computation techniques were discussed
The policy analysis was greatly enriched by considering transitional effects together with longer
run steadymiddotstate equilibrium paths generalized equal yield strategies accomodating different
financial crowding out impacts and generalized dynamic policy evaluation indicators
An important set of issues have been addresed by the models surveyed Traditional issues
focusing on the effects of capital taxation and consumption taxes have been pursued In terms of
capital taxation for example the intertemporal nature of the issue was enhanced by allowing an
optimal evolution of capital stock in the economy and forward looking optimal investment decisions
In Andersson (1987) GouldermiddotSummers (1987) and Pereira (1986b 1987d) this is coupled
with a improved treatment of several tax provisions like investment tax credits and depreciation
allowances In terms of the consumption taxes developments in the specification of Intertemporal
household behavior the introduction of overlapping generations and the modelling of
intergenerational links via bequest motives as well as a closer attention to demographic evolution
provided a much improved economic setting for the understanding of the several aspects of the
problem
Furthermore dynamic modelling has permitted the addressing of a set of new issues Policy
Issues ranging from dynamic tax policy analysis to the evaluation of exogenous changes in
government expenditures to the impact of different methods of financing government expenditure to
4 1
the importance of tinanciai croHrii1g out to 1he relevance of consumorS expectations in tne policy
results have been addressed
Despile all of the progress and in part due to such progress several avenues are wide open for
much neded additional research The following seem to be the most Interesting and promising
areasmiddot First the specincation of liquidity constraints to household behavior may help to obtain a
better understanding of policy chenges that affect directly the interest rates in the economy Second
major attention needs to be paid to the incorporafion of a roraign sector (with open capital markets)
Third some efforts are needed in terms of finding adequate computation techniques now that
dynamic modelling eally makes the curse of dimensionality worse than eVOrbull
The modelling of financial markets is the single most unsatisfactory speet of the dynamic
models surveyed here~ The problems associated with provlding a mo~ adequate treatment of
financial markets and in general endogenous and optimal financial deci$ions bull both at the corporate
and at the personal levelsmiddot have 10 be addressed That necessarily requires Introducing uncertainty
into the models To b adequate this must go well beyond the mere assumption of point expeetalions
with the whole array of modelling implementalion and evaluation problems thai generates
bull I i
42
REFERENCES
Abel A and O Blanchard 1983 An Intertemporal Model of Saving and Interest Econometrica
51 pp 675-92
~ Andersson K 1987 Tar-ation of Capital in Sweden - A General Equilibrium Model Ph D
Dissenation UnIversity of Lund Sweder
Auerbach A L KotlikoH 1983 National Savings Economic Welfare and the Structure of
Taxation in M Feldstein ed Behavioral Simulation Methods in Tax Policy- Analysis Chicago
University of Chicago Press
Auerbach A l Kotlikoff 1984 Social Security and the Economics of Demographic-
Transition in H Aaron and G Burtless eds Retirement and Economic Behavior Washington DC
Brookings Institution
Auerbach A L Kotlikoff 1987 Dynamic Fiscal Policy Cambridge Cambridge University
Press
Auerbach A L Kotlikof and J Skinner 1983 The Efficiency Gains from Dynamic Tax
Reform Internatiooaj Economic Reyiew 24 pp 81middot100
Ballard C 1983 Evaluation of the Consumption Tax with Dynamic General Equilibrium
Models Ph D Dissertation Stanford University
Ballard C 1987 Tax Policy and Consumer Foresight A General Equilibrium Simulation
Study Economic Inquiry 25 pp 267-284
Ballard C and L Goulder 1985 Consumption Taxes ForeSight and Welfare a Computatonal
General Equilibrium Analysis in J Piggott and J Whalley eds New Deyelooments in Apolied
General Equilibrium Analysis Cambridge Cambridge University Press
Ballard C D Fullerton J Shoven and J Whalley 1985 A General Equilibrium Model for
t 43
Tax Eoliey Eyalua~iQO NBER University or Chicago Press
Borges A (1987) bull A Survey of Compula1ional General Equilibnum Models CECD Economic
Slud~ 8
Bovenbig L 19B3 imperfect Capital Mobility in a Perfect Foresight Herberger Model
Mimea UC Br~eley
Bovenberg L 1984 Capitel Accumulation and Capital Immobility Qmiddottheory In a Dynamic General Equilibrium Framework PhD dissertation University of California at Berkeley t
Bovenberg L 19850 Dynamic General Equilibrium Tax Models with Adjustment Costs in A iManne ed bull 1985 EtQnomrc ECUjilbrium Model Formulation and Solution Mathematical
lProgramming Siudy 23 Arnsterdam NorthmiddotHoliand
Bovenberg L t 1986 Capital Income Taxation in Growing Open Eeondmies Journa of Pubtic
Eamp(omlcs 3t pp 347middot376 IBovenberg L ald W Keller 1981 Dynamics and Nonlinearlties in Applied General
Equilibrium Models Internal Report Mlmiddot81middot208 Department for Statistical Methods CBS bull
Vooiburg The Netherlands
Charnley C 1981 The Welfare Cost of C~pltal income Taxation In a Growing Economy
IJournal of Political ECQnQm~ 89 3 pp 468middot496 I
Charnley C 19810 Efficient Tax Reform in a Dynamic Model of General Equilibrium I Mimeo Yale University I Charnley C 1982 On the Optimal Taxation In Stylized Models 01 Inlartemporal General
Equilibrium Mimeo Yale University I Decaluw B and A Manens 1985 Pays en Oeveloppement at Modeles Calculables OEquilibre
General - Une Revue de al litterature empirique Monographie Sene 9 Volume 3 Centre de
Recherche et Developpemont Economlque Unrslt de Montreal
44
bull
I I
I
I
1
Erlich $ V Ginsburgh and L van der Heyden 1987 Wher~ do Real Wage Policies Lead
Belgium bull A General Equilibrium Analysis European Economic Reyiew (forthcoming)
air R and J Taylor 1983 Solution and Maximum Likelihood Estimation of Dynamic
Nonlinear Ratioant Expeclations Models Ecooometrica 51 pp 1169middot1186
Feltenstein A 1984 Money and Bonds In a Disaggregated Ec~nonYin Searl 1-1 J Shoven
Eds Aoolied ~eoeral EqJiliDCiulD tioalYsis Cambridge University Press Cambridge
Feltenstein A 19B6 A Dynamic General Equllibrlu Analysis of Financial Crowding Out
Theory with an Application to Australia Journal of Public Economics 31 No1
Fullerlon D R Gorden 1983 A Reexamination of Tax Distorllons in General Equilibrium
Models in M Feldstein Ed Behayora Simulation Methods in Tax pOlicy Analysis Chicago
University of Chipgo Press
Fvllerlon D J Shaven and J Whalley 1983 Replacing the US Income Tax with a
Progressive Consumption Tax JoulOal of Public ECQnomics 20 pp3middot23
Fullerton D Y Henderson J Shaven 1984 A COmparision of Methodologies in Empirical
General Equilibrium Models of Taxation in Searl H J Shoven Eds ~pQlied Geoaral Jguilibrium
lIoalyss Cambridge University Press Cambridge
Fullerton 0 A King J Shaven and J Whaney 1981 Corporate Tax Integration in the United
Statesmiddot a General Equiiibdufi1 Approach American Economic Review 71 pp~ 677~691
Galper H R Lucke E Tader 1985 Taxation portlor Choice and the Allocation of Capital A
General Equilibrium Approach Mimeo
Gill p W Murray M Saunders M Wright 1986 Users Guide for NPSOL (Version 40) A
Fortran Package for Nonlinear Programming T echnicsl Report SOL 86middot2 Deparlment of
Operations Research Stanford University
Glnsbur~h V and L van der Heyden 1985 General Equilibrium with Wage Rigidities An
45
AppJication to Belgium in A Manne ed t Economic EQumprit1m~ Made Formulation and $oJioo
Mathematical Programming Study 23 Amsterdam NorthmiddotHolland ~
Goulet J 1968 1 Adjustrent CO$1S in the Theory o(the Firm Reyiew of EconomiC Studies 35 bull
pp47middot55
Goulder L 1985 Tax-Financed versus BondmiddotFinancedGhanges in Governent Spending
bull Mimeo Harva(d UniversilYmiddot
Goulder l J Shaven and J Whalley 1983 Domestic Tax Policy and the Foreign Sector The
importance of altemative foreign policy formulations to results from a general equilibrium tax
analysis model In ~n methods in Ibe antico of income from caoia1 ed Martin
Feldstein Chicago University of Chicago riSS
Goulder l L Summers 1987 Tax Policy Asset Prices and 3rowtly ~
A General Equilibrium Analysis NBER Working Pper No 212B
Grandmont J 1982 Temporary General Equilibrium Theory Ch 20 in Handbook 01
Mathemalical Eccnornics Vol II NorthmiddotHoliand
Harberger A 1959 The Corporate Income Tax An empirical Appraisal In lax Revision
Comoendjum Voll House Cornmitee on Ways and Means Washington DC Govemment Printing
Office
Harberger A 1962 The Incidence of the Corporate Income Tax JOULn1 of Political
Economy 70 pp 215middot240
Harberger A 1964 Taxation Resource Aliocalrlon and Welfare fn lb Role of Dir bullbulll and
Indjresl Taxes io tbe Fede[al Reserve s~I~Ill Princeton Princeton Unlveresity Press
James J 1985 New Developments in the Applications of General Equilibrium models to
Economic History in Piggott aand Whalley Eds
Jorgenson D and K Yun 1984 Tax Policy and Capital Allocation Harvard Inslitute of
bull
I
I f
Economic Research WP 1107
Judd K 1965 ShonmiddotRun Analysis of Fiscal Policy in a Simple Perfect Foresight Model
Journal of poJiHcal 5CCJiW1l pp 298middot319 r
Upton D J Potbc J Sachsa nd L Summers 1982 Multiple Shooting in Rational
Expectations Models fuoometrlc 50 1329middot1333
Manne A ed 1985 fconQrnjc Eoullibrium Model FOrIDujaljQo and Solution Mathematical
Pogrammlng Study 23 Amsterdam NorthmiddotHolland
Manne A 1985 Or the Formulation and Solution 01 Economic Equilibrium Models In A
Manne ed ECQOQlli-Jr~ibmiddotum ~tQdel EQlJulailon and Solution Malhematical Programming
Study 23 Amsterdam NorthmiddotHolland
Mansur A and J Whalley 1984 Numerical Specification of Applied General Equilibrium
Models in Scarf H J Shoven Eds Aoplied General Eouillbrium AnaiysectIsect Cambridge Universily
Press Cambridge
Merrill 0 1972 Applications and Extensions of an Algorithm thaI Computes FixedmiddotPoints of
Cenain Upper Semimiddotcominuou Point to Set Mappings PhD Disserlation Unlvel$ity 01 Michigan
Negishi T 1960 Welfar Economics and Exislence I an Equnlbrlum for a Competitive
Economy MelroeCOOQ11ka 12 92middot97
Pereira A 1985 On the Existence and Uniqueness of Optimal Ou1put Path lor CRTS Firms in
Adjustment Costs Technologies Mimeo Stanford University
Pereira A 1985a Consumer Expectations and Ihe COSI 01 UllUty In an Intertemporat
Framework Mlmeo Stanford University
Pereira A 1985b Corporale Tax Integrallon and Government Deficits A Dynamic General
Equilibrium Anaiysls Mlmeo Stanford University
Pereir A 1987 On Ihe Computation of the Users Costs of Stocks Ecooomi (forthcoming)
47
Pereira A bull 1987b Equal Yield Alternatives and Government Oeficlts Mimeof Stanford
University
Pereira A 1987c A Dynamic Applied General ~uHlbrium Model for Tax Policy Evaluation
Comlletamp Docurnent2tion Mimeo University of California San Diego (in progress)
Pereira A 1987d Should the investment Tax Credit be Rei(ltroduced Mlmeo University
of California San Diego (ir progreSs)
Piggott J and J Whalley eds 1985 tJew QeyeQpments in Applied General EQuilibrium
bullenalysjs Cambridge Cambridge University Press
Preckel Pbull 1985 Alternative algorithms for computing economic equilibria In A Manne Ed
Radnor R 19amp3 Equllibrium under Uncertainty In K Arrow M Intrilligaor ods
I ibull
Robinson S 1 g86 Mullisectorai Models of Oeveloping CountrieS a Survey Oivision of
Agricullure and Resources UC Borkeley WP 401
Scarf H 1967 The approximation of fixed points of a continuous mapping ~MM Jouwal of
epplid Mathornallcs 15 pp 1328middot43
Scarf H with T Hansen 1973 00 the Computation of EconomiC Eguilibria New Haven Yale
University Press
Scarf H J Shoven Eds 1984 61l1llied Genera Elujlibrium 6nalyssect Cambridge University
~ Press Cambridge
Shoven J 1976 The Incidence and Efficiency Effects of Taxe on Income from Capital
Jurnal of ramie1 Economy 84 pp 1261-83
Shoven J 1983 Applied General Equilibrium Tax Modemng IMF Slaff PaoerS 30
Shoven J and LB Simon 1986 Share Repurchases and Acquisitions An Analysis of which
Firms Participate NBER Working Paper No 2243
bull
48
IShaven J and J Whalley 1972 ft General Equmbrium Calculation of lhe Effects of
Differential Taxation of Incoma from Capital in the US Jauroe of public Econgmics 1
pp281middot321 (
Shoven J bull and J Whalley 1973 General Equilibrium with Tax$s a computation procedure
and an existence pr00f ReviEhi of Economic Studies 40 pp 475middot490
Shoven J J Whalley 1977 Equal Yield Tax Alternatives A General Equilibrium
Computa~ui0nal Tfchnique J~HJrnal of PUblic EconomIcs 8 pp 211~224 bull
Shaven J J Whalley 1984 Applied GeneralmiddotEquilibrlum MOdels of Taxalion and
International Tr8de An Introduction and Survey Journal of Economic literature 23t pp
1007middot1051
Slemrod J 1980 A General Equilibrium Model of Capital Income Taxation Ph D bullDissertation Harvard Univnrsity
Slemrod J 1983 A General Equilibrium Model of Taxation with Endogenous Financial
Behavior in M Feldstein edbull Behavioral Simulation Methods In Tax Policy Analysis Chicago
UniVerSity of Chicago Press
Stone J 1985 Sequential Optimization and Complementarily Techniques for computing
Economic equilibrium in A Manne Ed
Summers L 1981 Taxation and Corporate Investment a q-Theory Approach BroQkinas
Summers L 1985 Taxation and the Size and ComposHion 01 the Capital Stock An Asset Price
Approach Mlmeo Harvard University
Whalley J 1975 A Geneal Equiloibrium Assessment of the 1973 United Kingdom tax
reform Economica 42 pp 139middot61
1JLTIllOS JIQll~)ltG PAPERS PUBLICADOS bull
I
I I
j
112 70 - LUCENA iJiCIIJo ~ IIEnvironrilent Honitoring and Orqanization Structure T (Setellibro 1987)
nI 71 - L~ENA Diogo A Note on the Representation of Information Stxueture (SeteJObro 1987)
nQ 72 - LUCENA D1CgO Envuorment 111)nitoring and Organization
Structure II (Setelilhro 1987)
ril 73 - AlmmE3 Antonio Pais e CAsPAR Yitor bull Tributacao Inoentivos e InvesUlllentos Analise Qualitativa noYellibro 19S7)
n2 74 - liAlltOSA rose Pedro Opt1lllal lIage R1gidity A SUggestedtlthodology to Teet tb Theory and an Application (Outubro 1987)
nI 75 - CABRAL L 11 B Three lIotes on SYJIIlletric Gemes witt itsYJIIlletric Equilibri (NoveJObro 1987)
n2 76 - LUCENlt D10gOTo Search or Not to Search (Fevereiro1967)
nQ 17 - SA Jorge asconcelos euroI A Theory of SynergyB (Fevereiro1987)
n2 76 - VILARES tanuei Jose Os Be IntenaMios IlIIportados ColIIO Factor de Produao (Julho 1987)
nQ 79 - SA J~rge VasCQncel03 e HoY to Copete and Co~unicate in tature Industrial Products_ (aio 1988)
n2 80 - ROIl Rafael Learning and Capacity Expansion in a New larket Under Uncertainty (Fevereiro 1988)
n2 81 PEREIRA Alfredo llarvilo Survey of Dynamic Computational General EquilibriUlll Hodel for Tax Policy Evaluation (Outubro 1987)
Qualquer in ormacao sabre o 1IIorlltlng Papers ja publlcados sera prestada pelo Secretariado de Apoio aos Docentes podendo os ~e3OSf
ser adquirido5 ne Secyo de Vlndas da Faculdade de EconollLia UNL na Travessa Esteviio Pinto Cltpolide - 1000 LISllOA
bull
35
54 The role 01 consumer expectatlons
The expectations analysis has uncovered one of the benefits of dynamic modelling
Ballard-Goulder (1985) find that the appeal of adopting a consumption tax depends on the level of
foresight possessed by the consumers Furthermore additional foresight may be welfare worsening
This is a second best type of result This tends 10 occur under policies that lead to capital deepening
and declining rate of return to capital over time To the extent that consumers have more foresight
they will be beller equippod to anticipate the fall in the rental price of capital Consequently if a
saving incentive is enacted people save more with myopia than with perfect foresight Given the
exi~tence of taxes on capita and the discrepancy between the private an~--sccial returns to capital
the greater savings with myopia is beller socially Ballard-Goulder find that the welfare gains
from a consumption tax are reduced by about 10 when we move from myopia to farleet foresight
Therefore the attractivenness of adopting a consumption tax seems fairly [obust across these
foresight specifications
I
bull
36
6 Tbe power and weaknesses 01 dynamic modelling the example 01 corporale
tax Integrallon
The Issue of corporate tax integration allows us to sh~w the stre-ngths and weaknesses or the
dynamic approach Empirical evidence using CGE techniques indicates that depending on the precise
scheme of integration and the 1ax replacement methods integratiOn may have substantial effects In
Ihe work of FuliertonmiddotKingmiddotShoven-Whalley (1980 1981) total integration was found to yield an bull
annual static eHielency gain of $4 to $8 billion in 1973 dollars Simulated dynamic gains may be as
large as $695 billion or abcu 14 of the present value of tulUre consumption and leisure in the
US economy These gains result primarily from interindustry reallocalions of Investment and an
improved inter~emporal allocation of consumption
Mora recent work emphasizes hOw consumers asset portfolio decisions and firms finanCial
decisions affect the efficiency gains from integration Slemrod (1980) focusing on consumers
asset ponfoio deCisions finds static efficiency gains which are about twice -as Iarge as those
repOrted iiy Fulierton-King-ShovenWhalley (1961) FuUrlon-Gurdon (1983) focus on the
firms financial decisions They roporl efficiency gains of 6 of GNP from the elimination of the
tax distor1ions favoring debt However when they eliminate the corporate tax and repiace it with
increaSed persona income taxes addlHonal dls10t1ions are created in the optimal labot~leisure
decisions These distortions tend to dominate the analysis slJ(h that the overall effects of complete
integration are very modest Galpr-LuckmiddotTodr (1986) address simultaneously consumers
asset pOrlfollo decisions and the firms financial decisions They also find very modest efficiency
gains from integralion
The above results are important but may be severely biased There are severa aspects of
economic behavior and modelling crucial for the study of Income tax Integration whiCh have no been
i 37
captured innny of the above papers One aspect is the absence of government deficitsmiddot a balanced
budget is assumed It is true Ihal resource crowding OUI is caplured In these mOdels but financial
crowding out induced by government spending is nOI Second investment is not derived from
optimtztlcn behavior Investment behavior passively accomodates endogenous saving decisions As
a consequence the differential impacts of policies in the incentives to save and to invest are not
captured Aso full capital mobility across secors is assumed with instantaneous capital
adjustmenlS towards optimal levels This assumption rules qut different costs of capital across
seclors and therefore differentiated reactions to tax policies changes Finallyhe modelling of both
gov9rnmerlt deficits and of endogenous real and financial investment decisions necessitates the
~nsideraiion of a dynamilt framework and the introduction of financiaJ assets govemment bonds and
iprivate financial assets A dynamic framework also highlights the efficiency effects of Integration on r
ithe ptimal intertemporal decisions The above models are either sIalic (Slemrods and
GalpermiddotltJckemiddotTode(sj or a dynamic sequ~nce of otherwise Slatic model bullbull I Pereir a (1986b) develops a dynamic applied general equilibrium model bull to study the tbull
efficiency effects of integration on the growth and allocation of investmenl across sectors Special
Iattention is paid to the structural affeelS of resource and financial crowding oul induced by
gove mect spending and deficits and to the real and financial Investment reactions across industries
to both tax integration and flnancial crowding out His mode departs from previous work on income
lax Integraticn and for that matler from most 01 the CGIE literalure for tax policy evaluation in
several direcllons It encompasses an endogenous sequential equilibrium struelure founded on
dynamic behavior with flexible expectations Government deficits are optlmaUy determined
Investment decisions are tward looking and the resull of optimizing behavior Several financial
assetsmiddot public and privatemiddot are considered
Simulation resulls Indicale that the welfare gains from Integration under large deficits are at
32
best lJlodest when measured in terms of the GNP The elimination of the corporate tax and its I replacemenl by increased income lax rates yields long run benefits which are never larger than
2 of the presenl yalue ct telr consumption and leisure onder the previous tax regime and 1
under the current tax regime However the short run w~lare effects of full integration tend 10 be
very low and in in some scenarios even negative This is a flew intertemporar pattern of efficiency Ieffects which reflects an adjJsiment lag in the interindustry investment decisions due to the
existence of costs of adjustment On the of her hand partial integration achieved by excluding I dividends from the corporale lax base syslematically yields negalive efficlencyeffeclS In Ihe long Inm these can be as high as 3 of present value of the future value of consumption and leisure This
is a new second best effect suggesllng Ihat less than complete integration may have perverse I efficiency effects
The dynamic slructure of PerclrBs model snowing for an improved freatment of real investment
declsions and government deficts provides a potential setting for a more accurate measure of the
costs and benefits of inlegration The simulation resullS suggest lower benefits and an intertempora Ibull
pattern of growing benefits
Simulation results also clearly suggest robust negative effects from partial integration How
reliable is this result If dJvidends were deductible from the corporate lax base (as are interest
payments) the preferenlial Ireatmant of debl over equily would be ellminaled Corporallons should
be expecled 10 react by decreasing the optimal deWeqully rallo However corporale financial
gecisions are exogenously set In Pereiras model as in aft of the other models surveyed that go as far
as dealing with the issue
Also withdrawing dividends from Ihe corporate tax base is a way of eliminating Ihe double
taxation of dividends at belh Ihe corporate and personnal Income levels How will the optimal
aliocellon of household saving accomodate that change in Ihe relalive return to corporale equity
39 I
t However hOusehold portfolio decisions are exogenouly set in Pereiras model as in all of the other
models surveyeltj go as far as dealing with the issue
IIt is sate to say that the evaluation 01 the benefits 01 partial Integration as defined above are
sevrey underestima1eO Meanwhile the distortionSoenerated by the increase in the marginal
personal tax rates designed to ralse the revenue foregone by the dividend exclusion are fully
aCCltlunted for A good measure of the costs of integration together With a poor evaluation of the
benefils may very well be the reason why partial integration is Simulated 10 yield neg alive
eHikiency gains
On a different vein as most of Ihe model surveyed here Pereiras is a closed economy model It
is legitimate to wonder how the resullS would change If international capital flows were anowed
IThis almost certainly would affect financial crowding out since a good d~al of the capital inflows
could be used to finance public debt
For these various reasons the results should be treated as preliminary but they strongly I suggosl that the dynamic structure provides valuable new insights into the corporate tax integration SSU2
40
7 Concluding remarks I I I i
What has been acccmplished Ihis far The success of Ihe CGE research in laelding the challenge
of dynamic modelling can not be denied
I iGreat progress in the modelling of household behavior has been achieved Promising
developm~nts in the modeUing of production and government behavior have also been accomplished
Interesting innovations in the concept of equilibrium and computation techniques were discussed
The policy analysis was greatly enriched by considering transitional effects together with longer
run steadymiddotstate equilibrium paths generalized equal yield strategies accomodating different
financial crowding out impacts and generalized dynamic policy evaluation indicators
An important set of issues have been addresed by the models surveyed Traditional issues
focusing on the effects of capital taxation and consumption taxes have been pursued In terms of
capital taxation for example the intertemporal nature of the issue was enhanced by allowing an
optimal evolution of capital stock in the economy and forward looking optimal investment decisions
In Andersson (1987) GouldermiddotSummers (1987) and Pereira (1986b 1987d) this is coupled
with a improved treatment of several tax provisions like investment tax credits and depreciation
allowances In terms of the consumption taxes developments in the specification of Intertemporal
household behavior the introduction of overlapping generations and the modelling of
intergenerational links via bequest motives as well as a closer attention to demographic evolution
provided a much improved economic setting for the understanding of the several aspects of the
problem
Furthermore dynamic modelling has permitted the addressing of a set of new issues Policy
Issues ranging from dynamic tax policy analysis to the evaluation of exogenous changes in
government expenditures to the impact of different methods of financing government expenditure to
4 1
the importance of tinanciai croHrii1g out to 1he relevance of consumorS expectations in tne policy
results have been addressed
Despile all of the progress and in part due to such progress several avenues are wide open for
much neded additional research The following seem to be the most Interesting and promising
areasmiddot First the specincation of liquidity constraints to household behavior may help to obtain a
better understanding of policy chenges that affect directly the interest rates in the economy Second
major attention needs to be paid to the incorporafion of a roraign sector (with open capital markets)
Third some efforts are needed in terms of finding adequate computation techniques now that
dynamic modelling eally makes the curse of dimensionality worse than eVOrbull
The modelling of financial markets is the single most unsatisfactory speet of the dynamic
models surveyed here~ The problems associated with provlding a mo~ adequate treatment of
financial markets and in general endogenous and optimal financial deci$ions bull both at the corporate
and at the personal levelsmiddot have 10 be addressed That necessarily requires Introducing uncertainty
into the models To b adequate this must go well beyond the mere assumption of point expeetalions
with the whole array of modelling implementalion and evaluation problems thai generates
bull I i
42
REFERENCES
Abel A and O Blanchard 1983 An Intertemporal Model of Saving and Interest Econometrica
51 pp 675-92
~ Andersson K 1987 Tar-ation of Capital in Sweden - A General Equilibrium Model Ph D
Dissenation UnIversity of Lund Sweder
Auerbach A L KotlikoH 1983 National Savings Economic Welfare and the Structure of
Taxation in M Feldstein ed Behavioral Simulation Methods in Tax Policy- Analysis Chicago
University of Chicago Press
Auerbach A l Kotlikoff 1984 Social Security and the Economics of Demographic-
Transition in H Aaron and G Burtless eds Retirement and Economic Behavior Washington DC
Brookings Institution
Auerbach A L Kotlikoff 1987 Dynamic Fiscal Policy Cambridge Cambridge University
Press
Auerbach A L Kotlikof and J Skinner 1983 The Efficiency Gains from Dynamic Tax
Reform Internatiooaj Economic Reyiew 24 pp 81middot100
Ballard C 1983 Evaluation of the Consumption Tax with Dynamic General Equilibrium
Models Ph D Dissertation Stanford University
Ballard C 1987 Tax Policy and Consumer Foresight A General Equilibrium Simulation
Study Economic Inquiry 25 pp 267-284
Ballard C and L Goulder 1985 Consumption Taxes ForeSight and Welfare a Computatonal
General Equilibrium Analysis in J Piggott and J Whalley eds New Deyelooments in Apolied
General Equilibrium Analysis Cambridge Cambridge University Press
Ballard C D Fullerton J Shoven and J Whalley 1985 A General Equilibrium Model for
t 43
Tax Eoliey Eyalua~iQO NBER University or Chicago Press
Borges A (1987) bull A Survey of Compula1ional General Equilibnum Models CECD Economic
Slud~ 8
Bovenbig L 19B3 imperfect Capital Mobility in a Perfect Foresight Herberger Model
Mimea UC Br~eley
Bovenberg L 1984 Capitel Accumulation and Capital Immobility Qmiddottheory In a Dynamic General Equilibrium Framework PhD dissertation University of California at Berkeley t
Bovenberg L 19850 Dynamic General Equilibrium Tax Models with Adjustment Costs in A iManne ed bull 1985 EtQnomrc ECUjilbrium Model Formulation and Solution Mathematical
lProgramming Siudy 23 Arnsterdam NorthmiddotHoliand
Bovenberg L t 1986 Capital Income Taxation in Growing Open Eeondmies Journa of Pubtic
Eamp(omlcs 3t pp 347middot376 IBovenberg L ald W Keller 1981 Dynamics and Nonlinearlties in Applied General
Equilibrium Models Internal Report Mlmiddot81middot208 Department for Statistical Methods CBS bull
Vooiburg The Netherlands
Charnley C 1981 The Welfare Cost of C~pltal income Taxation In a Growing Economy
IJournal of Political ECQnQm~ 89 3 pp 468middot496 I
Charnley C 19810 Efficient Tax Reform in a Dynamic Model of General Equilibrium I Mimeo Yale University I Charnley C 1982 On the Optimal Taxation In Stylized Models 01 Inlartemporal General
Equilibrium Mimeo Yale University I Decaluw B and A Manens 1985 Pays en Oeveloppement at Modeles Calculables OEquilibre
General - Une Revue de al litterature empirique Monographie Sene 9 Volume 3 Centre de
Recherche et Developpemont Economlque Unrslt de Montreal
44
bull
I I
I
I
1
Erlich $ V Ginsburgh and L van der Heyden 1987 Wher~ do Real Wage Policies Lead
Belgium bull A General Equilibrium Analysis European Economic Reyiew (forthcoming)
air R and J Taylor 1983 Solution and Maximum Likelihood Estimation of Dynamic
Nonlinear Ratioant Expeclations Models Ecooometrica 51 pp 1169middot1186
Feltenstein A 1984 Money and Bonds In a Disaggregated Ec~nonYin Searl 1-1 J Shoven
Eds Aoolied ~eoeral EqJiliDCiulD tioalYsis Cambridge University Press Cambridge
Feltenstein A 19B6 A Dynamic General Equllibrlu Analysis of Financial Crowding Out
Theory with an Application to Australia Journal of Public Economics 31 No1
Fullerlon D R Gorden 1983 A Reexamination of Tax Distorllons in General Equilibrium
Models in M Feldstein Ed Behayora Simulation Methods in Tax pOlicy Analysis Chicago
University of Chipgo Press
Fvllerlon D J Shaven and J Whalley 1983 Replacing the US Income Tax with a
Progressive Consumption Tax JoulOal of Public ECQnomics 20 pp3middot23
Fullerton D Y Henderson J Shaven 1984 A COmparision of Methodologies in Empirical
General Equilibrium Models of Taxation in Searl H J Shoven Eds ~pQlied Geoaral Jguilibrium
lIoalyss Cambridge University Press Cambridge
Fullerton 0 A King J Shaven and J Whaney 1981 Corporate Tax Integration in the United
Statesmiddot a General Equiiibdufi1 Approach American Economic Review 71 pp~ 677~691
Galper H R Lucke E Tader 1985 Taxation portlor Choice and the Allocation of Capital A
General Equilibrium Approach Mimeo
Gill p W Murray M Saunders M Wright 1986 Users Guide for NPSOL (Version 40) A
Fortran Package for Nonlinear Programming T echnicsl Report SOL 86middot2 Deparlment of
Operations Research Stanford University
Glnsbur~h V and L van der Heyden 1985 General Equilibrium with Wage Rigidities An
45
AppJication to Belgium in A Manne ed t Economic EQumprit1m~ Made Formulation and $oJioo
Mathematical Programming Study 23 Amsterdam NorthmiddotHolland ~
Goulet J 1968 1 Adjustrent CO$1S in the Theory o(the Firm Reyiew of EconomiC Studies 35 bull
pp47middot55
Goulder L 1985 Tax-Financed versus BondmiddotFinancedGhanges in Governent Spending
bull Mimeo Harva(d UniversilYmiddot
Goulder l J Shaven and J Whalley 1983 Domestic Tax Policy and the Foreign Sector The
importance of altemative foreign policy formulations to results from a general equilibrium tax
analysis model In ~n methods in Ibe antico of income from caoia1 ed Martin
Feldstein Chicago University of Chicago riSS
Goulder l L Summers 1987 Tax Policy Asset Prices and 3rowtly ~
A General Equilibrium Analysis NBER Working Pper No 212B
Grandmont J 1982 Temporary General Equilibrium Theory Ch 20 in Handbook 01
Mathemalical Eccnornics Vol II NorthmiddotHoliand
Harberger A 1959 The Corporate Income Tax An empirical Appraisal In lax Revision
Comoendjum Voll House Cornmitee on Ways and Means Washington DC Govemment Printing
Office
Harberger A 1962 The Incidence of the Corporate Income Tax JOULn1 of Political
Economy 70 pp 215middot240
Harberger A 1964 Taxation Resource Aliocalrlon and Welfare fn lb Role of Dir bullbulll and
Indjresl Taxes io tbe Fede[al Reserve s~I~Ill Princeton Princeton Unlveresity Press
James J 1985 New Developments in the Applications of General Equilibrium models to
Economic History in Piggott aand Whalley Eds
Jorgenson D and K Yun 1984 Tax Policy and Capital Allocation Harvard Inslitute of
bull
I
I f
Economic Research WP 1107
Judd K 1965 ShonmiddotRun Analysis of Fiscal Policy in a Simple Perfect Foresight Model
Journal of poJiHcal 5CCJiW1l pp 298middot319 r
Upton D J Potbc J Sachsa nd L Summers 1982 Multiple Shooting in Rational
Expectations Models fuoometrlc 50 1329middot1333
Manne A ed 1985 fconQrnjc Eoullibrium Model FOrIDujaljQo and Solution Mathematical
Pogrammlng Study 23 Amsterdam NorthmiddotHolland
Manne A 1985 Or the Formulation and Solution 01 Economic Equilibrium Models In A
Manne ed ECQOQlli-Jr~ibmiddotum ~tQdel EQlJulailon and Solution Malhematical Programming
Study 23 Amsterdam NorthmiddotHolland
Mansur A and J Whalley 1984 Numerical Specification of Applied General Equilibrium
Models in Scarf H J Shoven Eds Aoplied General Eouillbrium AnaiysectIsect Cambridge Universily
Press Cambridge
Merrill 0 1972 Applications and Extensions of an Algorithm thaI Computes FixedmiddotPoints of
Cenain Upper Semimiddotcominuou Point to Set Mappings PhD Disserlation Unlvel$ity 01 Michigan
Negishi T 1960 Welfar Economics and Exislence I an Equnlbrlum for a Competitive
Economy MelroeCOOQ11ka 12 92middot97
Pereira A 1985 On the Existence and Uniqueness of Optimal Ou1put Path lor CRTS Firms in
Adjustment Costs Technologies Mimeo Stanford University
Pereira A 1985a Consumer Expectations and Ihe COSI 01 UllUty In an Intertemporat
Framework Mlmeo Stanford University
Pereira A 1985b Corporale Tax Integrallon and Government Deficits A Dynamic General
Equilibrium Anaiysls Mlmeo Stanford University
Pereir A 1987 On Ihe Computation of the Users Costs of Stocks Ecooomi (forthcoming)
47
Pereira A bull 1987b Equal Yield Alternatives and Government Oeficlts Mimeof Stanford
University
Pereira A 1987c A Dynamic Applied General ~uHlbrium Model for Tax Policy Evaluation
Comlletamp Docurnent2tion Mimeo University of California San Diego (in progress)
Pereira A 1987d Should the investment Tax Credit be Rei(ltroduced Mlmeo University
of California San Diego (ir progreSs)
Piggott J and J Whalley eds 1985 tJew QeyeQpments in Applied General EQuilibrium
bullenalysjs Cambridge Cambridge University Press
Preckel Pbull 1985 Alternative algorithms for computing economic equilibria In A Manne Ed
Radnor R 19amp3 Equllibrium under Uncertainty In K Arrow M Intrilligaor ods
I ibull
Robinson S 1 g86 Mullisectorai Models of Oeveloping CountrieS a Survey Oivision of
Agricullure and Resources UC Borkeley WP 401
Scarf H 1967 The approximation of fixed points of a continuous mapping ~MM Jouwal of
epplid Mathornallcs 15 pp 1328middot43
Scarf H with T Hansen 1973 00 the Computation of EconomiC Eguilibria New Haven Yale
University Press
Scarf H J Shoven Eds 1984 61l1llied Genera Elujlibrium 6nalyssect Cambridge University
~ Press Cambridge
Shoven J 1976 The Incidence and Efficiency Effects of Taxe on Income from Capital
Jurnal of ramie1 Economy 84 pp 1261-83
Shoven J 1983 Applied General Equilibrium Tax Modemng IMF Slaff PaoerS 30
Shoven J and LB Simon 1986 Share Repurchases and Acquisitions An Analysis of which
Firms Participate NBER Working Paper No 2243
bull
48
IShaven J and J Whalley 1972 ft General Equmbrium Calculation of lhe Effects of
Differential Taxation of Incoma from Capital in the US Jauroe of public Econgmics 1
pp281middot321 (
Shoven J bull and J Whalley 1973 General Equilibrium with Tax$s a computation procedure
and an existence pr00f ReviEhi of Economic Studies 40 pp 475middot490
Shoven J J Whalley 1977 Equal Yield Tax Alternatives A General Equilibrium
Computa~ui0nal Tfchnique J~HJrnal of PUblic EconomIcs 8 pp 211~224 bull
Shaven J J Whalley 1984 Applied GeneralmiddotEquilibrlum MOdels of Taxalion and
International Tr8de An Introduction and Survey Journal of Economic literature 23t pp
1007middot1051
Slemrod J 1980 A General Equilibrium Model of Capital Income Taxation Ph D bullDissertation Harvard Univnrsity
Slemrod J 1983 A General Equilibrium Model of Taxation with Endogenous Financial
Behavior in M Feldstein edbull Behavioral Simulation Methods In Tax Policy Analysis Chicago
UniVerSity of Chicago Press
Stone J 1985 Sequential Optimization and Complementarily Techniques for computing
Economic equilibrium in A Manne Ed
Summers L 1981 Taxation and Corporate Investment a q-Theory Approach BroQkinas
Summers L 1985 Taxation and the Size and ComposHion 01 the Capital Stock An Asset Price
Approach Mlmeo Harvard University
Whalley J 1975 A Geneal Equiloibrium Assessment of the 1973 United Kingdom tax
reform Economica 42 pp 139middot61
1JLTIllOS JIQll~)ltG PAPERS PUBLICADOS bull
I
I I
j
112 70 - LUCENA iJiCIIJo ~ IIEnvironrilent Honitoring and Orqanization Structure T (Setellibro 1987)
nI 71 - L~ENA Diogo A Note on the Representation of Information Stxueture (SeteJObro 1987)
nQ 72 - LUCENA D1CgO Envuorment 111)nitoring and Organization
Structure II (Setelilhro 1987)
ril 73 - AlmmE3 Antonio Pais e CAsPAR Yitor bull Tributacao Inoentivos e InvesUlllentos Analise Qualitativa noYellibro 19S7)
n2 74 - liAlltOSA rose Pedro Opt1lllal lIage R1gidity A SUggestedtlthodology to Teet tb Theory and an Application (Outubro 1987)
nI 75 - CABRAL L 11 B Three lIotes on SYJIIlletric Gemes witt itsYJIIlletric Equilibri (NoveJObro 1987)
n2 76 - LUCENlt D10gOTo Search or Not to Search (Fevereiro1967)
nQ 17 - SA Jorge asconcelos euroI A Theory of SynergyB (Fevereiro1987)
n2 76 - VILARES tanuei Jose Os Be IntenaMios IlIIportados ColIIO Factor de Produao (Julho 1987)
nQ 79 - SA J~rge VasCQncel03 e HoY to Copete and Co~unicate in tature Industrial Products_ (aio 1988)
n2 80 - ROIl Rafael Learning and Capacity Expansion in a New larket Under Uncertainty (Fevereiro 1988)
n2 81 PEREIRA Alfredo llarvilo Survey of Dynamic Computational General EquilibriUlll Hodel for Tax Policy Evaluation (Outubro 1987)
Qualquer in ormacao sabre o 1IIorlltlng Papers ja publlcados sera prestada pelo Secretariado de Apoio aos Docentes podendo os ~e3OSf
ser adquirido5 ne Secyo de Vlndas da Faculdade de EconollLia UNL na Travessa Esteviio Pinto Cltpolide - 1000 LISllOA
I
bull
36
6 Tbe power and weaknesses 01 dynamic modelling the example 01 corporale
tax Integrallon
The Issue of corporate tax integration allows us to sh~w the stre-ngths and weaknesses or the
dynamic approach Empirical evidence using CGE techniques indicates that depending on the precise
scheme of integration and the 1ax replacement methods integratiOn may have substantial effects In
Ihe work of FuliertonmiddotKingmiddotShoven-Whalley (1980 1981) total integration was found to yield an bull
annual static eHielency gain of $4 to $8 billion in 1973 dollars Simulated dynamic gains may be as
large as $695 billion or abcu 14 of the present value of tulUre consumption and leisure in the
US economy These gains result primarily from interindustry reallocalions of Investment and an
improved inter~emporal allocation of consumption
Mora recent work emphasizes hOw consumers asset portfolio decisions and firms finanCial
decisions affect the efficiency gains from integration Slemrod (1980) focusing on consumers
asset ponfoio deCisions finds static efficiency gains which are about twice -as Iarge as those
repOrted iiy Fulierton-King-ShovenWhalley (1961) FuUrlon-Gurdon (1983) focus on the
firms financial decisions They roporl efficiency gains of 6 of GNP from the elimination of the
tax distor1ions favoring debt However when they eliminate the corporate tax and repiace it with
increaSed persona income taxes addlHonal dls10t1ions are created in the optimal labot~leisure
decisions These distortions tend to dominate the analysis slJ(h that the overall effects of complete
integration are very modest Galpr-LuckmiddotTodr (1986) address simultaneously consumers
asset pOrlfollo decisions and the firms financial decisions They also find very modest efficiency
gains from integralion
The above results are important but may be severely biased There are severa aspects of
economic behavior and modelling crucial for the study of Income tax Integration whiCh have no been
i 37
captured innny of the above papers One aspect is the absence of government deficitsmiddot a balanced
budget is assumed It is true Ihal resource crowding OUI is caplured In these mOdels but financial
crowding out induced by government spending is nOI Second investment is not derived from
optimtztlcn behavior Investment behavior passively accomodates endogenous saving decisions As
a consequence the differential impacts of policies in the incentives to save and to invest are not
captured Aso full capital mobility across secors is assumed with instantaneous capital
adjustmenlS towards optimal levels This assumption rules qut different costs of capital across
seclors and therefore differentiated reactions to tax policies changes Finallyhe modelling of both
gov9rnmerlt deficits and of endogenous real and financial investment decisions necessitates the
~nsideraiion of a dynamilt framework and the introduction of financiaJ assets govemment bonds and
iprivate financial assets A dynamic framework also highlights the efficiency effects of Integration on r
ithe ptimal intertemporal decisions The above models are either sIalic (Slemrods and
GalpermiddotltJckemiddotTode(sj or a dynamic sequ~nce of otherwise Slatic model bullbull I Pereir a (1986b) develops a dynamic applied general equilibrium model bull to study the tbull
efficiency effects of integration on the growth and allocation of investmenl across sectors Special
Iattention is paid to the structural affeelS of resource and financial crowding oul induced by
gove mect spending and deficits and to the real and financial Investment reactions across industries
to both tax integration and flnancial crowding out His mode departs from previous work on income
lax Integraticn and for that matler from most 01 the CGIE literalure for tax policy evaluation in
several direcllons It encompasses an endogenous sequential equilibrium struelure founded on
dynamic behavior with flexible expectations Government deficits are optlmaUy determined
Investment decisions are tward looking and the resull of optimizing behavior Several financial
assetsmiddot public and privatemiddot are considered
Simulation resulls Indicale that the welfare gains from Integration under large deficits are at
32
best lJlodest when measured in terms of the GNP The elimination of the corporate tax and its I replacemenl by increased income lax rates yields long run benefits which are never larger than
2 of the presenl yalue ct telr consumption and leisure onder the previous tax regime and 1
under the current tax regime However the short run w~lare effects of full integration tend 10 be
very low and in in some scenarios even negative This is a flew intertemporar pattern of efficiency Ieffects which reflects an adjJsiment lag in the interindustry investment decisions due to the
existence of costs of adjustment On the of her hand partial integration achieved by excluding I dividends from the corporale lax base syslematically yields negalive efficlencyeffeclS In Ihe long Inm these can be as high as 3 of present value of the future value of consumption and leisure This
is a new second best effect suggesllng Ihat less than complete integration may have perverse I efficiency effects
The dynamic slructure of PerclrBs model snowing for an improved freatment of real investment
declsions and government deficts provides a potential setting for a more accurate measure of the
costs and benefits of inlegration The simulation resullS suggest lower benefits and an intertempora Ibull
pattern of growing benefits
Simulation results also clearly suggest robust negative effects from partial integration How
reliable is this result If dJvidends were deductible from the corporate lax base (as are interest
payments) the preferenlial Ireatmant of debl over equily would be ellminaled Corporallons should
be expecled 10 react by decreasing the optimal deWeqully rallo However corporale financial
gecisions are exogenously set In Pereiras model as in aft of the other models surveyed that go as far
as dealing with the issue
Also withdrawing dividends from Ihe corporate tax base is a way of eliminating Ihe double
taxation of dividends at belh Ihe corporate and personnal Income levels How will the optimal
aliocellon of household saving accomodate that change in Ihe relalive return to corporale equity
39 I
t However hOusehold portfolio decisions are exogenouly set in Pereiras model as in all of the other
models surveyeltj go as far as dealing with the issue
IIt is sate to say that the evaluation 01 the benefits 01 partial Integration as defined above are
sevrey underestima1eO Meanwhile the distortionSoenerated by the increase in the marginal
personal tax rates designed to ralse the revenue foregone by the dividend exclusion are fully
aCCltlunted for A good measure of the costs of integration together With a poor evaluation of the
benefils may very well be the reason why partial integration is Simulated 10 yield neg alive
eHikiency gains
On a different vein as most of Ihe model surveyed here Pereiras is a closed economy model It
is legitimate to wonder how the resullS would change If international capital flows were anowed
IThis almost certainly would affect financial crowding out since a good d~al of the capital inflows
could be used to finance public debt
For these various reasons the results should be treated as preliminary but they strongly I suggosl that the dynamic structure provides valuable new insights into the corporate tax integration SSU2
40
7 Concluding remarks I I I i
What has been acccmplished Ihis far The success of Ihe CGE research in laelding the challenge
of dynamic modelling can not be denied
I iGreat progress in the modelling of household behavior has been achieved Promising
developm~nts in the modeUing of production and government behavior have also been accomplished
Interesting innovations in the concept of equilibrium and computation techniques were discussed
The policy analysis was greatly enriched by considering transitional effects together with longer
run steadymiddotstate equilibrium paths generalized equal yield strategies accomodating different
financial crowding out impacts and generalized dynamic policy evaluation indicators
An important set of issues have been addresed by the models surveyed Traditional issues
focusing on the effects of capital taxation and consumption taxes have been pursued In terms of
capital taxation for example the intertemporal nature of the issue was enhanced by allowing an
optimal evolution of capital stock in the economy and forward looking optimal investment decisions
In Andersson (1987) GouldermiddotSummers (1987) and Pereira (1986b 1987d) this is coupled
with a improved treatment of several tax provisions like investment tax credits and depreciation
allowances In terms of the consumption taxes developments in the specification of Intertemporal
household behavior the introduction of overlapping generations and the modelling of
intergenerational links via bequest motives as well as a closer attention to demographic evolution
provided a much improved economic setting for the understanding of the several aspects of the
problem
Furthermore dynamic modelling has permitted the addressing of a set of new issues Policy
Issues ranging from dynamic tax policy analysis to the evaluation of exogenous changes in
government expenditures to the impact of different methods of financing government expenditure to
4 1
the importance of tinanciai croHrii1g out to 1he relevance of consumorS expectations in tne policy
results have been addressed
Despile all of the progress and in part due to such progress several avenues are wide open for
much neded additional research The following seem to be the most Interesting and promising
areasmiddot First the specincation of liquidity constraints to household behavior may help to obtain a
better understanding of policy chenges that affect directly the interest rates in the economy Second
major attention needs to be paid to the incorporafion of a roraign sector (with open capital markets)
Third some efforts are needed in terms of finding adequate computation techniques now that
dynamic modelling eally makes the curse of dimensionality worse than eVOrbull
The modelling of financial markets is the single most unsatisfactory speet of the dynamic
models surveyed here~ The problems associated with provlding a mo~ adequate treatment of
financial markets and in general endogenous and optimal financial deci$ions bull both at the corporate
and at the personal levelsmiddot have 10 be addressed That necessarily requires Introducing uncertainty
into the models To b adequate this must go well beyond the mere assumption of point expeetalions
with the whole array of modelling implementalion and evaluation problems thai generates
bull I i
42
REFERENCES
Abel A and O Blanchard 1983 An Intertemporal Model of Saving and Interest Econometrica
51 pp 675-92
~ Andersson K 1987 Tar-ation of Capital in Sweden - A General Equilibrium Model Ph D
Dissenation UnIversity of Lund Sweder
Auerbach A L KotlikoH 1983 National Savings Economic Welfare and the Structure of
Taxation in M Feldstein ed Behavioral Simulation Methods in Tax Policy- Analysis Chicago
University of Chicago Press
Auerbach A l Kotlikoff 1984 Social Security and the Economics of Demographic-
Transition in H Aaron and G Burtless eds Retirement and Economic Behavior Washington DC
Brookings Institution
Auerbach A L Kotlikoff 1987 Dynamic Fiscal Policy Cambridge Cambridge University
Press
Auerbach A L Kotlikof and J Skinner 1983 The Efficiency Gains from Dynamic Tax
Reform Internatiooaj Economic Reyiew 24 pp 81middot100
Ballard C 1983 Evaluation of the Consumption Tax with Dynamic General Equilibrium
Models Ph D Dissertation Stanford University
Ballard C 1987 Tax Policy and Consumer Foresight A General Equilibrium Simulation
Study Economic Inquiry 25 pp 267-284
Ballard C and L Goulder 1985 Consumption Taxes ForeSight and Welfare a Computatonal
General Equilibrium Analysis in J Piggott and J Whalley eds New Deyelooments in Apolied
General Equilibrium Analysis Cambridge Cambridge University Press
Ballard C D Fullerton J Shoven and J Whalley 1985 A General Equilibrium Model for
t 43
Tax Eoliey Eyalua~iQO NBER University or Chicago Press
Borges A (1987) bull A Survey of Compula1ional General Equilibnum Models CECD Economic
Slud~ 8
Bovenbig L 19B3 imperfect Capital Mobility in a Perfect Foresight Herberger Model
Mimea UC Br~eley
Bovenberg L 1984 Capitel Accumulation and Capital Immobility Qmiddottheory In a Dynamic General Equilibrium Framework PhD dissertation University of California at Berkeley t
Bovenberg L 19850 Dynamic General Equilibrium Tax Models with Adjustment Costs in A iManne ed bull 1985 EtQnomrc ECUjilbrium Model Formulation and Solution Mathematical
lProgramming Siudy 23 Arnsterdam NorthmiddotHoliand
Bovenberg L t 1986 Capital Income Taxation in Growing Open Eeondmies Journa of Pubtic
Eamp(omlcs 3t pp 347middot376 IBovenberg L ald W Keller 1981 Dynamics and Nonlinearlties in Applied General
Equilibrium Models Internal Report Mlmiddot81middot208 Department for Statistical Methods CBS bull
Vooiburg The Netherlands
Charnley C 1981 The Welfare Cost of C~pltal income Taxation In a Growing Economy
IJournal of Political ECQnQm~ 89 3 pp 468middot496 I
Charnley C 19810 Efficient Tax Reform in a Dynamic Model of General Equilibrium I Mimeo Yale University I Charnley C 1982 On the Optimal Taxation In Stylized Models 01 Inlartemporal General
Equilibrium Mimeo Yale University I Decaluw B and A Manens 1985 Pays en Oeveloppement at Modeles Calculables OEquilibre
General - Une Revue de al litterature empirique Monographie Sene 9 Volume 3 Centre de
Recherche et Developpemont Economlque Unrslt de Montreal
44
bull
I I
I
I
1
Erlich $ V Ginsburgh and L van der Heyden 1987 Wher~ do Real Wage Policies Lead
Belgium bull A General Equilibrium Analysis European Economic Reyiew (forthcoming)
air R and J Taylor 1983 Solution and Maximum Likelihood Estimation of Dynamic
Nonlinear Ratioant Expeclations Models Ecooometrica 51 pp 1169middot1186
Feltenstein A 1984 Money and Bonds In a Disaggregated Ec~nonYin Searl 1-1 J Shoven
Eds Aoolied ~eoeral EqJiliDCiulD tioalYsis Cambridge University Press Cambridge
Feltenstein A 19B6 A Dynamic General Equllibrlu Analysis of Financial Crowding Out
Theory with an Application to Australia Journal of Public Economics 31 No1
Fullerlon D R Gorden 1983 A Reexamination of Tax Distorllons in General Equilibrium
Models in M Feldstein Ed Behayora Simulation Methods in Tax pOlicy Analysis Chicago
University of Chipgo Press
Fvllerlon D J Shaven and J Whalley 1983 Replacing the US Income Tax with a
Progressive Consumption Tax JoulOal of Public ECQnomics 20 pp3middot23
Fullerton D Y Henderson J Shaven 1984 A COmparision of Methodologies in Empirical
General Equilibrium Models of Taxation in Searl H J Shoven Eds ~pQlied Geoaral Jguilibrium
lIoalyss Cambridge University Press Cambridge
Fullerton 0 A King J Shaven and J Whaney 1981 Corporate Tax Integration in the United
Statesmiddot a General Equiiibdufi1 Approach American Economic Review 71 pp~ 677~691
Galper H R Lucke E Tader 1985 Taxation portlor Choice and the Allocation of Capital A
General Equilibrium Approach Mimeo
Gill p W Murray M Saunders M Wright 1986 Users Guide for NPSOL (Version 40) A
Fortran Package for Nonlinear Programming T echnicsl Report SOL 86middot2 Deparlment of
Operations Research Stanford University
Glnsbur~h V and L van der Heyden 1985 General Equilibrium with Wage Rigidities An
45
AppJication to Belgium in A Manne ed t Economic EQumprit1m~ Made Formulation and $oJioo
Mathematical Programming Study 23 Amsterdam NorthmiddotHolland ~
Goulet J 1968 1 Adjustrent CO$1S in the Theory o(the Firm Reyiew of EconomiC Studies 35 bull
pp47middot55
Goulder L 1985 Tax-Financed versus BondmiddotFinancedGhanges in Governent Spending
bull Mimeo Harva(d UniversilYmiddot
Goulder l J Shaven and J Whalley 1983 Domestic Tax Policy and the Foreign Sector The
importance of altemative foreign policy formulations to results from a general equilibrium tax
analysis model In ~n methods in Ibe antico of income from caoia1 ed Martin
Feldstein Chicago University of Chicago riSS
Goulder l L Summers 1987 Tax Policy Asset Prices and 3rowtly ~
A General Equilibrium Analysis NBER Working Pper No 212B
Grandmont J 1982 Temporary General Equilibrium Theory Ch 20 in Handbook 01
Mathemalical Eccnornics Vol II NorthmiddotHoliand
Harberger A 1959 The Corporate Income Tax An empirical Appraisal In lax Revision
Comoendjum Voll House Cornmitee on Ways and Means Washington DC Govemment Printing
Office
Harberger A 1962 The Incidence of the Corporate Income Tax JOULn1 of Political
Economy 70 pp 215middot240
Harberger A 1964 Taxation Resource Aliocalrlon and Welfare fn lb Role of Dir bullbulll and
Indjresl Taxes io tbe Fede[al Reserve s~I~Ill Princeton Princeton Unlveresity Press
James J 1985 New Developments in the Applications of General Equilibrium models to
Economic History in Piggott aand Whalley Eds
Jorgenson D and K Yun 1984 Tax Policy and Capital Allocation Harvard Inslitute of
bull
I
I f
Economic Research WP 1107
Judd K 1965 ShonmiddotRun Analysis of Fiscal Policy in a Simple Perfect Foresight Model
Journal of poJiHcal 5CCJiW1l pp 298middot319 r
Upton D J Potbc J Sachsa nd L Summers 1982 Multiple Shooting in Rational
Expectations Models fuoometrlc 50 1329middot1333
Manne A ed 1985 fconQrnjc Eoullibrium Model FOrIDujaljQo and Solution Mathematical
Pogrammlng Study 23 Amsterdam NorthmiddotHolland
Manne A 1985 Or the Formulation and Solution 01 Economic Equilibrium Models In A
Manne ed ECQOQlli-Jr~ibmiddotum ~tQdel EQlJulailon and Solution Malhematical Programming
Study 23 Amsterdam NorthmiddotHolland
Mansur A and J Whalley 1984 Numerical Specification of Applied General Equilibrium
Models in Scarf H J Shoven Eds Aoplied General Eouillbrium AnaiysectIsect Cambridge Universily
Press Cambridge
Merrill 0 1972 Applications and Extensions of an Algorithm thaI Computes FixedmiddotPoints of
Cenain Upper Semimiddotcominuou Point to Set Mappings PhD Disserlation Unlvel$ity 01 Michigan
Negishi T 1960 Welfar Economics and Exislence I an Equnlbrlum for a Competitive
Economy MelroeCOOQ11ka 12 92middot97
Pereira A 1985 On the Existence and Uniqueness of Optimal Ou1put Path lor CRTS Firms in
Adjustment Costs Technologies Mimeo Stanford University
Pereira A 1985a Consumer Expectations and Ihe COSI 01 UllUty In an Intertemporat
Framework Mlmeo Stanford University
Pereira A 1985b Corporale Tax Integrallon and Government Deficits A Dynamic General
Equilibrium Anaiysls Mlmeo Stanford University
Pereir A 1987 On Ihe Computation of the Users Costs of Stocks Ecooomi (forthcoming)
47
Pereira A bull 1987b Equal Yield Alternatives and Government Oeficlts Mimeof Stanford
University
Pereira A 1987c A Dynamic Applied General ~uHlbrium Model for Tax Policy Evaluation
Comlletamp Docurnent2tion Mimeo University of California San Diego (in progress)
Pereira A 1987d Should the investment Tax Credit be Rei(ltroduced Mlmeo University
of California San Diego (ir progreSs)
Piggott J and J Whalley eds 1985 tJew QeyeQpments in Applied General EQuilibrium
bullenalysjs Cambridge Cambridge University Press
Preckel Pbull 1985 Alternative algorithms for computing economic equilibria In A Manne Ed
Radnor R 19amp3 Equllibrium under Uncertainty In K Arrow M Intrilligaor ods
I ibull
Robinson S 1 g86 Mullisectorai Models of Oeveloping CountrieS a Survey Oivision of
Agricullure and Resources UC Borkeley WP 401
Scarf H 1967 The approximation of fixed points of a continuous mapping ~MM Jouwal of
epplid Mathornallcs 15 pp 1328middot43
Scarf H with T Hansen 1973 00 the Computation of EconomiC Eguilibria New Haven Yale
University Press
Scarf H J Shoven Eds 1984 61l1llied Genera Elujlibrium 6nalyssect Cambridge University
~ Press Cambridge
Shoven J 1976 The Incidence and Efficiency Effects of Taxe on Income from Capital
Jurnal of ramie1 Economy 84 pp 1261-83
Shoven J 1983 Applied General Equilibrium Tax Modemng IMF Slaff PaoerS 30
Shoven J and LB Simon 1986 Share Repurchases and Acquisitions An Analysis of which
Firms Participate NBER Working Paper No 2243
bull
48
IShaven J and J Whalley 1972 ft General Equmbrium Calculation of lhe Effects of
Differential Taxation of Incoma from Capital in the US Jauroe of public Econgmics 1
pp281middot321 (
Shoven J bull and J Whalley 1973 General Equilibrium with Tax$s a computation procedure
and an existence pr00f ReviEhi of Economic Studies 40 pp 475middot490
Shoven J J Whalley 1977 Equal Yield Tax Alternatives A General Equilibrium
Computa~ui0nal Tfchnique J~HJrnal of PUblic EconomIcs 8 pp 211~224 bull
Shaven J J Whalley 1984 Applied GeneralmiddotEquilibrlum MOdels of Taxalion and
International Tr8de An Introduction and Survey Journal of Economic literature 23t pp
1007middot1051
Slemrod J 1980 A General Equilibrium Model of Capital Income Taxation Ph D bullDissertation Harvard Univnrsity
Slemrod J 1983 A General Equilibrium Model of Taxation with Endogenous Financial
Behavior in M Feldstein edbull Behavioral Simulation Methods In Tax Policy Analysis Chicago
UniVerSity of Chicago Press
Stone J 1985 Sequential Optimization and Complementarily Techniques for computing
Economic equilibrium in A Manne Ed
Summers L 1981 Taxation and Corporate Investment a q-Theory Approach BroQkinas
Summers L 1985 Taxation and the Size and ComposHion 01 the Capital Stock An Asset Price
Approach Mlmeo Harvard University
Whalley J 1975 A Geneal Equiloibrium Assessment of the 1973 United Kingdom tax
reform Economica 42 pp 139middot61
1JLTIllOS JIQll~)ltG PAPERS PUBLICADOS bull
I
I I
j
112 70 - LUCENA iJiCIIJo ~ IIEnvironrilent Honitoring and Orqanization Structure T (Setellibro 1987)
nI 71 - L~ENA Diogo A Note on the Representation of Information Stxueture (SeteJObro 1987)
nQ 72 - LUCENA D1CgO Envuorment 111)nitoring and Organization
Structure II (Setelilhro 1987)
ril 73 - AlmmE3 Antonio Pais e CAsPAR Yitor bull Tributacao Inoentivos e InvesUlllentos Analise Qualitativa noYellibro 19S7)
n2 74 - liAlltOSA rose Pedro Opt1lllal lIage R1gidity A SUggestedtlthodology to Teet tb Theory and an Application (Outubro 1987)
nI 75 - CABRAL L 11 B Three lIotes on SYJIIlletric Gemes witt itsYJIIlletric Equilibri (NoveJObro 1987)
n2 76 - LUCENlt D10gOTo Search or Not to Search (Fevereiro1967)
nQ 17 - SA Jorge asconcelos euroI A Theory of SynergyB (Fevereiro1987)
n2 76 - VILARES tanuei Jose Os Be IntenaMios IlIIportados ColIIO Factor de Produao (Julho 1987)
nQ 79 - SA J~rge VasCQncel03 e HoY to Copete and Co~unicate in tature Industrial Products_ (aio 1988)
n2 80 - ROIl Rafael Learning and Capacity Expansion in a New larket Under Uncertainty (Fevereiro 1988)
n2 81 PEREIRA Alfredo llarvilo Survey of Dynamic Computational General EquilibriUlll Hodel for Tax Policy Evaluation (Outubro 1987)
Qualquer in ormacao sabre o 1IIorlltlng Papers ja publlcados sera prestada pelo Secretariado de Apoio aos Docentes podendo os ~e3OSf
ser adquirido5 ne Secyo de Vlndas da Faculdade de EconollLia UNL na Travessa Esteviio Pinto Cltpolide - 1000 LISllOA
i 37
captured innny of the above papers One aspect is the absence of government deficitsmiddot a balanced
budget is assumed It is true Ihal resource crowding OUI is caplured In these mOdels but financial
crowding out induced by government spending is nOI Second investment is not derived from
optimtztlcn behavior Investment behavior passively accomodates endogenous saving decisions As
a consequence the differential impacts of policies in the incentives to save and to invest are not
captured Aso full capital mobility across secors is assumed with instantaneous capital
adjustmenlS towards optimal levels This assumption rules qut different costs of capital across
seclors and therefore differentiated reactions to tax policies changes Finallyhe modelling of both
gov9rnmerlt deficits and of endogenous real and financial investment decisions necessitates the
~nsideraiion of a dynamilt framework and the introduction of financiaJ assets govemment bonds and
iprivate financial assets A dynamic framework also highlights the efficiency effects of Integration on r
ithe ptimal intertemporal decisions The above models are either sIalic (Slemrods and
GalpermiddotltJckemiddotTode(sj or a dynamic sequ~nce of otherwise Slatic model bullbull I Pereir a (1986b) develops a dynamic applied general equilibrium model bull to study the tbull
efficiency effects of integration on the growth and allocation of investmenl across sectors Special
Iattention is paid to the structural affeelS of resource and financial crowding oul induced by
gove mect spending and deficits and to the real and financial Investment reactions across industries
to both tax integration and flnancial crowding out His mode departs from previous work on income
lax Integraticn and for that matler from most 01 the CGIE literalure for tax policy evaluation in
several direcllons It encompasses an endogenous sequential equilibrium struelure founded on
dynamic behavior with flexible expectations Government deficits are optlmaUy determined
Investment decisions are tward looking and the resull of optimizing behavior Several financial
assetsmiddot public and privatemiddot are considered
Simulation resulls Indicale that the welfare gains from Integration under large deficits are at
32
best lJlodest when measured in terms of the GNP The elimination of the corporate tax and its I replacemenl by increased income lax rates yields long run benefits which are never larger than
2 of the presenl yalue ct telr consumption and leisure onder the previous tax regime and 1
under the current tax regime However the short run w~lare effects of full integration tend 10 be
very low and in in some scenarios even negative This is a flew intertemporar pattern of efficiency Ieffects which reflects an adjJsiment lag in the interindustry investment decisions due to the
existence of costs of adjustment On the of her hand partial integration achieved by excluding I dividends from the corporale lax base syslematically yields negalive efficlencyeffeclS In Ihe long Inm these can be as high as 3 of present value of the future value of consumption and leisure This
is a new second best effect suggesllng Ihat less than complete integration may have perverse I efficiency effects
The dynamic slructure of PerclrBs model snowing for an improved freatment of real investment
declsions and government deficts provides a potential setting for a more accurate measure of the
costs and benefits of inlegration The simulation resullS suggest lower benefits and an intertempora Ibull
pattern of growing benefits
Simulation results also clearly suggest robust negative effects from partial integration How
reliable is this result If dJvidends were deductible from the corporate lax base (as are interest
payments) the preferenlial Ireatmant of debl over equily would be ellminaled Corporallons should
be expecled 10 react by decreasing the optimal deWeqully rallo However corporale financial
gecisions are exogenously set In Pereiras model as in aft of the other models surveyed that go as far
as dealing with the issue
Also withdrawing dividends from Ihe corporate tax base is a way of eliminating Ihe double
taxation of dividends at belh Ihe corporate and personnal Income levels How will the optimal
aliocellon of household saving accomodate that change in Ihe relalive return to corporale equity
39 I
t However hOusehold portfolio decisions are exogenouly set in Pereiras model as in all of the other
models surveyeltj go as far as dealing with the issue
IIt is sate to say that the evaluation 01 the benefits 01 partial Integration as defined above are
sevrey underestima1eO Meanwhile the distortionSoenerated by the increase in the marginal
personal tax rates designed to ralse the revenue foregone by the dividend exclusion are fully
aCCltlunted for A good measure of the costs of integration together With a poor evaluation of the
benefils may very well be the reason why partial integration is Simulated 10 yield neg alive
eHikiency gains
On a different vein as most of Ihe model surveyed here Pereiras is a closed economy model It
is legitimate to wonder how the resullS would change If international capital flows were anowed
IThis almost certainly would affect financial crowding out since a good d~al of the capital inflows
could be used to finance public debt
For these various reasons the results should be treated as preliminary but they strongly I suggosl that the dynamic structure provides valuable new insights into the corporate tax integration SSU2
40
7 Concluding remarks I I I i
What has been acccmplished Ihis far The success of Ihe CGE research in laelding the challenge
of dynamic modelling can not be denied
I iGreat progress in the modelling of household behavior has been achieved Promising
developm~nts in the modeUing of production and government behavior have also been accomplished
Interesting innovations in the concept of equilibrium and computation techniques were discussed
The policy analysis was greatly enriched by considering transitional effects together with longer
run steadymiddotstate equilibrium paths generalized equal yield strategies accomodating different
financial crowding out impacts and generalized dynamic policy evaluation indicators
An important set of issues have been addresed by the models surveyed Traditional issues
focusing on the effects of capital taxation and consumption taxes have been pursued In terms of
capital taxation for example the intertemporal nature of the issue was enhanced by allowing an
optimal evolution of capital stock in the economy and forward looking optimal investment decisions
In Andersson (1987) GouldermiddotSummers (1987) and Pereira (1986b 1987d) this is coupled
with a improved treatment of several tax provisions like investment tax credits and depreciation
allowances In terms of the consumption taxes developments in the specification of Intertemporal
household behavior the introduction of overlapping generations and the modelling of
intergenerational links via bequest motives as well as a closer attention to demographic evolution
provided a much improved economic setting for the understanding of the several aspects of the
problem
Furthermore dynamic modelling has permitted the addressing of a set of new issues Policy
Issues ranging from dynamic tax policy analysis to the evaluation of exogenous changes in
government expenditures to the impact of different methods of financing government expenditure to
4 1
the importance of tinanciai croHrii1g out to 1he relevance of consumorS expectations in tne policy
results have been addressed
Despile all of the progress and in part due to such progress several avenues are wide open for
much neded additional research The following seem to be the most Interesting and promising
areasmiddot First the specincation of liquidity constraints to household behavior may help to obtain a
better understanding of policy chenges that affect directly the interest rates in the economy Second
major attention needs to be paid to the incorporafion of a roraign sector (with open capital markets)
Third some efforts are needed in terms of finding adequate computation techniques now that
dynamic modelling eally makes the curse of dimensionality worse than eVOrbull
The modelling of financial markets is the single most unsatisfactory speet of the dynamic
models surveyed here~ The problems associated with provlding a mo~ adequate treatment of
financial markets and in general endogenous and optimal financial deci$ions bull both at the corporate
and at the personal levelsmiddot have 10 be addressed That necessarily requires Introducing uncertainty
into the models To b adequate this must go well beyond the mere assumption of point expeetalions
with the whole array of modelling implementalion and evaluation problems thai generates
bull I i
42
REFERENCES
Abel A and O Blanchard 1983 An Intertemporal Model of Saving and Interest Econometrica
51 pp 675-92
~ Andersson K 1987 Tar-ation of Capital in Sweden - A General Equilibrium Model Ph D
Dissenation UnIversity of Lund Sweder
Auerbach A L KotlikoH 1983 National Savings Economic Welfare and the Structure of
Taxation in M Feldstein ed Behavioral Simulation Methods in Tax Policy- Analysis Chicago
University of Chicago Press
Auerbach A l Kotlikoff 1984 Social Security and the Economics of Demographic-
Transition in H Aaron and G Burtless eds Retirement and Economic Behavior Washington DC
Brookings Institution
Auerbach A L Kotlikoff 1987 Dynamic Fiscal Policy Cambridge Cambridge University
Press
Auerbach A L Kotlikof and J Skinner 1983 The Efficiency Gains from Dynamic Tax
Reform Internatiooaj Economic Reyiew 24 pp 81middot100
Ballard C 1983 Evaluation of the Consumption Tax with Dynamic General Equilibrium
Models Ph D Dissertation Stanford University
Ballard C 1987 Tax Policy and Consumer Foresight A General Equilibrium Simulation
Study Economic Inquiry 25 pp 267-284
Ballard C and L Goulder 1985 Consumption Taxes ForeSight and Welfare a Computatonal
General Equilibrium Analysis in J Piggott and J Whalley eds New Deyelooments in Apolied
General Equilibrium Analysis Cambridge Cambridge University Press
Ballard C D Fullerton J Shoven and J Whalley 1985 A General Equilibrium Model for
t 43
Tax Eoliey Eyalua~iQO NBER University or Chicago Press
Borges A (1987) bull A Survey of Compula1ional General Equilibnum Models CECD Economic
Slud~ 8
Bovenbig L 19B3 imperfect Capital Mobility in a Perfect Foresight Herberger Model
Mimea UC Br~eley
Bovenberg L 1984 Capitel Accumulation and Capital Immobility Qmiddottheory In a Dynamic General Equilibrium Framework PhD dissertation University of California at Berkeley t
Bovenberg L 19850 Dynamic General Equilibrium Tax Models with Adjustment Costs in A iManne ed bull 1985 EtQnomrc ECUjilbrium Model Formulation and Solution Mathematical
lProgramming Siudy 23 Arnsterdam NorthmiddotHoliand
Bovenberg L t 1986 Capital Income Taxation in Growing Open Eeondmies Journa of Pubtic
Eamp(omlcs 3t pp 347middot376 IBovenberg L ald W Keller 1981 Dynamics and Nonlinearlties in Applied General
Equilibrium Models Internal Report Mlmiddot81middot208 Department for Statistical Methods CBS bull
Vooiburg The Netherlands
Charnley C 1981 The Welfare Cost of C~pltal income Taxation In a Growing Economy
IJournal of Political ECQnQm~ 89 3 pp 468middot496 I
Charnley C 19810 Efficient Tax Reform in a Dynamic Model of General Equilibrium I Mimeo Yale University I Charnley C 1982 On the Optimal Taxation In Stylized Models 01 Inlartemporal General
Equilibrium Mimeo Yale University I Decaluw B and A Manens 1985 Pays en Oeveloppement at Modeles Calculables OEquilibre
General - Une Revue de al litterature empirique Monographie Sene 9 Volume 3 Centre de
Recherche et Developpemont Economlque Unrslt de Montreal
44
bull
I I
I
I
1
Erlich $ V Ginsburgh and L van der Heyden 1987 Wher~ do Real Wage Policies Lead
Belgium bull A General Equilibrium Analysis European Economic Reyiew (forthcoming)
air R and J Taylor 1983 Solution and Maximum Likelihood Estimation of Dynamic
Nonlinear Ratioant Expeclations Models Ecooometrica 51 pp 1169middot1186
Feltenstein A 1984 Money and Bonds In a Disaggregated Ec~nonYin Searl 1-1 J Shoven
Eds Aoolied ~eoeral EqJiliDCiulD tioalYsis Cambridge University Press Cambridge
Feltenstein A 19B6 A Dynamic General Equllibrlu Analysis of Financial Crowding Out
Theory with an Application to Australia Journal of Public Economics 31 No1
Fullerlon D R Gorden 1983 A Reexamination of Tax Distorllons in General Equilibrium
Models in M Feldstein Ed Behayora Simulation Methods in Tax pOlicy Analysis Chicago
University of Chipgo Press
Fvllerlon D J Shaven and J Whalley 1983 Replacing the US Income Tax with a
Progressive Consumption Tax JoulOal of Public ECQnomics 20 pp3middot23
Fullerton D Y Henderson J Shaven 1984 A COmparision of Methodologies in Empirical
General Equilibrium Models of Taxation in Searl H J Shoven Eds ~pQlied Geoaral Jguilibrium
lIoalyss Cambridge University Press Cambridge
Fullerton 0 A King J Shaven and J Whaney 1981 Corporate Tax Integration in the United
Statesmiddot a General Equiiibdufi1 Approach American Economic Review 71 pp~ 677~691
Galper H R Lucke E Tader 1985 Taxation portlor Choice and the Allocation of Capital A
General Equilibrium Approach Mimeo
Gill p W Murray M Saunders M Wright 1986 Users Guide for NPSOL (Version 40) A
Fortran Package for Nonlinear Programming T echnicsl Report SOL 86middot2 Deparlment of
Operations Research Stanford University
Glnsbur~h V and L van der Heyden 1985 General Equilibrium with Wage Rigidities An
45
AppJication to Belgium in A Manne ed t Economic EQumprit1m~ Made Formulation and $oJioo
Mathematical Programming Study 23 Amsterdam NorthmiddotHolland ~
Goulet J 1968 1 Adjustrent CO$1S in the Theory o(the Firm Reyiew of EconomiC Studies 35 bull
pp47middot55
Goulder L 1985 Tax-Financed versus BondmiddotFinancedGhanges in Governent Spending
bull Mimeo Harva(d UniversilYmiddot
Goulder l J Shaven and J Whalley 1983 Domestic Tax Policy and the Foreign Sector The
importance of altemative foreign policy formulations to results from a general equilibrium tax
analysis model In ~n methods in Ibe antico of income from caoia1 ed Martin
Feldstein Chicago University of Chicago riSS
Goulder l L Summers 1987 Tax Policy Asset Prices and 3rowtly ~
A General Equilibrium Analysis NBER Working Pper No 212B
Grandmont J 1982 Temporary General Equilibrium Theory Ch 20 in Handbook 01
Mathemalical Eccnornics Vol II NorthmiddotHoliand
Harberger A 1959 The Corporate Income Tax An empirical Appraisal In lax Revision
Comoendjum Voll House Cornmitee on Ways and Means Washington DC Govemment Printing
Office
Harberger A 1962 The Incidence of the Corporate Income Tax JOULn1 of Political
Economy 70 pp 215middot240
Harberger A 1964 Taxation Resource Aliocalrlon and Welfare fn lb Role of Dir bullbulll and
Indjresl Taxes io tbe Fede[al Reserve s~I~Ill Princeton Princeton Unlveresity Press
James J 1985 New Developments in the Applications of General Equilibrium models to
Economic History in Piggott aand Whalley Eds
Jorgenson D and K Yun 1984 Tax Policy and Capital Allocation Harvard Inslitute of
bull
I
I f
Economic Research WP 1107
Judd K 1965 ShonmiddotRun Analysis of Fiscal Policy in a Simple Perfect Foresight Model
Journal of poJiHcal 5CCJiW1l pp 298middot319 r
Upton D J Potbc J Sachsa nd L Summers 1982 Multiple Shooting in Rational
Expectations Models fuoometrlc 50 1329middot1333
Manne A ed 1985 fconQrnjc Eoullibrium Model FOrIDujaljQo and Solution Mathematical
Pogrammlng Study 23 Amsterdam NorthmiddotHolland
Manne A 1985 Or the Formulation and Solution 01 Economic Equilibrium Models In A
Manne ed ECQOQlli-Jr~ibmiddotum ~tQdel EQlJulailon and Solution Malhematical Programming
Study 23 Amsterdam NorthmiddotHolland
Mansur A and J Whalley 1984 Numerical Specification of Applied General Equilibrium
Models in Scarf H J Shoven Eds Aoplied General Eouillbrium AnaiysectIsect Cambridge Universily
Press Cambridge
Merrill 0 1972 Applications and Extensions of an Algorithm thaI Computes FixedmiddotPoints of
Cenain Upper Semimiddotcominuou Point to Set Mappings PhD Disserlation Unlvel$ity 01 Michigan
Negishi T 1960 Welfar Economics and Exislence I an Equnlbrlum for a Competitive
Economy MelroeCOOQ11ka 12 92middot97
Pereira A 1985 On the Existence and Uniqueness of Optimal Ou1put Path lor CRTS Firms in
Adjustment Costs Technologies Mimeo Stanford University
Pereira A 1985a Consumer Expectations and Ihe COSI 01 UllUty In an Intertemporat
Framework Mlmeo Stanford University
Pereira A 1985b Corporale Tax Integrallon and Government Deficits A Dynamic General
Equilibrium Anaiysls Mlmeo Stanford University
Pereir A 1987 On Ihe Computation of the Users Costs of Stocks Ecooomi (forthcoming)
47
Pereira A bull 1987b Equal Yield Alternatives and Government Oeficlts Mimeof Stanford
University
Pereira A 1987c A Dynamic Applied General ~uHlbrium Model for Tax Policy Evaluation
Comlletamp Docurnent2tion Mimeo University of California San Diego (in progress)
Pereira A 1987d Should the investment Tax Credit be Rei(ltroduced Mlmeo University
of California San Diego (ir progreSs)
Piggott J and J Whalley eds 1985 tJew QeyeQpments in Applied General EQuilibrium
bullenalysjs Cambridge Cambridge University Press
Preckel Pbull 1985 Alternative algorithms for computing economic equilibria In A Manne Ed
Radnor R 19amp3 Equllibrium under Uncertainty In K Arrow M Intrilligaor ods
I ibull
Robinson S 1 g86 Mullisectorai Models of Oeveloping CountrieS a Survey Oivision of
Agricullure and Resources UC Borkeley WP 401
Scarf H 1967 The approximation of fixed points of a continuous mapping ~MM Jouwal of
epplid Mathornallcs 15 pp 1328middot43
Scarf H with T Hansen 1973 00 the Computation of EconomiC Eguilibria New Haven Yale
University Press
Scarf H J Shoven Eds 1984 61l1llied Genera Elujlibrium 6nalyssect Cambridge University
~ Press Cambridge
Shoven J 1976 The Incidence and Efficiency Effects of Taxe on Income from Capital
Jurnal of ramie1 Economy 84 pp 1261-83
Shoven J 1983 Applied General Equilibrium Tax Modemng IMF Slaff PaoerS 30
Shoven J and LB Simon 1986 Share Repurchases and Acquisitions An Analysis of which
Firms Participate NBER Working Paper No 2243
bull
48
IShaven J and J Whalley 1972 ft General Equmbrium Calculation of lhe Effects of
Differential Taxation of Incoma from Capital in the US Jauroe of public Econgmics 1
pp281middot321 (
Shoven J bull and J Whalley 1973 General Equilibrium with Tax$s a computation procedure
and an existence pr00f ReviEhi of Economic Studies 40 pp 475middot490
Shoven J J Whalley 1977 Equal Yield Tax Alternatives A General Equilibrium
Computa~ui0nal Tfchnique J~HJrnal of PUblic EconomIcs 8 pp 211~224 bull
Shaven J J Whalley 1984 Applied GeneralmiddotEquilibrlum MOdels of Taxalion and
International Tr8de An Introduction and Survey Journal of Economic literature 23t pp
1007middot1051
Slemrod J 1980 A General Equilibrium Model of Capital Income Taxation Ph D bullDissertation Harvard Univnrsity
Slemrod J 1983 A General Equilibrium Model of Taxation with Endogenous Financial
Behavior in M Feldstein edbull Behavioral Simulation Methods In Tax Policy Analysis Chicago
UniVerSity of Chicago Press
Stone J 1985 Sequential Optimization and Complementarily Techniques for computing
Economic equilibrium in A Manne Ed
Summers L 1981 Taxation and Corporate Investment a q-Theory Approach BroQkinas
Summers L 1985 Taxation and the Size and ComposHion 01 the Capital Stock An Asset Price
Approach Mlmeo Harvard University
Whalley J 1975 A Geneal Equiloibrium Assessment of the 1973 United Kingdom tax
reform Economica 42 pp 139middot61
1JLTIllOS JIQll~)ltG PAPERS PUBLICADOS bull
I
I I
j
112 70 - LUCENA iJiCIIJo ~ IIEnvironrilent Honitoring and Orqanization Structure T (Setellibro 1987)
nI 71 - L~ENA Diogo A Note on the Representation of Information Stxueture (SeteJObro 1987)
nQ 72 - LUCENA D1CgO Envuorment 111)nitoring and Organization
Structure II (Setelilhro 1987)
ril 73 - AlmmE3 Antonio Pais e CAsPAR Yitor bull Tributacao Inoentivos e InvesUlllentos Analise Qualitativa noYellibro 19S7)
n2 74 - liAlltOSA rose Pedro Opt1lllal lIage R1gidity A SUggestedtlthodology to Teet tb Theory and an Application (Outubro 1987)
nI 75 - CABRAL L 11 B Three lIotes on SYJIIlletric Gemes witt itsYJIIlletric Equilibri (NoveJObro 1987)
n2 76 - LUCENlt D10gOTo Search or Not to Search (Fevereiro1967)
nQ 17 - SA Jorge asconcelos euroI A Theory of SynergyB (Fevereiro1987)
n2 76 - VILARES tanuei Jose Os Be IntenaMios IlIIportados ColIIO Factor de Produao (Julho 1987)
nQ 79 - SA J~rge VasCQncel03 e HoY to Copete and Co~unicate in tature Industrial Products_ (aio 1988)
n2 80 - ROIl Rafael Learning and Capacity Expansion in a New larket Under Uncertainty (Fevereiro 1988)
n2 81 PEREIRA Alfredo llarvilo Survey of Dynamic Computational General EquilibriUlll Hodel for Tax Policy Evaluation (Outubro 1987)
Qualquer in ormacao sabre o 1IIorlltlng Papers ja publlcados sera prestada pelo Secretariado de Apoio aos Docentes podendo os ~e3OSf
ser adquirido5 ne Secyo de Vlndas da Faculdade de EconollLia UNL na Travessa Esteviio Pinto Cltpolide - 1000 LISllOA
32
best lJlodest when measured in terms of the GNP The elimination of the corporate tax and its I replacemenl by increased income lax rates yields long run benefits which are never larger than
2 of the presenl yalue ct telr consumption and leisure onder the previous tax regime and 1
under the current tax regime However the short run w~lare effects of full integration tend 10 be
very low and in in some scenarios even negative This is a flew intertemporar pattern of efficiency Ieffects which reflects an adjJsiment lag in the interindustry investment decisions due to the
existence of costs of adjustment On the of her hand partial integration achieved by excluding I dividends from the corporale lax base syslematically yields negalive efficlencyeffeclS In Ihe long Inm these can be as high as 3 of present value of the future value of consumption and leisure This
is a new second best effect suggesllng Ihat less than complete integration may have perverse I efficiency effects
The dynamic slructure of PerclrBs model snowing for an improved freatment of real investment
declsions and government deficts provides a potential setting for a more accurate measure of the
costs and benefits of inlegration The simulation resullS suggest lower benefits and an intertempora Ibull
pattern of growing benefits
Simulation results also clearly suggest robust negative effects from partial integration How
reliable is this result If dJvidends were deductible from the corporate lax base (as are interest
payments) the preferenlial Ireatmant of debl over equily would be ellminaled Corporallons should
be expecled 10 react by decreasing the optimal deWeqully rallo However corporale financial
gecisions are exogenously set In Pereiras model as in aft of the other models surveyed that go as far
as dealing with the issue
Also withdrawing dividends from Ihe corporate tax base is a way of eliminating Ihe double
taxation of dividends at belh Ihe corporate and personnal Income levels How will the optimal
aliocellon of household saving accomodate that change in Ihe relalive return to corporale equity
39 I
t However hOusehold portfolio decisions are exogenouly set in Pereiras model as in all of the other
models surveyeltj go as far as dealing with the issue
IIt is sate to say that the evaluation 01 the benefits 01 partial Integration as defined above are
sevrey underestima1eO Meanwhile the distortionSoenerated by the increase in the marginal
personal tax rates designed to ralse the revenue foregone by the dividend exclusion are fully
aCCltlunted for A good measure of the costs of integration together With a poor evaluation of the
benefils may very well be the reason why partial integration is Simulated 10 yield neg alive
eHikiency gains
On a different vein as most of Ihe model surveyed here Pereiras is a closed economy model It
is legitimate to wonder how the resullS would change If international capital flows were anowed
IThis almost certainly would affect financial crowding out since a good d~al of the capital inflows
could be used to finance public debt
For these various reasons the results should be treated as preliminary but they strongly I suggosl that the dynamic structure provides valuable new insights into the corporate tax integration SSU2
40
7 Concluding remarks I I I i
What has been acccmplished Ihis far The success of Ihe CGE research in laelding the challenge
of dynamic modelling can not be denied
I iGreat progress in the modelling of household behavior has been achieved Promising
developm~nts in the modeUing of production and government behavior have also been accomplished
Interesting innovations in the concept of equilibrium and computation techniques were discussed
The policy analysis was greatly enriched by considering transitional effects together with longer
run steadymiddotstate equilibrium paths generalized equal yield strategies accomodating different
financial crowding out impacts and generalized dynamic policy evaluation indicators
An important set of issues have been addresed by the models surveyed Traditional issues
focusing on the effects of capital taxation and consumption taxes have been pursued In terms of
capital taxation for example the intertemporal nature of the issue was enhanced by allowing an
optimal evolution of capital stock in the economy and forward looking optimal investment decisions
In Andersson (1987) GouldermiddotSummers (1987) and Pereira (1986b 1987d) this is coupled
with a improved treatment of several tax provisions like investment tax credits and depreciation
allowances In terms of the consumption taxes developments in the specification of Intertemporal
household behavior the introduction of overlapping generations and the modelling of
intergenerational links via bequest motives as well as a closer attention to demographic evolution
provided a much improved economic setting for the understanding of the several aspects of the
problem
Furthermore dynamic modelling has permitted the addressing of a set of new issues Policy
Issues ranging from dynamic tax policy analysis to the evaluation of exogenous changes in
government expenditures to the impact of different methods of financing government expenditure to
4 1
the importance of tinanciai croHrii1g out to 1he relevance of consumorS expectations in tne policy
results have been addressed
Despile all of the progress and in part due to such progress several avenues are wide open for
much neded additional research The following seem to be the most Interesting and promising
areasmiddot First the specincation of liquidity constraints to household behavior may help to obtain a
better understanding of policy chenges that affect directly the interest rates in the economy Second
major attention needs to be paid to the incorporafion of a roraign sector (with open capital markets)
Third some efforts are needed in terms of finding adequate computation techniques now that
dynamic modelling eally makes the curse of dimensionality worse than eVOrbull
The modelling of financial markets is the single most unsatisfactory speet of the dynamic
models surveyed here~ The problems associated with provlding a mo~ adequate treatment of
financial markets and in general endogenous and optimal financial deci$ions bull both at the corporate
and at the personal levelsmiddot have 10 be addressed That necessarily requires Introducing uncertainty
into the models To b adequate this must go well beyond the mere assumption of point expeetalions
with the whole array of modelling implementalion and evaluation problems thai generates
bull I i
42
REFERENCES
Abel A and O Blanchard 1983 An Intertemporal Model of Saving and Interest Econometrica
51 pp 675-92
~ Andersson K 1987 Tar-ation of Capital in Sweden - A General Equilibrium Model Ph D
Dissenation UnIversity of Lund Sweder
Auerbach A L KotlikoH 1983 National Savings Economic Welfare and the Structure of
Taxation in M Feldstein ed Behavioral Simulation Methods in Tax Policy- Analysis Chicago
University of Chicago Press
Auerbach A l Kotlikoff 1984 Social Security and the Economics of Demographic-
Transition in H Aaron and G Burtless eds Retirement and Economic Behavior Washington DC
Brookings Institution
Auerbach A L Kotlikoff 1987 Dynamic Fiscal Policy Cambridge Cambridge University
Press
Auerbach A L Kotlikof and J Skinner 1983 The Efficiency Gains from Dynamic Tax
Reform Internatiooaj Economic Reyiew 24 pp 81middot100
Ballard C 1983 Evaluation of the Consumption Tax with Dynamic General Equilibrium
Models Ph D Dissertation Stanford University
Ballard C 1987 Tax Policy and Consumer Foresight A General Equilibrium Simulation
Study Economic Inquiry 25 pp 267-284
Ballard C and L Goulder 1985 Consumption Taxes ForeSight and Welfare a Computatonal
General Equilibrium Analysis in J Piggott and J Whalley eds New Deyelooments in Apolied
General Equilibrium Analysis Cambridge Cambridge University Press
Ballard C D Fullerton J Shoven and J Whalley 1985 A General Equilibrium Model for
t 43
Tax Eoliey Eyalua~iQO NBER University or Chicago Press
Borges A (1987) bull A Survey of Compula1ional General Equilibnum Models CECD Economic
Slud~ 8
Bovenbig L 19B3 imperfect Capital Mobility in a Perfect Foresight Herberger Model
Mimea UC Br~eley
Bovenberg L 1984 Capitel Accumulation and Capital Immobility Qmiddottheory In a Dynamic General Equilibrium Framework PhD dissertation University of California at Berkeley t
Bovenberg L 19850 Dynamic General Equilibrium Tax Models with Adjustment Costs in A iManne ed bull 1985 EtQnomrc ECUjilbrium Model Formulation and Solution Mathematical
lProgramming Siudy 23 Arnsterdam NorthmiddotHoliand
Bovenberg L t 1986 Capital Income Taxation in Growing Open Eeondmies Journa of Pubtic
Eamp(omlcs 3t pp 347middot376 IBovenberg L ald W Keller 1981 Dynamics and Nonlinearlties in Applied General
Equilibrium Models Internal Report Mlmiddot81middot208 Department for Statistical Methods CBS bull
Vooiburg The Netherlands
Charnley C 1981 The Welfare Cost of C~pltal income Taxation In a Growing Economy
IJournal of Political ECQnQm~ 89 3 pp 468middot496 I
Charnley C 19810 Efficient Tax Reform in a Dynamic Model of General Equilibrium I Mimeo Yale University I Charnley C 1982 On the Optimal Taxation In Stylized Models 01 Inlartemporal General
Equilibrium Mimeo Yale University I Decaluw B and A Manens 1985 Pays en Oeveloppement at Modeles Calculables OEquilibre
General - Une Revue de al litterature empirique Monographie Sene 9 Volume 3 Centre de
Recherche et Developpemont Economlque Unrslt de Montreal
44
bull
I I
I
I
1
Erlich $ V Ginsburgh and L van der Heyden 1987 Wher~ do Real Wage Policies Lead
Belgium bull A General Equilibrium Analysis European Economic Reyiew (forthcoming)
air R and J Taylor 1983 Solution and Maximum Likelihood Estimation of Dynamic
Nonlinear Ratioant Expeclations Models Ecooometrica 51 pp 1169middot1186
Feltenstein A 1984 Money and Bonds In a Disaggregated Ec~nonYin Searl 1-1 J Shoven
Eds Aoolied ~eoeral EqJiliDCiulD tioalYsis Cambridge University Press Cambridge
Feltenstein A 19B6 A Dynamic General Equllibrlu Analysis of Financial Crowding Out
Theory with an Application to Australia Journal of Public Economics 31 No1
Fullerlon D R Gorden 1983 A Reexamination of Tax Distorllons in General Equilibrium
Models in M Feldstein Ed Behayora Simulation Methods in Tax pOlicy Analysis Chicago
University of Chipgo Press
Fvllerlon D J Shaven and J Whalley 1983 Replacing the US Income Tax with a
Progressive Consumption Tax JoulOal of Public ECQnomics 20 pp3middot23
Fullerton D Y Henderson J Shaven 1984 A COmparision of Methodologies in Empirical
General Equilibrium Models of Taxation in Searl H J Shoven Eds ~pQlied Geoaral Jguilibrium
lIoalyss Cambridge University Press Cambridge
Fullerton 0 A King J Shaven and J Whaney 1981 Corporate Tax Integration in the United
Statesmiddot a General Equiiibdufi1 Approach American Economic Review 71 pp~ 677~691
Galper H R Lucke E Tader 1985 Taxation portlor Choice and the Allocation of Capital A
General Equilibrium Approach Mimeo
Gill p W Murray M Saunders M Wright 1986 Users Guide for NPSOL (Version 40) A
Fortran Package for Nonlinear Programming T echnicsl Report SOL 86middot2 Deparlment of
Operations Research Stanford University
Glnsbur~h V and L van der Heyden 1985 General Equilibrium with Wage Rigidities An
45
AppJication to Belgium in A Manne ed t Economic EQumprit1m~ Made Formulation and $oJioo
Mathematical Programming Study 23 Amsterdam NorthmiddotHolland ~
Goulet J 1968 1 Adjustrent CO$1S in the Theory o(the Firm Reyiew of EconomiC Studies 35 bull
pp47middot55
Goulder L 1985 Tax-Financed versus BondmiddotFinancedGhanges in Governent Spending
bull Mimeo Harva(d UniversilYmiddot
Goulder l J Shaven and J Whalley 1983 Domestic Tax Policy and the Foreign Sector The
importance of altemative foreign policy formulations to results from a general equilibrium tax
analysis model In ~n methods in Ibe antico of income from caoia1 ed Martin
Feldstein Chicago University of Chicago riSS
Goulder l L Summers 1987 Tax Policy Asset Prices and 3rowtly ~
A General Equilibrium Analysis NBER Working Pper No 212B
Grandmont J 1982 Temporary General Equilibrium Theory Ch 20 in Handbook 01
Mathemalical Eccnornics Vol II NorthmiddotHoliand
Harberger A 1959 The Corporate Income Tax An empirical Appraisal In lax Revision
Comoendjum Voll House Cornmitee on Ways and Means Washington DC Govemment Printing
Office
Harberger A 1962 The Incidence of the Corporate Income Tax JOULn1 of Political
Economy 70 pp 215middot240
Harberger A 1964 Taxation Resource Aliocalrlon and Welfare fn lb Role of Dir bullbulll and
Indjresl Taxes io tbe Fede[al Reserve s~I~Ill Princeton Princeton Unlveresity Press
James J 1985 New Developments in the Applications of General Equilibrium models to
Economic History in Piggott aand Whalley Eds
Jorgenson D and K Yun 1984 Tax Policy and Capital Allocation Harvard Inslitute of
bull
I
I f
Economic Research WP 1107
Judd K 1965 ShonmiddotRun Analysis of Fiscal Policy in a Simple Perfect Foresight Model
Journal of poJiHcal 5CCJiW1l pp 298middot319 r
Upton D J Potbc J Sachsa nd L Summers 1982 Multiple Shooting in Rational
Expectations Models fuoometrlc 50 1329middot1333
Manne A ed 1985 fconQrnjc Eoullibrium Model FOrIDujaljQo and Solution Mathematical
Pogrammlng Study 23 Amsterdam NorthmiddotHolland
Manne A 1985 Or the Formulation and Solution 01 Economic Equilibrium Models In A
Manne ed ECQOQlli-Jr~ibmiddotum ~tQdel EQlJulailon and Solution Malhematical Programming
Study 23 Amsterdam NorthmiddotHolland
Mansur A and J Whalley 1984 Numerical Specification of Applied General Equilibrium
Models in Scarf H J Shoven Eds Aoplied General Eouillbrium AnaiysectIsect Cambridge Universily
Press Cambridge
Merrill 0 1972 Applications and Extensions of an Algorithm thaI Computes FixedmiddotPoints of
Cenain Upper Semimiddotcominuou Point to Set Mappings PhD Disserlation Unlvel$ity 01 Michigan
Negishi T 1960 Welfar Economics and Exislence I an Equnlbrlum for a Competitive
Economy MelroeCOOQ11ka 12 92middot97
Pereira A 1985 On the Existence and Uniqueness of Optimal Ou1put Path lor CRTS Firms in
Adjustment Costs Technologies Mimeo Stanford University
Pereira A 1985a Consumer Expectations and Ihe COSI 01 UllUty In an Intertemporat
Framework Mlmeo Stanford University
Pereira A 1985b Corporale Tax Integrallon and Government Deficits A Dynamic General
Equilibrium Anaiysls Mlmeo Stanford University
Pereir A 1987 On Ihe Computation of the Users Costs of Stocks Ecooomi (forthcoming)
47
Pereira A bull 1987b Equal Yield Alternatives and Government Oeficlts Mimeof Stanford
University
Pereira A 1987c A Dynamic Applied General ~uHlbrium Model for Tax Policy Evaluation
Comlletamp Docurnent2tion Mimeo University of California San Diego (in progress)
Pereira A 1987d Should the investment Tax Credit be Rei(ltroduced Mlmeo University
of California San Diego (ir progreSs)
Piggott J and J Whalley eds 1985 tJew QeyeQpments in Applied General EQuilibrium
bullenalysjs Cambridge Cambridge University Press
Preckel Pbull 1985 Alternative algorithms for computing economic equilibria In A Manne Ed
Radnor R 19amp3 Equllibrium under Uncertainty In K Arrow M Intrilligaor ods
I ibull
Robinson S 1 g86 Mullisectorai Models of Oeveloping CountrieS a Survey Oivision of
Agricullure and Resources UC Borkeley WP 401
Scarf H 1967 The approximation of fixed points of a continuous mapping ~MM Jouwal of
epplid Mathornallcs 15 pp 1328middot43
Scarf H with T Hansen 1973 00 the Computation of EconomiC Eguilibria New Haven Yale
University Press
Scarf H J Shoven Eds 1984 61l1llied Genera Elujlibrium 6nalyssect Cambridge University
~ Press Cambridge
Shoven J 1976 The Incidence and Efficiency Effects of Taxe on Income from Capital
Jurnal of ramie1 Economy 84 pp 1261-83
Shoven J 1983 Applied General Equilibrium Tax Modemng IMF Slaff PaoerS 30
Shoven J and LB Simon 1986 Share Repurchases and Acquisitions An Analysis of which
Firms Participate NBER Working Paper No 2243
bull
48
IShaven J and J Whalley 1972 ft General Equmbrium Calculation of lhe Effects of
Differential Taxation of Incoma from Capital in the US Jauroe of public Econgmics 1
pp281middot321 (
Shoven J bull and J Whalley 1973 General Equilibrium with Tax$s a computation procedure
and an existence pr00f ReviEhi of Economic Studies 40 pp 475middot490
Shoven J J Whalley 1977 Equal Yield Tax Alternatives A General Equilibrium
Computa~ui0nal Tfchnique J~HJrnal of PUblic EconomIcs 8 pp 211~224 bull
Shaven J J Whalley 1984 Applied GeneralmiddotEquilibrlum MOdels of Taxalion and
International Tr8de An Introduction and Survey Journal of Economic literature 23t pp
1007middot1051
Slemrod J 1980 A General Equilibrium Model of Capital Income Taxation Ph D bullDissertation Harvard Univnrsity
Slemrod J 1983 A General Equilibrium Model of Taxation with Endogenous Financial
Behavior in M Feldstein edbull Behavioral Simulation Methods In Tax Policy Analysis Chicago
UniVerSity of Chicago Press
Stone J 1985 Sequential Optimization and Complementarily Techniques for computing
Economic equilibrium in A Manne Ed
Summers L 1981 Taxation and Corporate Investment a q-Theory Approach BroQkinas
Summers L 1985 Taxation and the Size and ComposHion 01 the Capital Stock An Asset Price
Approach Mlmeo Harvard University
Whalley J 1975 A Geneal Equiloibrium Assessment of the 1973 United Kingdom tax
reform Economica 42 pp 139middot61
1JLTIllOS JIQll~)ltG PAPERS PUBLICADOS bull
I
I I
j
112 70 - LUCENA iJiCIIJo ~ IIEnvironrilent Honitoring and Orqanization Structure T (Setellibro 1987)
nI 71 - L~ENA Diogo A Note on the Representation of Information Stxueture (SeteJObro 1987)
nQ 72 - LUCENA D1CgO Envuorment 111)nitoring and Organization
Structure II (Setelilhro 1987)
ril 73 - AlmmE3 Antonio Pais e CAsPAR Yitor bull Tributacao Inoentivos e InvesUlllentos Analise Qualitativa noYellibro 19S7)
n2 74 - liAlltOSA rose Pedro Opt1lllal lIage R1gidity A SUggestedtlthodology to Teet tb Theory and an Application (Outubro 1987)
nI 75 - CABRAL L 11 B Three lIotes on SYJIIlletric Gemes witt itsYJIIlletric Equilibri (NoveJObro 1987)
n2 76 - LUCENlt D10gOTo Search or Not to Search (Fevereiro1967)
nQ 17 - SA Jorge asconcelos euroI A Theory of SynergyB (Fevereiro1987)
n2 76 - VILARES tanuei Jose Os Be IntenaMios IlIIportados ColIIO Factor de Produao (Julho 1987)
nQ 79 - SA J~rge VasCQncel03 e HoY to Copete and Co~unicate in tature Industrial Products_ (aio 1988)
n2 80 - ROIl Rafael Learning and Capacity Expansion in a New larket Under Uncertainty (Fevereiro 1988)
n2 81 PEREIRA Alfredo llarvilo Survey of Dynamic Computational General EquilibriUlll Hodel for Tax Policy Evaluation (Outubro 1987)
Qualquer in ormacao sabre o 1IIorlltlng Papers ja publlcados sera prestada pelo Secretariado de Apoio aos Docentes podendo os ~e3OSf
ser adquirido5 ne Secyo de Vlndas da Faculdade de EconollLia UNL na Travessa Esteviio Pinto Cltpolide - 1000 LISllOA
39 I
t However hOusehold portfolio decisions are exogenouly set in Pereiras model as in all of the other
models surveyeltj go as far as dealing with the issue
IIt is sate to say that the evaluation 01 the benefits 01 partial Integration as defined above are
sevrey underestima1eO Meanwhile the distortionSoenerated by the increase in the marginal
personal tax rates designed to ralse the revenue foregone by the dividend exclusion are fully
aCCltlunted for A good measure of the costs of integration together With a poor evaluation of the
benefils may very well be the reason why partial integration is Simulated 10 yield neg alive
eHikiency gains
On a different vein as most of Ihe model surveyed here Pereiras is a closed economy model It
is legitimate to wonder how the resullS would change If international capital flows were anowed
IThis almost certainly would affect financial crowding out since a good d~al of the capital inflows
could be used to finance public debt
For these various reasons the results should be treated as preliminary but they strongly I suggosl that the dynamic structure provides valuable new insights into the corporate tax integration SSU2
40
7 Concluding remarks I I I i
What has been acccmplished Ihis far The success of Ihe CGE research in laelding the challenge
of dynamic modelling can not be denied
I iGreat progress in the modelling of household behavior has been achieved Promising
developm~nts in the modeUing of production and government behavior have also been accomplished
Interesting innovations in the concept of equilibrium and computation techniques were discussed
The policy analysis was greatly enriched by considering transitional effects together with longer
run steadymiddotstate equilibrium paths generalized equal yield strategies accomodating different
financial crowding out impacts and generalized dynamic policy evaluation indicators
An important set of issues have been addresed by the models surveyed Traditional issues
focusing on the effects of capital taxation and consumption taxes have been pursued In terms of
capital taxation for example the intertemporal nature of the issue was enhanced by allowing an
optimal evolution of capital stock in the economy and forward looking optimal investment decisions
In Andersson (1987) GouldermiddotSummers (1987) and Pereira (1986b 1987d) this is coupled
with a improved treatment of several tax provisions like investment tax credits and depreciation
allowances In terms of the consumption taxes developments in the specification of Intertemporal
household behavior the introduction of overlapping generations and the modelling of
intergenerational links via bequest motives as well as a closer attention to demographic evolution
provided a much improved economic setting for the understanding of the several aspects of the
problem
Furthermore dynamic modelling has permitted the addressing of a set of new issues Policy
Issues ranging from dynamic tax policy analysis to the evaluation of exogenous changes in
government expenditures to the impact of different methods of financing government expenditure to
4 1
the importance of tinanciai croHrii1g out to 1he relevance of consumorS expectations in tne policy
results have been addressed
Despile all of the progress and in part due to such progress several avenues are wide open for
much neded additional research The following seem to be the most Interesting and promising
areasmiddot First the specincation of liquidity constraints to household behavior may help to obtain a
better understanding of policy chenges that affect directly the interest rates in the economy Second
major attention needs to be paid to the incorporafion of a roraign sector (with open capital markets)
Third some efforts are needed in terms of finding adequate computation techniques now that
dynamic modelling eally makes the curse of dimensionality worse than eVOrbull
The modelling of financial markets is the single most unsatisfactory speet of the dynamic
models surveyed here~ The problems associated with provlding a mo~ adequate treatment of
financial markets and in general endogenous and optimal financial deci$ions bull both at the corporate
and at the personal levelsmiddot have 10 be addressed That necessarily requires Introducing uncertainty
into the models To b adequate this must go well beyond the mere assumption of point expeetalions
with the whole array of modelling implementalion and evaluation problems thai generates
bull I i
42
REFERENCES
Abel A and O Blanchard 1983 An Intertemporal Model of Saving and Interest Econometrica
51 pp 675-92
~ Andersson K 1987 Tar-ation of Capital in Sweden - A General Equilibrium Model Ph D
Dissenation UnIversity of Lund Sweder
Auerbach A L KotlikoH 1983 National Savings Economic Welfare and the Structure of
Taxation in M Feldstein ed Behavioral Simulation Methods in Tax Policy- Analysis Chicago
University of Chicago Press
Auerbach A l Kotlikoff 1984 Social Security and the Economics of Demographic-
Transition in H Aaron and G Burtless eds Retirement and Economic Behavior Washington DC
Brookings Institution
Auerbach A L Kotlikoff 1987 Dynamic Fiscal Policy Cambridge Cambridge University
Press
Auerbach A L Kotlikof and J Skinner 1983 The Efficiency Gains from Dynamic Tax
Reform Internatiooaj Economic Reyiew 24 pp 81middot100
Ballard C 1983 Evaluation of the Consumption Tax with Dynamic General Equilibrium
Models Ph D Dissertation Stanford University
Ballard C 1987 Tax Policy and Consumer Foresight A General Equilibrium Simulation
Study Economic Inquiry 25 pp 267-284
Ballard C and L Goulder 1985 Consumption Taxes ForeSight and Welfare a Computatonal
General Equilibrium Analysis in J Piggott and J Whalley eds New Deyelooments in Apolied
General Equilibrium Analysis Cambridge Cambridge University Press
Ballard C D Fullerton J Shoven and J Whalley 1985 A General Equilibrium Model for
t 43
Tax Eoliey Eyalua~iQO NBER University or Chicago Press
Borges A (1987) bull A Survey of Compula1ional General Equilibnum Models CECD Economic
Slud~ 8
Bovenbig L 19B3 imperfect Capital Mobility in a Perfect Foresight Herberger Model
Mimea UC Br~eley
Bovenberg L 1984 Capitel Accumulation and Capital Immobility Qmiddottheory In a Dynamic General Equilibrium Framework PhD dissertation University of California at Berkeley t
Bovenberg L 19850 Dynamic General Equilibrium Tax Models with Adjustment Costs in A iManne ed bull 1985 EtQnomrc ECUjilbrium Model Formulation and Solution Mathematical
lProgramming Siudy 23 Arnsterdam NorthmiddotHoliand
Bovenberg L t 1986 Capital Income Taxation in Growing Open Eeondmies Journa of Pubtic
Eamp(omlcs 3t pp 347middot376 IBovenberg L ald W Keller 1981 Dynamics and Nonlinearlties in Applied General
Equilibrium Models Internal Report Mlmiddot81middot208 Department for Statistical Methods CBS bull
Vooiburg The Netherlands
Charnley C 1981 The Welfare Cost of C~pltal income Taxation In a Growing Economy
IJournal of Political ECQnQm~ 89 3 pp 468middot496 I
Charnley C 19810 Efficient Tax Reform in a Dynamic Model of General Equilibrium I Mimeo Yale University I Charnley C 1982 On the Optimal Taxation In Stylized Models 01 Inlartemporal General
Equilibrium Mimeo Yale University I Decaluw B and A Manens 1985 Pays en Oeveloppement at Modeles Calculables OEquilibre
General - Une Revue de al litterature empirique Monographie Sene 9 Volume 3 Centre de
Recherche et Developpemont Economlque Unrslt de Montreal
44
bull
I I
I
I
1
Erlich $ V Ginsburgh and L van der Heyden 1987 Wher~ do Real Wage Policies Lead
Belgium bull A General Equilibrium Analysis European Economic Reyiew (forthcoming)
air R and J Taylor 1983 Solution and Maximum Likelihood Estimation of Dynamic
Nonlinear Ratioant Expeclations Models Ecooometrica 51 pp 1169middot1186
Feltenstein A 1984 Money and Bonds In a Disaggregated Ec~nonYin Searl 1-1 J Shoven
Eds Aoolied ~eoeral EqJiliDCiulD tioalYsis Cambridge University Press Cambridge
Feltenstein A 19B6 A Dynamic General Equllibrlu Analysis of Financial Crowding Out
Theory with an Application to Australia Journal of Public Economics 31 No1
Fullerlon D R Gorden 1983 A Reexamination of Tax Distorllons in General Equilibrium
Models in M Feldstein Ed Behayora Simulation Methods in Tax pOlicy Analysis Chicago
University of Chipgo Press
Fvllerlon D J Shaven and J Whalley 1983 Replacing the US Income Tax with a
Progressive Consumption Tax JoulOal of Public ECQnomics 20 pp3middot23
Fullerton D Y Henderson J Shaven 1984 A COmparision of Methodologies in Empirical
General Equilibrium Models of Taxation in Searl H J Shoven Eds ~pQlied Geoaral Jguilibrium
lIoalyss Cambridge University Press Cambridge
Fullerton 0 A King J Shaven and J Whaney 1981 Corporate Tax Integration in the United
Statesmiddot a General Equiiibdufi1 Approach American Economic Review 71 pp~ 677~691
Galper H R Lucke E Tader 1985 Taxation portlor Choice and the Allocation of Capital A
General Equilibrium Approach Mimeo
Gill p W Murray M Saunders M Wright 1986 Users Guide for NPSOL (Version 40) A
Fortran Package for Nonlinear Programming T echnicsl Report SOL 86middot2 Deparlment of
Operations Research Stanford University
Glnsbur~h V and L van der Heyden 1985 General Equilibrium with Wage Rigidities An
45
AppJication to Belgium in A Manne ed t Economic EQumprit1m~ Made Formulation and $oJioo
Mathematical Programming Study 23 Amsterdam NorthmiddotHolland ~
Goulet J 1968 1 Adjustrent CO$1S in the Theory o(the Firm Reyiew of EconomiC Studies 35 bull
pp47middot55
Goulder L 1985 Tax-Financed versus BondmiddotFinancedGhanges in Governent Spending
bull Mimeo Harva(d UniversilYmiddot
Goulder l J Shaven and J Whalley 1983 Domestic Tax Policy and the Foreign Sector The
importance of altemative foreign policy formulations to results from a general equilibrium tax
analysis model In ~n methods in Ibe antico of income from caoia1 ed Martin
Feldstein Chicago University of Chicago riSS
Goulder l L Summers 1987 Tax Policy Asset Prices and 3rowtly ~
A General Equilibrium Analysis NBER Working Pper No 212B
Grandmont J 1982 Temporary General Equilibrium Theory Ch 20 in Handbook 01
Mathemalical Eccnornics Vol II NorthmiddotHoliand
Harberger A 1959 The Corporate Income Tax An empirical Appraisal In lax Revision
Comoendjum Voll House Cornmitee on Ways and Means Washington DC Govemment Printing
Office
Harberger A 1962 The Incidence of the Corporate Income Tax JOULn1 of Political
Economy 70 pp 215middot240
Harberger A 1964 Taxation Resource Aliocalrlon and Welfare fn lb Role of Dir bullbulll and
Indjresl Taxes io tbe Fede[al Reserve s~I~Ill Princeton Princeton Unlveresity Press
James J 1985 New Developments in the Applications of General Equilibrium models to
Economic History in Piggott aand Whalley Eds
Jorgenson D and K Yun 1984 Tax Policy and Capital Allocation Harvard Inslitute of
bull
I
I f
Economic Research WP 1107
Judd K 1965 ShonmiddotRun Analysis of Fiscal Policy in a Simple Perfect Foresight Model
Journal of poJiHcal 5CCJiW1l pp 298middot319 r
Upton D J Potbc J Sachsa nd L Summers 1982 Multiple Shooting in Rational
Expectations Models fuoometrlc 50 1329middot1333
Manne A ed 1985 fconQrnjc Eoullibrium Model FOrIDujaljQo and Solution Mathematical
Pogrammlng Study 23 Amsterdam NorthmiddotHolland
Manne A 1985 Or the Formulation and Solution 01 Economic Equilibrium Models In A
Manne ed ECQOQlli-Jr~ibmiddotum ~tQdel EQlJulailon and Solution Malhematical Programming
Study 23 Amsterdam NorthmiddotHolland
Mansur A and J Whalley 1984 Numerical Specification of Applied General Equilibrium
Models in Scarf H J Shoven Eds Aoplied General Eouillbrium AnaiysectIsect Cambridge Universily
Press Cambridge
Merrill 0 1972 Applications and Extensions of an Algorithm thaI Computes FixedmiddotPoints of
Cenain Upper Semimiddotcominuou Point to Set Mappings PhD Disserlation Unlvel$ity 01 Michigan
Negishi T 1960 Welfar Economics and Exislence I an Equnlbrlum for a Competitive
Economy MelroeCOOQ11ka 12 92middot97
Pereira A 1985 On the Existence and Uniqueness of Optimal Ou1put Path lor CRTS Firms in
Adjustment Costs Technologies Mimeo Stanford University
Pereira A 1985a Consumer Expectations and Ihe COSI 01 UllUty In an Intertemporat
Framework Mlmeo Stanford University
Pereira A 1985b Corporale Tax Integrallon and Government Deficits A Dynamic General
Equilibrium Anaiysls Mlmeo Stanford University
Pereir A 1987 On Ihe Computation of the Users Costs of Stocks Ecooomi (forthcoming)
47
Pereira A bull 1987b Equal Yield Alternatives and Government Oeficlts Mimeof Stanford
University
Pereira A 1987c A Dynamic Applied General ~uHlbrium Model for Tax Policy Evaluation
Comlletamp Docurnent2tion Mimeo University of California San Diego (in progress)
Pereira A 1987d Should the investment Tax Credit be Rei(ltroduced Mlmeo University
of California San Diego (ir progreSs)
Piggott J and J Whalley eds 1985 tJew QeyeQpments in Applied General EQuilibrium
bullenalysjs Cambridge Cambridge University Press
Preckel Pbull 1985 Alternative algorithms for computing economic equilibria In A Manne Ed
Radnor R 19amp3 Equllibrium under Uncertainty In K Arrow M Intrilligaor ods
I ibull
Robinson S 1 g86 Mullisectorai Models of Oeveloping CountrieS a Survey Oivision of
Agricullure and Resources UC Borkeley WP 401
Scarf H 1967 The approximation of fixed points of a continuous mapping ~MM Jouwal of
epplid Mathornallcs 15 pp 1328middot43
Scarf H with T Hansen 1973 00 the Computation of EconomiC Eguilibria New Haven Yale
University Press
Scarf H J Shoven Eds 1984 61l1llied Genera Elujlibrium 6nalyssect Cambridge University
~ Press Cambridge
Shoven J 1976 The Incidence and Efficiency Effects of Taxe on Income from Capital
Jurnal of ramie1 Economy 84 pp 1261-83
Shoven J 1983 Applied General Equilibrium Tax Modemng IMF Slaff PaoerS 30
Shoven J and LB Simon 1986 Share Repurchases and Acquisitions An Analysis of which
Firms Participate NBER Working Paper No 2243
bull
48
IShaven J and J Whalley 1972 ft General Equmbrium Calculation of lhe Effects of
Differential Taxation of Incoma from Capital in the US Jauroe of public Econgmics 1
pp281middot321 (
Shoven J bull and J Whalley 1973 General Equilibrium with Tax$s a computation procedure
and an existence pr00f ReviEhi of Economic Studies 40 pp 475middot490
Shoven J J Whalley 1977 Equal Yield Tax Alternatives A General Equilibrium
Computa~ui0nal Tfchnique J~HJrnal of PUblic EconomIcs 8 pp 211~224 bull
Shaven J J Whalley 1984 Applied GeneralmiddotEquilibrlum MOdels of Taxalion and
International Tr8de An Introduction and Survey Journal of Economic literature 23t pp
1007middot1051
Slemrod J 1980 A General Equilibrium Model of Capital Income Taxation Ph D bullDissertation Harvard Univnrsity
Slemrod J 1983 A General Equilibrium Model of Taxation with Endogenous Financial
Behavior in M Feldstein edbull Behavioral Simulation Methods In Tax Policy Analysis Chicago
UniVerSity of Chicago Press
Stone J 1985 Sequential Optimization and Complementarily Techniques for computing
Economic equilibrium in A Manne Ed
Summers L 1981 Taxation and Corporate Investment a q-Theory Approach BroQkinas
Summers L 1985 Taxation and the Size and ComposHion 01 the Capital Stock An Asset Price
Approach Mlmeo Harvard University
Whalley J 1975 A Geneal Equiloibrium Assessment of the 1973 United Kingdom tax
reform Economica 42 pp 139middot61
1JLTIllOS JIQll~)ltG PAPERS PUBLICADOS bull
I
I I
j
112 70 - LUCENA iJiCIIJo ~ IIEnvironrilent Honitoring and Orqanization Structure T (Setellibro 1987)
nI 71 - L~ENA Diogo A Note on the Representation of Information Stxueture (SeteJObro 1987)
nQ 72 - LUCENA D1CgO Envuorment 111)nitoring and Organization
Structure II (Setelilhro 1987)
ril 73 - AlmmE3 Antonio Pais e CAsPAR Yitor bull Tributacao Inoentivos e InvesUlllentos Analise Qualitativa noYellibro 19S7)
n2 74 - liAlltOSA rose Pedro Opt1lllal lIage R1gidity A SUggestedtlthodology to Teet tb Theory and an Application (Outubro 1987)
nI 75 - CABRAL L 11 B Three lIotes on SYJIIlletric Gemes witt itsYJIIlletric Equilibri (NoveJObro 1987)
n2 76 - LUCENlt D10gOTo Search or Not to Search (Fevereiro1967)
nQ 17 - SA Jorge asconcelos euroI A Theory of SynergyB (Fevereiro1987)
n2 76 - VILARES tanuei Jose Os Be IntenaMios IlIIportados ColIIO Factor de Produao (Julho 1987)
nQ 79 - SA J~rge VasCQncel03 e HoY to Copete and Co~unicate in tature Industrial Products_ (aio 1988)
n2 80 - ROIl Rafael Learning and Capacity Expansion in a New larket Under Uncertainty (Fevereiro 1988)
n2 81 PEREIRA Alfredo llarvilo Survey of Dynamic Computational General EquilibriUlll Hodel for Tax Policy Evaluation (Outubro 1987)
Qualquer in ormacao sabre o 1IIorlltlng Papers ja publlcados sera prestada pelo Secretariado de Apoio aos Docentes podendo os ~e3OSf
ser adquirido5 ne Secyo de Vlndas da Faculdade de EconollLia UNL na Travessa Esteviio Pinto Cltpolide - 1000 LISllOA
40
7 Concluding remarks I I I i
What has been acccmplished Ihis far The success of Ihe CGE research in laelding the challenge
of dynamic modelling can not be denied
I iGreat progress in the modelling of household behavior has been achieved Promising
developm~nts in the modeUing of production and government behavior have also been accomplished
Interesting innovations in the concept of equilibrium and computation techniques were discussed
The policy analysis was greatly enriched by considering transitional effects together with longer
run steadymiddotstate equilibrium paths generalized equal yield strategies accomodating different
financial crowding out impacts and generalized dynamic policy evaluation indicators
An important set of issues have been addresed by the models surveyed Traditional issues
focusing on the effects of capital taxation and consumption taxes have been pursued In terms of
capital taxation for example the intertemporal nature of the issue was enhanced by allowing an
optimal evolution of capital stock in the economy and forward looking optimal investment decisions
In Andersson (1987) GouldermiddotSummers (1987) and Pereira (1986b 1987d) this is coupled
with a improved treatment of several tax provisions like investment tax credits and depreciation
allowances In terms of the consumption taxes developments in the specification of Intertemporal
household behavior the introduction of overlapping generations and the modelling of
intergenerational links via bequest motives as well as a closer attention to demographic evolution
provided a much improved economic setting for the understanding of the several aspects of the
problem
Furthermore dynamic modelling has permitted the addressing of a set of new issues Policy
Issues ranging from dynamic tax policy analysis to the evaluation of exogenous changes in
government expenditures to the impact of different methods of financing government expenditure to
4 1
the importance of tinanciai croHrii1g out to 1he relevance of consumorS expectations in tne policy
results have been addressed
Despile all of the progress and in part due to such progress several avenues are wide open for
much neded additional research The following seem to be the most Interesting and promising
areasmiddot First the specincation of liquidity constraints to household behavior may help to obtain a
better understanding of policy chenges that affect directly the interest rates in the economy Second
major attention needs to be paid to the incorporafion of a roraign sector (with open capital markets)
Third some efforts are needed in terms of finding adequate computation techniques now that
dynamic modelling eally makes the curse of dimensionality worse than eVOrbull
The modelling of financial markets is the single most unsatisfactory speet of the dynamic
models surveyed here~ The problems associated with provlding a mo~ adequate treatment of
financial markets and in general endogenous and optimal financial deci$ions bull both at the corporate
and at the personal levelsmiddot have 10 be addressed That necessarily requires Introducing uncertainty
into the models To b adequate this must go well beyond the mere assumption of point expeetalions
with the whole array of modelling implementalion and evaluation problems thai generates
bull I i
42
REFERENCES
Abel A and O Blanchard 1983 An Intertemporal Model of Saving and Interest Econometrica
51 pp 675-92
~ Andersson K 1987 Tar-ation of Capital in Sweden - A General Equilibrium Model Ph D
Dissenation UnIversity of Lund Sweder
Auerbach A L KotlikoH 1983 National Savings Economic Welfare and the Structure of
Taxation in M Feldstein ed Behavioral Simulation Methods in Tax Policy- Analysis Chicago
University of Chicago Press
Auerbach A l Kotlikoff 1984 Social Security and the Economics of Demographic-
Transition in H Aaron and G Burtless eds Retirement and Economic Behavior Washington DC
Brookings Institution
Auerbach A L Kotlikoff 1987 Dynamic Fiscal Policy Cambridge Cambridge University
Press
Auerbach A L Kotlikof and J Skinner 1983 The Efficiency Gains from Dynamic Tax
Reform Internatiooaj Economic Reyiew 24 pp 81middot100
Ballard C 1983 Evaluation of the Consumption Tax with Dynamic General Equilibrium
Models Ph D Dissertation Stanford University
Ballard C 1987 Tax Policy and Consumer Foresight A General Equilibrium Simulation
Study Economic Inquiry 25 pp 267-284
Ballard C and L Goulder 1985 Consumption Taxes ForeSight and Welfare a Computatonal
General Equilibrium Analysis in J Piggott and J Whalley eds New Deyelooments in Apolied
General Equilibrium Analysis Cambridge Cambridge University Press
Ballard C D Fullerton J Shoven and J Whalley 1985 A General Equilibrium Model for
t 43
Tax Eoliey Eyalua~iQO NBER University or Chicago Press
Borges A (1987) bull A Survey of Compula1ional General Equilibnum Models CECD Economic
Slud~ 8
Bovenbig L 19B3 imperfect Capital Mobility in a Perfect Foresight Herberger Model
Mimea UC Br~eley
Bovenberg L 1984 Capitel Accumulation and Capital Immobility Qmiddottheory In a Dynamic General Equilibrium Framework PhD dissertation University of California at Berkeley t
Bovenberg L 19850 Dynamic General Equilibrium Tax Models with Adjustment Costs in A iManne ed bull 1985 EtQnomrc ECUjilbrium Model Formulation and Solution Mathematical
lProgramming Siudy 23 Arnsterdam NorthmiddotHoliand
Bovenberg L t 1986 Capital Income Taxation in Growing Open Eeondmies Journa of Pubtic
Eamp(omlcs 3t pp 347middot376 IBovenberg L ald W Keller 1981 Dynamics and Nonlinearlties in Applied General
Equilibrium Models Internal Report Mlmiddot81middot208 Department for Statistical Methods CBS bull
Vooiburg The Netherlands
Charnley C 1981 The Welfare Cost of C~pltal income Taxation In a Growing Economy
IJournal of Political ECQnQm~ 89 3 pp 468middot496 I
Charnley C 19810 Efficient Tax Reform in a Dynamic Model of General Equilibrium I Mimeo Yale University I Charnley C 1982 On the Optimal Taxation In Stylized Models 01 Inlartemporal General
Equilibrium Mimeo Yale University I Decaluw B and A Manens 1985 Pays en Oeveloppement at Modeles Calculables OEquilibre
General - Une Revue de al litterature empirique Monographie Sene 9 Volume 3 Centre de
Recherche et Developpemont Economlque Unrslt de Montreal
44
bull
I I
I
I
1
Erlich $ V Ginsburgh and L van der Heyden 1987 Wher~ do Real Wage Policies Lead
Belgium bull A General Equilibrium Analysis European Economic Reyiew (forthcoming)
air R and J Taylor 1983 Solution and Maximum Likelihood Estimation of Dynamic
Nonlinear Ratioant Expeclations Models Ecooometrica 51 pp 1169middot1186
Feltenstein A 1984 Money and Bonds In a Disaggregated Ec~nonYin Searl 1-1 J Shoven
Eds Aoolied ~eoeral EqJiliDCiulD tioalYsis Cambridge University Press Cambridge
Feltenstein A 19B6 A Dynamic General Equllibrlu Analysis of Financial Crowding Out
Theory with an Application to Australia Journal of Public Economics 31 No1
Fullerlon D R Gorden 1983 A Reexamination of Tax Distorllons in General Equilibrium
Models in M Feldstein Ed Behayora Simulation Methods in Tax pOlicy Analysis Chicago
University of Chipgo Press
Fvllerlon D J Shaven and J Whalley 1983 Replacing the US Income Tax with a
Progressive Consumption Tax JoulOal of Public ECQnomics 20 pp3middot23
Fullerton D Y Henderson J Shaven 1984 A COmparision of Methodologies in Empirical
General Equilibrium Models of Taxation in Searl H J Shoven Eds ~pQlied Geoaral Jguilibrium
lIoalyss Cambridge University Press Cambridge
Fullerton 0 A King J Shaven and J Whaney 1981 Corporate Tax Integration in the United
Statesmiddot a General Equiiibdufi1 Approach American Economic Review 71 pp~ 677~691
Galper H R Lucke E Tader 1985 Taxation portlor Choice and the Allocation of Capital A
General Equilibrium Approach Mimeo
Gill p W Murray M Saunders M Wright 1986 Users Guide for NPSOL (Version 40) A
Fortran Package for Nonlinear Programming T echnicsl Report SOL 86middot2 Deparlment of
Operations Research Stanford University
Glnsbur~h V and L van der Heyden 1985 General Equilibrium with Wage Rigidities An
45
AppJication to Belgium in A Manne ed t Economic EQumprit1m~ Made Formulation and $oJioo
Mathematical Programming Study 23 Amsterdam NorthmiddotHolland ~
Goulet J 1968 1 Adjustrent CO$1S in the Theory o(the Firm Reyiew of EconomiC Studies 35 bull
pp47middot55
Goulder L 1985 Tax-Financed versus BondmiddotFinancedGhanges in Governent Spending
bull Mimeo Harva(d UniversilYmiddot
Goulder l J Shaven and J Whalley 1983 Domestic Tax Policy and the Foreign Sector The
importance of altemative foreign policy formulations to results from a general equilibrium tax
analysis model In ~n methods in Ibe antico of income from caoia1 ed Martin
Feldstein Chicago University of Chicago riSS
Goulder l L Summers 1987 Tax Policy Asset Prices and 3rowtly ~
A General Equilibrium Analysis NBER Working Pper No 212B
Grandmont J 1982 Temporary General Equilibrium Theory Ch 20 in Handbook 01
Mathemalical Eccnornics Vol II NorthmiddotHoliand
Harberger A 1959 The Corporate Income Tax An empirical Appraisal In lax Revision
Comoendjum Voll House Cornmitee on Ways and Means Washington DC Govemment Printing
Office
Harberger A 1962 The Incidence of the Corporate Income Tax JOULn1 of Political
Economy 70 pp 215middot240
Harberger A 1964 Taxation Resource Aliocalrlon and Welfare fn lb Role of Dir bullbulll and
Indjresl Taxes io tbe Fede[al Reserve s~I~Ill Princeton Princeton Unlveresity Press
James J 1985 New Developments in the Applications of General Equilibrium models to
Economic History in Piggott aand Whalley Eds
Jorgenson D and K Yun 1984 Tax Policy and Capital Allocation Harvard Inslitute of
bull
I
I f
Economic Research WP 1107
Judd K 1965 ShonmiddotRun Analysis of Fiscal Policy in a Simple Perfect Foresight Model
Journal of poJiHcal 5CCJiW1l pp 298middot319 r
Upton D J Potbc J Sachsa nd L Summers 1982 Multiple Shooting in Rational
Expectations Models fuoometrlc 50 1329middot1333
Manne A ed 1985 fconQrnjc Eoullibrium Model FOrIDujaljQo and Solution Mathematical
Pogrammlng Study 23 Amsterdam NorthmiddotHolland
Manne A 1985 Or the Formulation and Solution 01 Economic Equilibrium Models In A
Manne ed ECQOQlli-Jr~ibmiddotum ~tQdel EQlJulailon and Solution Malhematical Programming
Study 23 Amsterdam NorthmiddotHolland
Mansur A and J Whalley 1984 Numerical Specification of Applied General Equilibrium
Models in Scarf H J Shoven Eds Aoplied General Eouillbrium AnaiysectIsect Cambridge Universily
Press Cambridge
Merrill 0 1972 Applications and Extensions of an Algorithm thaI Computes FixedmiddotPoints of
Cenain Upper Semimiddotcominuou Point to Set Mappings PhD Disserlation Unlvel$ity 01 Michigan
Negishi T 1960 Welfar Economics and Exislence I an Equnlbrlum for a Competitive
Economy MelroeCOOQ11ka 12 92middot97
Pereira A 1985 On the Existence and Uniqueness of Optimal Ou1put Path lor CRTS Firms in
Adjustment Costs Technologies Mimeo Stanford University
Pereira A 1985a Consumer Expectations and Ihe COSI 01 UllUty In an Intertemporat
Framework Mlmeo Stanford University
Pereira A 1985b Corporale Tax Integrallon and Government Deficits A Dynamic General
Equilibrium Anaiysls Mlmeo Stanford University
Pereir A 1987 On Ihe Computation of the Users Costs of Stocks Ecooomi (forthcoming)
47
Pereira A bull 1987b Equal Yield Alternatives and Government Oeficlts Mimeof Stanford
University
Pereira A 1987c A Dynamic Applied General ~uHlbrium Model for Tax Policy Evaluation
Comlletamp Docurnent2tion Mimeo University of California San Diego (in progress)
Pereira A 1987d Should the investment Tax Credit be Rei(ltroduced Mlmeo University
of California San Diego (ir progreSs)
Piggott J and J Whalley eds 1985 tJew QeyeQpments in Applied General EQuilibrium
bullenalysjs Cambridge Cambridge University Press
Preckel Pbull 1985 Alternative algorithms for computing economic equilibria In A Manne Ed
Radnor R 19amp3 Equllibrium under Uncertainty In K Arrow M Intrilligaor ods
I ibull
Robinson S 1 g86 Mullisectorai Models of Oeveloping CountrieS a Survey Oivision of
Agricullure and Resources UC Borkeley WP 401
Scarf H 1967 The approximation of fixed points of a continuous mapping ~MM Jouwal of
epplid Mathornallcs 15 pp 1328middot43
Scarf H with T Hansen 1973 00 the Computation of EconomiC Eguilibria New Haven Yale
University Press
Scarf H J Shoven Eds 1984 61l1llied Genera Elujlibrium 6nalyssect Cambridge University
~ Press Cambridge
Shoven J 1976 The Incidence and Efficiency Effects of Taxe on Income from Capital
Jurnal of ramie1 Economy 84 pp 1261-83
Shoven J 1983 Applied General Equilibrium Tax Modemng IMF Slaff PaoerS 30
Shoven J and LB Simon 1986 Share Repurchases and Acquisitions An Analysis of which
Firms Participate NBER Working Paper No 2243
bull
48
IShaven J and J Whalley 1972 ft General Equmbrium Calculation of lhe Effects of
Differential Taxation of Incoma from Capital in the US Jauroe of public Econgmics 1
pp281middot321 (
Shoven J bull and J Whalley 1973 General Equilibrium with Tax$s a computation procedure
and an existence pr00f ReviEhi of Economic Studies 40 pp 475middot490
Shoven J J Whalley 1977 Equal Yield Tax Alternatives A General Equilibrium
Computa~ui0nal Tfchnique J~HJrnal of PUblic EconomIcs 8 pp 211~224 bull
Shaven J J Whalley 1984 Applied GeneralmiddotEquilibrlum MOdels of Taxalion and
International Tr8de An Introduction and Survey Journal of Economic literature 23t pp
1007middot1051
Slemrod J 1980 A General Equilibrium Model of Capital Income Taxation Ph D bullDissertation Harvard Univnrsity
Slemrod J 1983 A General Equilibrium Model of Taxation with Endogenous Financial
Behavior in M Feldstein edbull Behavioral Simulation Methods In Tax Policy Analysis Chicago
UniVerSity of Chicago Press
Stone J 1985 Sequential Optimization and Complementarily Techniques for computing
Economic equilibrium in A Manne Ed
Summers L 1981 Taxation and Corporate Investment a q-Theory Approach BroQkinas
Summers L 1985 Taxation and the Size and ComposHion 01 the Capital Stock An Asset Price
Approach Mlmeo Harvard University
Whalley J 1975 A Geneal Equiloibrium Assessment of the 1973 United Kingdom tax
reform Economica 42 pp 139middot61
1JLTIllOS JIQll~)ltG PAPERS PUBLICADOS bull
I
I I
j
112 70 - LUCENA iJiCIIJo ~ IIEnvironrilent Honitoring and Orqanization Structure T (Setellibro 1987)
nI 71 - L~ENA Diogo A Note on the Representation of Information Stxueture (SeteJObro 1987)
nQ 72 - LUCENA D1CgO Envuorment 111)nitoring and Organization
Structure II (Setelilhro 1987)
ril 73 - AlmmE3 Antonio Pais e CAsPAR Yitor bull Tributacao Inoentivos e InvesUlllentos Analise Qualitativa noYellibro 19S7)
n2 74 - liAlltOSA rose Pedro Opt1lllal lIage R1gidity A SUggestedtlthodology to Teet tb Theory and an Application (Outubro 1987)
nI 75 - CABRAL L 11 B Three lIotes on SYJIIlletric Gemes witt itsYJIIlletric Equilibri (NoveJObro 1987)
n2 76 - LUCENlt D10gOTo Search or Not to Search (Fevereiro1967)
nQ 17 - SA Jorge asconcelos euroI A Theory of SynergyB (Fevereiro1987)
n2 76 - VILARES tanuei Jose Os Be IntenaMios IlIIportados ColIIO Factor de Produao (Julho 1987)
nQ 79 - SA J~rge VasCQncel03 e HoY to Copete and Co~unicate in tature Industrial Products_ (aio 1988)
n2 80 - ROIl Rafael Learning and Capacity Expansion in a New larket Under Uncertainty (Fevereiro 1988)
n2 81 PEREIRA Alfredo llarvilo Survey of Dynamic Computational General EquilibriUlll Hodel for Tax Policy Evaluation (Outubro 1987)
Qualquer in ormacao sabre o 1IIorlltlng Papers ja publlcados sera prestada pelo Secretariado de Apoio aos Docentes podendo os ~e3OSf
ser adquirido5 ne Secyo de Vlndas da Faculdade de EconollLia UNL na Travessa Esteviio Pinto Cltpolide - 1000 LISllOA
4 1
the importance of tinanciai croHrii1g out to 1he relevance of consumorS expectations in tne policy
results have been addressed
Despile all of the progress and in part due to such progress several avenues are wide open for
much neded additional research The following seem to be the most Interesting and promising
areasmiddot First the specincation of liquidity constraints to household behavior may help to obtain a
better understanding of policy chenges that affect directly the interest rates in the economy Second
major attention needs to be paid to the incorporafion of a roraign sector (with open capital markets)
Third some efforts are needed in terms of finding adequate computation techniques now that
dynamic modelling eally makes the curse of dimensionality worse than eVOrbull
The modelling of financial markets is the single most unsatisfactory speet of the dynamic
models surveyed here~ The problems associated with provlding a mo~ adequate treatment of
financial markets and in general endogenous and optimal financial deci$ions bull both at the corporate
and at the personal levelsmiddot have 10 be addressed That necessarily requires Introducing uncertainty
into the models To b adequate this must go well beyond the mere assumption of point expeetalions
with the whole array of modelling implementalion and evaluation problems thai generates
bull I i
42
REFERENCES
Abel A and O Blanchard 1983 An Intertemporal Model of Saving and Interest Econometrica
51 pp 675-92
~ Andersson K 1987 Tar-ation of Capital in Sweden - A General Equilibrium Model Ph D
Dissenation UnIversity of Lund Sweder
Auerbach A L KotlikoH 1983 National Savings Economic Welfare and the Structure of
Taxation in M Feldstein ed Behavioral Simulation Methods in Tax Policy- Analysis Chicago
University of Chicago Press
Auerbach A l Kotlikoff 1984 Social Security and the Economics of Demographic-
Transition in H Aaron and G Burtless eds Retirement and Economic Behavior Washington DC
Brookings Institution
Auerbach A L Kotlikoff 1987 Dynamic Fiscal Policy Cambridge Cambridge University
Press
Auerbach A L Kotlikof and J Skinner 1983 The Efficiency Gains from Dynamic Tax
Reform Internatiooaj Economic Reyiew 24 pp 81middot100
Ballard C 1983 Evaluation of the Consumption Tax with Dynamic General Equilibrium
Models Ph D Dissertation Stanford University
Ballard C 1987 Tax Policy and Consumer Foresight A General Equilibrium Simulation
Study Economic Inquiry 25 pp 267-284
Ballard C and L Goulder 1985 Consumption Taxes ForeSight and Welfare a Computatonal
General Equilibrium Analysis in J Piggott and J Whalley eds New Deyelooments in Apolied
General Equilibrium Analysis Cambridge Cambridge University Press
Ballard C D Fullerton J Shoven and J Whalley 1985 A General Equilibrium Model for
t 43
Tax Eoliey Eyalua~iQO NBER University or Chicago Press
Borges A (1987) bull A Survey of Compula1ional General Equilibnum Models CECD Economic
Slud~ 8
Bovenbig L 19B3 imperfect Capital Mobility in a Perfect Foresight Herberger Model
Mimea UC Br~eley
Bovenberg L 1984 Capitel Accumulation and Capital Immobility Qmiddottheory In a Dynamic General Equilibrium Framework PhD dissertation University of California at Berkeley t
Bovenberg L 19850 Dynamic General Equilibrium Tax Models with Adjustment Costs in A iManne ed bull 1985 EtQnomrc ECUjilbrium Model Formulation and Solution Mathematical
lProgramming Siudy 23 Arnsterdam NorthmiddotHoliand
Bovenberg L t 1986 Capital Income Taxation in Growing Open Eeondmies Journa of Pubtic
Eamp(omlcs 3t pp 347middot376 IBovenberg L ald W Keller 1981 Dynamics and Nonlinearlties in Applied General
Equilibrium Models Internal Report Mlmiddot81middot208 Department for Statistical Methods CBS bull
Vooiburg The Netherlands
Charnley C 1981 The Welfare Cost of C~pltal income Taxation In a Growing Economy
IJournal of Political ECQnQm~ 89 3 pp 468middot496 I
Charnley C 19810 Efficient Tax Reform in a Dynamic Model of General Equilibrium I Mimeo Yale University I Charnley C 1982 On the Optimal Taxation In Stylized Models 01 Inlartemporal General
Equilibrium Mimeo Yale University I Decaluw B and A Manens 1985 Pays en Oeveloppement at Modeles Calculables OEquilibre
General - Une Revue de al litterature empirique Monographie Sene 9 Volume 3 Centre de
Recherche et Developpemont Economlque Unrslt de Montreal
44
bull
I I
I
I
1
Erlich $ V Ginsburgh and L van der Heyden 1987 Wher~ do Real Wage Policies Lead
Belgium bull A General Equilibrium Analysis European Economic Reyiew (forthcoming)
air R and J Taylor 1983 Solution and Maximum Likelihood Estimation of Dynamic
Nonlinear Ratioant Expeclations Models Ecooometrica 51 pp 1169middot1186
Feltenstein A 1984 Money and Bonds In a Disaggregated Ec~nonYin Searl 1-1 J Shoven
Eds Aoolied ~eoeral EqJiliDCiulD tioalYsis Cambridge University Press Cambridge
Feltenstein A 19B6 A Dynamic General Equllibrlu Analysis of Financial Crowding Out
Theory with an Application to Australia Journal of Public Economics 31 No1
Fullerlon D R Gorden 1983 A Reexamination of Tax Distorllons in General Equilibrium
Models in M Feldstein Ed Behayora Simulation Methods in Tax pOlicy Analysis Chicago
University of Chipgo Press
Fvllerlon D J Shaven and J Whalley 1983 Replacing the US Income Tax with a
Progressive Consumption Tax JoulOal of Public ECQnomics 20 pp3middot23
Fullerton D Y Henderson J Shaven 1984 A COmparision of Methodologies in Empirical
General Equilibrium Models of Taxation in Searl H J Shoven Eds ~pQlied Geoaral Jguilibrium
lIoalyss Cambridge University Press Cambridge
Fullerton 0 A King J Shaven and J Whaney 1981 Corporate Tax Integration in the United
Statesmiddot a General Equiiibdufi1 Approach American Economic Review 71 pp~ 677~691
Galper H R Lucke E Tader 1985 Taxation portlor Choice and the Allocation of Capital A
General Equilibrium Approach Mimeo
Gill p W Murray M Saunders M Wright 1986 Users Guide for NPSOL (Version 40) A
Fortran Package for Nonlinear Programming T echnicsl Report SOL 86middot2 Deparlment of
Operations Research Stanford University
Glnsbur~h V and L van der Heyden 1985 General Equilibrium with Wage Rigidities An
45
AppJication to Belgium in A Manne ed t Economic EQumprit1m~ Made Formulation and $oJioo
Mathematical Programming Study 23 Amsterdam NorthmiddotHolland ~
Goulet J 1968 1 Adjustrent CO$1S in the Theory o(the Firm Reyiew of EconomiC Studies 35 bull
pp47middot55
Goulder L 1985 Tax-Financed versus BondmiddotFinancedGhanges in Governent Spending
bull Mimeo Harva(d UniversilYmiddot
Goulder l J Shaven and J Whalley 1983 Domestic Tax Policy and the Foreign Sector The
importance of altemative foreign policy formulations to results from a general equilibrium tax
analysis model In ~n methods in Ibe antico of income from caoia1 ed Martin
Feldstein Chicago University of Chicago riSS
Goulder l L Summers 1987 Tax Policy Asset Prices and 3rowtly ~
A General Equilibrium Analysis NBER Working Pper No 212B
Grandmont J 1982 Temporary General Equilibrium Theory Ch 20 in Handbook 01
Mathemalical Eccnornics Vol II NorthmiddotHoliand
Harberger A 1959 The Corporate Income Tax An empirical Appraisal In lax Revision
Comoendjum Voll House Cornmitee on Ways and Means Washington DC Govemment Printing
Office
Harberger A 1962 The Incidence of the Corporate Income Tax JOULn1 of Political
Economy 70 pp 215middot240
Harberger A 1964 Taxation Resource Aliocalrlon and Welfare fn lb Role of Dir bullbulll and
Indjresl Taxes io tbe Fede[al Reserve s~I~Ill Princeton Princeton Unlveresity Press
James J 1985 New Developments in the Applications of General Equilibrium models to
Economic History in Piggott aand Whalley Eds
Jorgenson D and K Yun 1984 Tax Policy and Capital Allocation Harvard Inslitute of
bull
I
I f
Economic Research WP 1107
Judd K 1965 ShonmiddotRun Analysis of Fiscal Policy in a Simple Perfect Foresight Model
Journal of poJiHcal 5CCJiW1l pp 298middot319 r
Upton D J Potbc J Sachsa nd L Summers 1982 Multiple Shooting in Rational
Expectations Models fuoometrlc 50 1329middot1333
Manne A ed 1985 fconQrnjc Eoullibrium Model FOrIDujaljQo and Solution Mathematical
Pogrammlng Study 23 Amsterdam NorthmiddotHolland
Manne A 1985 Or the Formulation and Solution 01 Economic Equilibrium Models In A
Manne ed ECQOQlli-Jr~ibmiddotum ~tQdel EQlJulailon and Solution Malhematical Programming
Study 23 Amsterdam NorthmiddotHolland
Mansur A and J Whalley 1984 Numerical Specification of Applied General Equilibrium
Models in Scarf H J Shoven Eds Aoplied General Eouillbrium AnaiysectIsect Cambridge Universily
Press Cambridge
Merrill 0 1972 Applications and Extensions of an Algorithm thaI Computes FixedmiddotPoints of
Cenain Upper Semimiddotcominuou Point to Set Mappings PhD Disserlation Unlvel$ity 01 Michigan
Negishi T 1960 Welfar Economics and Exislence I an Equnlbrlum for a Competitive
Economy MelroeCOOQ11ka 12 92middot97
Pereira A 1985 On the Existence and Uniqueness of Optimal Ou1put Path lor CRTS Firms in
Adjustment Costs Technologies Mimeo Stanford University
Pereira A 1985a Consumer Expectations and Ihe COSI 01 UllUty In an Intertemporat
Framework Mlmeo Stanford University
Pereira A 1985b Corporale Tax Integrallon and Government Deficits A Dynamic General
Equilibrium Anaiysls Mlmeo Stanford University
Pereir A 1987 On Ihe Computation of the Users Costs of Stocks Ecooomi (forthcoming)
47
Pereira A bull 1987b Equal Yield Alternatives and Government Oeficlts Mimeof Stanford
University
Pereira A 1987c A Dynamic Applied General ~uHlbrium Model for Tax Policy Evaluation
Comlletamp Docurnent2tion Mimeo University of California San Diego (in progress)
Pereira A 1987d Should the investment Tax Credit be Rei(ltroduced Mlmeo University
of California San Diego (ir progreSs)
Piggott J and J Whalley eds 1985 tJew QeyeQpments in Applied General EQuilibrium
bullenalysjs Cambridge Cambridge University Press
Preckel Pbull 1985 Alternative algorithms for computing economic equilibria In A Manne Ed
Radnor R 19amp3 Equllibrium under Uncertainty In K Arrow M Intrilligaor ods
I ibull
Robinson S 1 g86 Mullisectorai Models of Oeveloping CountrieS a Survey Oivision of
Agricullure and Resources UC Borkeley WP 401
Scarf H 1967 The approximation of fixed points of a continuous mapping ~MM Jouwal of
epplid Mathornallcs 15 pp 1328middot43
Scarf H with T Hansen 1973 00 the Computation of EconomiC Eguilibria New Haven Yale
University Press
Scarf H J Shoven Eds 1984 61l1llied Genera Elujlibrium 6nalyssect Cambridge University
~ Press Cambridge
Shoven J 1976 The Incidence and Efficiency Effects of Taxe on Income from Capital
Jurnal of ramie1 Economy 84 pp 1261-83
Shoven J 1983 Applied General Equilibrium Tax Modemng IMF Slaff PaoerS 30
Shoven J and LB Simon 1986 Share Repurchases and Acquisitions An Analysis of which
Firms Participate NBER Working Paper No 2243
bull
48
IShaven J and J Whalley 1972 ft General Equmbrium Calculation of lhe Effects of
Differential Taxation of Incoma from Capital in the US Jauroe of public Econgmics 1
pp281middot321 (
Shoven J bull and J Whalley 1973 General Equilibrium with Tax$s a computation procedure
and an existence pr00f ReviEhi of Economic Studies 40 pp 475middot490
Shoven J J Whalley 1977 Equal Yield Tax Alternatives A General Equilibrium
Computa~ui0nal Tfchnique J~HJrnal of PUblic EconomIcs 8 pp 211~224 bull
Shaven J J Whalley 1984 Applied GeneralmiddotEquilibrlum MOdels of Taxalion and
International Tr8de An Introduction and Survey Journal of Economic literature 23t pp
1007middot1051
Slemrod J 1980 A General Equilibrium Model of Capital Income Taxation Ph D bullDissertation Harvard Univnrsity
Slemrod J 1983 A General Equilibrium Model of Taxation with Endogenous Financial
Behavior in M Feldstein edbull Behavioral Simulation Methods In Tax Policy Analysis Chicago
UniVerSity of Chicago Press
Stone J 1985 Sequential Optimization and Complementarily Techniques for computing
Economic equilibrium in A Manne Ed
Summers L 1981 Taxation and Corporate Investment a q-Theory Approach BroQkinas
Summers L 1985 Taxation and the Size and ComposHion 01 the Capital Stock An Asset Price
Approach Mlmeo Harvard University
Whalley J 1975 A Geneal Equiloibrium Assessment of the 1973 United Kingdom tax
reform Economica 42 pp 139middot61
1JLTIllOS JIQll~)ltG PAPERS PUBLICADOS bull
I
I I
j
112 70 - LUCENA iJiCIIJo ~ IIEnvironrilent Honitoring and Orqanization Structure T (Setellibro 1987)
nI 71 - L~ENA Diogo A Note on the Representation of Information Stxueture (SeteJObro 1987)
nQ 72 - LUCENA D1CgO Envuorment 111)nitoring and Organization
Structure II (Setelilhro 1987)
ril 73 - AlmmE3 Antonio Pais e CAsPAR Yitor bull Tributacao Inoentivos e InvesUlllentos Analise Qualitativa noYellibro 19S7)
n2 74 - liAlltOSA rose Pedro Opt1lllal lIage R1gidity A SUggestedtlthodology to Teet tb Theory and an Application (Outubro 1987)
nI 75 - CABRAL L 11 B Three lIotes on SYJIIlletric Gemes witt itsYJIIlletric Equilibri (NoveJObro 1987)
n2 76 - LUCENlt D10gOTo Search or Not to Search (Fevereiro1967)
nQ 17 - SA Jorge asconcelos euroI A Theory of SynergyB (Fevereiro1987)
n2 76 - VILARES tanuei Jose Os Be IntenaMios IlIIportados ColIIO Factor de Produao (Julho 1987)
nQ 79 - SA J~rge VasCQncel03 e HoY to Copete and Co~unicate in tature Industrial Products_ (aio 1988)
n2 80 - ROIl Rafael Learning and Capacity Expansion in a New larket Under Uncertainty (Fevereiro 1988)
n2 81 PEREIRA Alfredo llarvilo Survey of Dynamic Computational General EquilibriUlll Hodel for Tax Policy Evaluation (Outubro 1987)
Qualquer in ormacao sabre o 1IIorlltlng Papers ja publlcados sera prestada pelo Secretariado de Apoio aos Docentes podendo os ~e3OSf
ser adquirido5 ne Secyo de Vlndas da Faculdade de EconollLia UNL na Travessa Esteviio Pinto Cltpolide - 1000 LISllOA
42
REFERENCES
Abel A and O Blanchard 1983 An Intertemporal Model of Saving and Interest Econometrica
51 pp 675-92
~ Andersson K 1987 Tar-ation of Capital in Sweden - A General Equilibrium Model Ph D
Dissenation UnIversity of Lund Sweder
Auerbach A L KotlikoH 1983 National Savings Economic Welfare and the Structure of
Taxation in M Feldstein ed Behavioral Simulation Methods in Tax Policy- Analysis Chicago
University of Chicago Press
Auerbach A l Kotlikoff 1984 Social Security and the Economics of Demographic-
Transition in H Aaron and G Burtless eds Retirement and Economic Behavior Washington DC
Brookings Institution
Auerbach A L Kotlikoff 1987 Dynamic Fiscal Policy Cambridge Cambridge University
Press
Auerbach A L Kotlikof and J Skinner 1983 The Efficiency Gains from Dynamic Tax
Reform Internatiooaj Economic Reyiew 24 pp 81middot100
Ballard C 1983 Evaluation of the Consumption Tax with Dynamic General Equilibrium
Models Ph D Dissertation Stanford University
Ballard C 1987 Tax Policy and Consumer Foresight A General Equilibrium Simulation
Study Economic Inquiry 25 pp 267-284
Ballard C and L Goulder 1985 Consumption Taxes ForeSight and Welfare a Computatonal
General Equilibrium Analysis in J Piggott and J Whalley eds New Deyelooments in Apolied
General Equilibrium Analysis Cambridge Cambridge University Press
Ballard C D Fullerton J Shoven and J Whalley 1985 A General Equilibrium Model for
t 43
Tax Eoliey Eyalua~iQO NBER University or Chicago Press
Borges A (1987) bull A Survey of Compula1ional General Equilibnum Models CECD Economic
Slud~ 8
Bovenbig L 19B3 imperfect Capital Mobility in a Perfect Foresight Herberger Model
Mimea UC Br~eley
Bovenberg L 1984 Capitel Accumulation and Capital Immobility Qmiddottheory In a Dynamic General Equilibrium Framework PhD dissertation University of California at Berkeley t
Bovenberg L 19850 Dynamic General Equilibrium Tax Models with Adjustment Costs in A iManne ed bull 1985 EtQnomrc ECUjilbrium Model Formulation and Solution Mathematical
lProgramming Siudy 23 Arnsterdam NorthmiddotHoliand
Bovenberg L t 1986 Capital Income Taxation in Growing Open Eeondmies Journa of Pubtic
Eamp(omlcs 3t pp 347middot376 IBovenberg L ald W Keller 1981 Dynamics and Nonlinearlties in Applied General
Equilibrium Models Internal Report Mlmiddot81middot208 Department for Statistical Methods CBS bull
Vooiburg The Netherlands
Charnley C 1981 The Welfare Cost of C~pltal income Taxation In a Growing Economy
IJournal of Political ECQnQm~ 89 3 pp 468middot496 I
Charnley C 19810 Efficient Tax Reform in a Dynamic Model of General Equilibrium I Mimeo Yale University I Charnley C 1982 On the Optimal Taxation In Stylized Models 01 Inlartemporal General
Equilibrium Mimeo Yale University I Decaluw B and A Manens 1985 Pays en Oeveloppement at Modeles Calculables OEquilibre
General - Une Revue de al litterature empirique Monographie Sene 9 Volume 3 Centre de
Recherche et Developpemont Economlque Unrslt de Montreal
44
bull
I I
I
I
1
Erlich $ V Ginsburgh and L van der Heyden 1987 Wher~ do Real Wage Policies Lead
Belgium bull A General Equilibrium Analysis European Economic Reyiew (forthcoming)
air R and J Taylor 1983 Solution and Maximum Likelihood Estimation of Dynamic
Nonlinear Ratioant Expeclations Models Ecooometrica 51 pp 1169middot1186
Feltenstein A 1984 Money and Bonds In a Disaggregated Ec~nonYin Searl 1-1 J Shoven
Eds Aoolied ~eoeral EqJiliDCiulD tioalYsis Cambridge University Press Cambridge
Feltenstein A 19B6 A Dynamic General Equllibrlu Analysis of Financial Crowding Out
Theory with an Application to Australia Journal of Public Economics 31 No1
Fullerlon D R Gorden 1983 A Reexamination of Tax Distorllons in General Equilibrium
Models in M Feldstein Ed Behayora Simulation Methods in Tax pOlicy Analysis Chicago
University of Chipgo Press
Fvllerlon D J Shaven and J Whalley 1983 Replacing the US Income Tax with a
Progressive Consumption Tax JoulOal of Public ECQnomics 20 pp3middot23
Fullerton D Y Henderson J Shaven 1984 A COmparision of Methodologies in Empirical
General Equilibrium Models of Taxation in Searl H J Shoven Eds ~pQlied Geoaral Jguilibrium
lIoalyss Cambridge University Press Cambridge
Fullerton 0 A King J Shaven and J Whaney 1981 Corporate Tax Integration in the United
Statesmiddot a General Equiiibdufi1 Approach American Economic Review 71 pp~ 677~691
Galper H R Lucke E Tader 1985 Taxation portlor Choice and the Allocation of Capital A
General Equilibrium Approach Mimeo
Gill p W Murray M Saunders M Wright 1986 Users Guide for NPSOL (Version 40) A
Fortran Package for Nonlinear Programming T echnicsl Report SOL 86middot2 Deparlment of
Operations Research Stanford University
Glnsbur~h V and L van der Heyden 1985 General Equilibrium with Wage Rigidities An
45
AppJication to Belgium in A Manne ed t Economic EQumprit1m~ Made Formulation and $oJioo
Mathematical Programming Study 23 Amsterdam NorthmiddotHolland ~
Goulet J 1968 1 Adjustrent CO$1S in the Theory o(the Firm Reyiew of EconomiC Studies 35 bull
pp47middot55
Goulder L 1985 Tax-Financed versus BondmiddotFinancedGhanges in Governent Spending
bull Mimeo Harva(d UniversilYmiddot
Goulder l J Shaven and J Whalley 1983 Domestic Tax Policy and the Foreign Sector The
importance of altemative foreign policy formulations to results from a general equilibrium tax
analysis model In ~n methods in Ibe antico of income from caoia1 ed Martin
Feldstein Chicago University of Chicago riSS
Goulder l L Summers 1987 Tax Policy Asset Prices and 3rowtly ~
A General Equilibrium Analysis NBER Working Pper No 212B
Grandmont J 1982 Temporary General Equilibrium Theory Ch 20 in Handbook 01
Mathemalical Eccnornics Vol II NorthmiddotHoliand
Harberger A 1959 The Corporate Income Tax An empirical Appraisal In lax Revision
Comoendjum Voll House Cornmitee on Ways and Means Washington DC Govemment Printing
Office
Harberger A 1962 The Incidence of the Corporate Income Tax JOULn1 of Political
Economy 70 pp 215middot240
Harberger A 1964 Taxation Resource Aliocalrlon and Welfare fn lb Role of Dir bullbulll and
Indjresl Taxes io tbe Fede[al Reserve s~I~Ill Princeton Princeton Unlveresity Press
James J 1985 New Developments in the Applications of General Equilibrium models to
Economic History in Piggott aand Whalley Eds
Jorgenson D and K Yun 1984 Tax Policy and Capital Allocation Harvard Inslitute of
bull
I
I f
Economic Research WP 1107
Judd K 1965 ShonmiddotRun Analysis of Fiscal Policy in a Simple Perfect Foresight Model
Journal of poJiHcal 5CCJiW1l pp 298middot319 r
Upton D J Potbc J Sachsa nd L Summers 1982 Multiple Shooting in Rational
Expectations Models fuoometrlc 50 1329middot1333
Manne A ed 1985 fconQrnjc Eoullibrium Model FOrIDujaljQo and Solution Mathematical
Pogrammlng Study 23 Amsterdam NorthmiddotHolland
Manne A 1985 Or the Formulation and Solution 01 Economic Equilibrium Models In A
Manne ed ECQOQlli-Jr~ibmiddotum ~tQdel EQlJulailon and Solution Malhematical Programming
Study 23 Amsterdam NorthmiddotHolland
Mansur A and J Whalley 1984 Numerical Specification of Applied General Equilibrium
Models in Scarf H J Shoven Eds Aoplied General Eouillbrium AnaiysectIsect Cambridge Universily
Press Cambridge
Merrill 0 1972 Applications and Extensions of an Algorithm thaI Computes FixedmiddotPoints of
Cenain Upper Semimiddotcominuou Point to Set Mappings PhD Disserlation Unlvel$ity 01 Michigan
Negishi T 1960 Welfar Economics and Exislence I an Equnlbrlum for a Competitive
Economy MelroeCOOQ11ka 12 92middot97
Pereira A 1985 On the Existence and Uniqueness of Optimal Ou1put Path lor CRTS Firms in
Adjustment Costs Technologies Mimeo Stanford University
Pereira A 1985a Consumer Expectations and Ihe COSI 01 UllUty In an Intertemporat
Framework Mlmeo Stanford University
Pereira A 1985b Corporale Tax Integrallon and Government Deficits A Dynamic General
Equilibrium Anaiysls Mlmeo Stanford University
Pereir A 1987 On Ihe Computation of the Users Costs of Stocks Ecooomi (forthcoming)
47
Pereira A bull 1987b Equal Yield Alternatives and Government Oeficlts Mimeof Stanford
University
Pereira A 1987c A Dynamic Applied General ~uHlbrium Model for Tax Policy Evaluation
Comlletamp Docurnent2tion Mimeo University of California San Diego (in progress)
Pereira A 1987d Should the investment Tax Credit be Rei(ltroduced Mlmeo University
of California San Diego (ir progreSs)
Piggott J and J Whalley eds 1985 tJew QeyeQpments in Applied General EQuilibrium
bullenalysjs Cambridge Cambridge University Press
Preckel Pbull 1985 Alternative algorithms for computing economic equilibria In A Manne Ed
Radnor R 19amp3 Equllibrium under Uncertainty In K Arrow M Intrilligaor ods
I ibull
Robinson S 1 g86 Mullisectorai Models of Oeveloping CountrieS a Survey Oivision of
Agricullure and Resources UC Borkeley WP 401
Scarf H 1967 The approximation of fixed points of a continuous mapping ~MM Jouwal of
epplid Mathornallcs 15 pp 1328middot43
Scarf H with T Hansen 1973 00 the Computation of EconomiC Eguilibria New Haven Yale
University Press
Scarf H J Shoven Eds 1984 61l1llied Genera Elujlibrium 6nalyssect Cambridge University
~ Press Cambridge
Shoven J 1976 The Incidence and Efficiency Effects of Taxe on Income from Capital
Jurnal of ramie1 Economy 84 pp 1261-83
Shoven J 1983 Applied General Equilibrium Tax Modemng IMF Slaff PaoerS 30
Shoven J and LB Simon 1986 Share Repurchases and Acquisitions An Analysis of which
Firms Participate NBER Working Paper No 2243
bull
48
IShaven J and J Whalley 1972 ft General Equmbrium Calculation of lhe Effects of
Differential Taxation of Incoma from Capital in the US Jauroe of public Econgmics 1
pp281middot321 (
Shoven J bull and J Whalley 1973 General Equilibrium with Tax$s a computation procedure
and an existence pr00f ReviEhi of Economic Studies 40 pp 475middot490
Shoven J J Whalley 1977 Equal Yield Tax Alternatives A General Equilibrium
Computa~ui0nal Tfchnique J~HJrnal of PUblic EconomIcs 8 pp 211~224 bull
Shaven J J Whalley 1984 Applied GeneralmiddotEquilibrlum MOdels of Taxalion and
International Tr8de An Introduction and Survey Journal of Economic literature 23t pp
1007middot1051
Slemrod J 1980 A General Equilibrium Model of Capital Income Taxation Ph D bullDissertation Harvard Univnrsity
Slemrod J 1983 A General Equilibrium Model of Taxation with Endogenous Financial
Behavior in M Feldstein edbull Behavioral Simulation Methods In Tax Policy Analysis Chicago
UniVerSity of Chicago Press
Stone J 1985 Sequential Optimization and Complementarily Techniques for computing
Economic equilibrium in A Manne Ed
Summers L 1981 Taxation and Corporate Investment a q-Theory Approach BroQkinas
Summers L 1985 Taxation and the Size and ComposHion 01 the Capital Stock An Asset Price
Approach Mlmeo Harvard University
Whalley J 1975 A Geneal Equiloibrium Assessment of the 1973 United Kingdom tax
reform Economica 42 pp 139middot61
1JLTIllOS JIQll~)ltG PAPERS PUBLICADOS bull
I
I I
j
112 70 - LUCENA iJiCIIJo ~ IIEnvironrilent Honitoring and Orqanization Structure T (Setellibro 1987)
nI 71 - L~ENA Diogo A Note on the Representation of Information Stxueture (SeteJObro 1987)
nQ 72 - LUCENA D1CgO Envuorment 111)nitoring and Organization
Structure II (Setelilhro 1987)
ril 73 - AlmmE3 Antonio Pais e CAsPAR Yitor bull Tributacao Inoentivos e InvesUlllentos Analise Qualitativa noYellibro 19S7)
n2 74 - liAlltOSA rose Pedro Opt1lllal lIage R1gidity A SUggestedtlthodology to Teet tb Theory and an Application (Outubro 1987)
nI 75 - CABRAL L 11 B Three lIotes on SYJIIlletric Gemes witt itsYJIIlletric Equilibri (NoveJObro 1987)
n2 76 - LUCENlt D10gOTo Search or Not to Search (Fevereiro1967)
nQ 17 - SA Jorge asconcelos euroI A Theory of SynergyB (Fevereiro1987)
n2 76 - VILARES tanuei Jose Os Be IntenaMios IlIIportados ColIIO Factor de Produao (Julho 1987)
nQ 79 - SA J~rge VasCQncel03 e HoY to Copete and Co~unicate in tature Industrial Products_ (aio 1988)
n2 80 - ROIl Rafael Learning and Capacity Expansion in a New larket Under Uncertainty (Fevereiro 1988)
n2 81 PEREIRA Alfredo llarvilo Survey of Dynamic Computational General EquilibriUlll Hodel for Tax Policy Evaluation (Outubro 1987)
Qualquer in ormacao sabre o 1IIorlltlng Papers ja publlcados sera prestada pelo Secretariado de Apoio aos Docentes podendo os ~e3OSf
ser adquirido5 ne Secyo de Vlndas da Faculdade de EconollLia UNL na Travessa Esteviio Pinto Cltpolide - 1000 LISllOA
t 43
Tax Eoliey Eyalua~iQO NBER University or Chicago Press
Borges A (1987) bull A Survey of Compula1ional General Equilibnum Models CECD Economic
Slud~ 8
Bovenbig L 19B3 imperfect Capital Mobility in a Perfect Foresight Herberger Model
Mimea UC Br~eley
Bovenberg L 1984 Capitel Accumulation and Capital Immobility Qmiddottheory In a Dynamic General Equilibrium Framework PhD dissertation University of California at Berkeley t
Bovenberg L 19850 Dynamic General Equilibrium Tax Models with Adjustment Costs in A iManne ed bull 1985 EtQnomrc ECUjilbrium Model Formulation and Solution Mathematical
lProgramming Siudy 23 Arnsterdam NorthmiddotHoliand
Bovenberg L t 1986 Capital Income Taxation in Growing Open Eeondmies Journa of Pubtic
Eamp(omlcs 3t pp 347middot376 IBovenberg L ald W Keller 1981 Dynamics and Nonlinearlties in Applied General
Equilibrium Models Internal Report Mlmiddot81middot208 Department for Statistical Methods CBS bull
Vooiburg The Netherlands
Charnley C 1981 The Welfare Cost of C~pltal income Taxation In a Growing Economy
IJournal of Political ECQnQm~ 89 3 pp 468middot496 I
Charnley C 19810 Efficient Tax Reform in a Dynamic Model of General Equilibrium I Mimeo Yale University I Charnley C 1982 On the Optimal Taxation In Stylized Models 01 Inlartemporal General
Equilibrium Mimeo Yale University I Decaluw B and A Manens 1985 Pays en Oeveloppement at Modeles Calculables OEquilibre
General - Une Revue de al litterature empirique Monographie Sene 9 Volume 3 Centre de
Recherche et Developpemont Economlque Unrslt de Montreal
44
bull
I I
I
I
1
Erlich $ V Ginsburgh and L van der Heyden 1987 Wher~ do Real Wage Policies Lead
Belgium bull A General Equilibrium Analysis European Economic Reyiew (forthcoming)
air R and J Taylor 1983 Solution and Maximum Likelihood Estimation of Dynamic
Nonlinear Ratioant Expeclations Models Ecooometrica 51 pp 1169middot1186
Feltenstein A 1984 Money and Bonds In a Disaggregated Ec~nonYin Searl 1-1 J Shoven
Eds Aoolied ~eoeral EqJiliDCiulD tioalYsis Cambridge University Press Cambridge
Feltenstein A 19B6 A Dynamic General Equllibrlu Analysis of Financial Crowding Out
Theory with an Application to Australia Journal of Public Economics 31 No1
Fullerlon D R Gorden 1983 A Reexamination of Tax Distorllons in General Equilibrium
Models in M Feldstein Ed Behayora Simulation Methods in Tax pOlicy Analysis Chicago
University of Chipgo Press
Fvllerlon D J Shaven and J Whalley 1983 Replacing the US Income Tax with a
Progressive Consumption Tax JoulOal of Public ECQnomics 20 pp3middot23
Fullerton D Y Henderson J Shaven 1984 A COmparision of Methodologies in Empirical
General Equilibrium Models of Taxation in Searl H J Shoven Eds ~pQlied Geoaral Jguilibrium
lIoalyss Cambridge University Press Cambridge
Fullerton 0 A King J Shaven and J Whaney 1981 Corporate Tax Integration in the United
Statesmiddot a General Equiiibdufi1 Approach American Economic Review 71 pp~ 677~691
Galper H R Lucke E Tader 1985 Taxation portlor Choice and the Allocation of Capital A
General Equilibrium Approach Mimeo
Gill p W Murray M Saunders M Wright 1986 Users Guide for NPSOL (Version 40) A
Fortran Package for Nonlinear Programming T echnicsl Report SOL 86middot2 Deparlment of
Operations Research Stanford University
Glnsbur~h V and L van der Heyden 1985 General Equilibrium with Wage Rigidities An
45
AppJication to Belgium in A Manne ed t Economic EQumprit1m~ Made Formulation and $oJioo
Mathematical Programming Study 23 Amsterdam NorthmiddotHolland ~
Goulet J 1968 1 Adjustrent CO$1S in the Theory o(the Firm Reyiew of EconomiC Studies 35 bull
pp47middot55
Goulder L 1985 Tax-Financed versus BondmiddotFinancedGhanges in Governent Spending
bull Mimeo Harva(d UniversilYmiddot
Goulder l J Shaven and J Whalley 1983 Domestic Tax Policy and the Foreign Sector The
importance of altemative foreign policy formulations to results from a general equilibrium tax
analysis model In ~n methods in Ibe antico of income from caoia1 ed Martin
Feldstein Chicago University of Chicago riSS
Goulder l L Summers 1987 Tax Policy Asset Prices and 3rowtly ~
A General Equilibrium Analysis NBER Working Pper No 212B
Grandmont J 1982 Temporary General Equilibrium Theory Ch 20 in Handbook 01
Mathemalical Eccnornics Vol II NorthmiddotHoliand
Harberger A 1959 The Corporate Income Tax An empirical Appraisal In lax Revision
Comoendjum Voll House Cornmitee on Ways and Means Washington DC Govemment Printing
Office
Harberger A 1962 The Incidence of the Corporate Income Tax JOULn1 of Political
Economy 70 pp 215middot240
Harberger A 1964 Taxation Resource Aliocalrlon and Welfare fn lb Role of Dir bullbulll and
Indjresl Taxes io tbe Fede[al Reserve s~I~Ill Princeton Princeton Unlveresity Press
James J 1985 New Developments in the Applications of General Equilibrium models to
Economic History in Piggott aand Whalley Eds
Jorgenson D and K Yun 1984 Tax Policy and Capital Allocation Harvard Inslitute of
bull
I
I f
Economic Research WP 1107
Judd K 1965 ShonmiddotRun Analysis of Fiscal Policy in a Simple Perfect Foresight Model
Journal of poJiHcal 5CCJiW1l pp 298middot319 r
Upton D J Potbc J Sachsa nd L Summers 1982 Multiple Shooting in Rational
Expectations Models fuoometrlc 50 1329middot1333
Manne A ed 1985 fconQrnjc Eoullibrium Model FOrIDujaljQo and Solution Mathematical
Pogrammlng Study 23 Amsterdam NorthmiddotHolland
Manne A 1985 Or the Formulation and Solution 01 Economic Equilibrium Models In A
Manne ed ECQOQlli-Jr~ibmiddotum ~tQdel EQlJulailon and Solution Malhematical Programming
Study 23 Amsterdam NorthmiddotHolland
Mansur A and J Whalley 1984 Numerical Specification of Applied General Equilibrium
Models in Scarf H J Shoven Eds Aoplied General Eouillbrium AnaiysectIsect Cambridge Universily
Press Cambridge
Merrill 0 1972 Applications and Extensions of an Algorithm thaI Computes FixedmiddotPoints of
Cenain Upper Semimiddotcominuou Point to Set Mappings PhD Disserlation Unlvel$ity 01 Michigan
Negishi T 1960 Welfar Economics and Exislence I an Equnlbrlum for a Competitive
Economy MelroeCOOQ11ka 12 92middot97
Pereira A 1985 On the Existence and Uniqueness of Optimal Ou1put Path lor CRTS Firms in
Adjustment Costs Technologies Mimeo Stanford University
Pereira A 1985a Consumer Expectations and Ihe COSI 01 UllUty In an Intertemporat
Framework Mlmeo Stanford University
Pereira A 1985b Corporale Tax Integrallon and Government Deficits A Dynamic General
Equilibrium Anaiysls Mlmeo Stanford University
Pereir A 1987 On Ihe Computation of the Users Costs of Stocks Ecooomi (forthcoming)
47
Pereira A bull 1987b Equal Yield Alternatives and Government Oeficlts Mimeof Stanford
University
Pereira A 1987c A Dynamic Applied General ~uHlbrium Model for Tax Policy Evaluation
Comlletamp Docurnent2tion Mimeo University of California San Diego (in progress)
Pereira A 1987d Should the investment Tax Credit be Rei(ltroduced Mlmeo University
of California San Diego (ir progreSs)
Piggott J and J Whalley eds 1985 tJew QeyeQpments in Applied General EQuilibrium
bullenalysjs Cambridge Cambridge University Press
Preckel Pbull 1985 Alternative algorithms for computing economic equilibria In A Manne Ed
Radnor R 19amp3 Equllibrium under Uncertainty In K Arrow M Intrilligaor ods
I ibull
Robinson S 1 g86 Mullisectorai Models of Oeveloping CountrieS a Survey Oivision of
Agricullure and Resources UC Borkeley WP 401
Scarf H 1967 The approximation of fixed points of a continuous mapping ~MM Jouwal of
epplid Mathornallcs 15 pp 1328middot43
Scarf H with T Hansen 1973 00 the Computation of EconomiC Eguilibria New Haven Yale
University Press
Scarf H J Shoven Eds 1984 61l1llied Genera Elujlibrium 6nalyssect Cambridge University
~ Press Cambridge
Shoven J 1976 The Incidence and Efficiency Effects of Taxe on Income from Capital
Jurnal of ramie1 Economy 84 pp 1261-83
Shoven J 1983 Applied General Equilibrium Tax Modemng IMF Slaff PaoerS 30
Shoven J and LB Simon 1986 Share Repurchases and Acquisitions An Analysis of which
Firms Participate NBER Working Paper No 2243
bull
48
IShaven J and J Whalley 1972 ft General Equmbrium Calculation of lhe Effects of
Differential Taxation of Incoma from Capital in the US Jauroe of public Econgmics 1
pp281middot321 (
Shoven J bull and J Whalley 1973 General Equilibrium with Tax$s a computation procedure
and an existence pr00f ReviEhi of Economic Studies 40 pp 475middot490
Shoven J J Whalley 1977 Equal Yield Tax Alternatives A General Equilibrium
Computa~ui0nal Tfchnique J~HJrnal of PUblic EconomIcs 8 pp 211~224 bull
Shaven J J Whalley 1984 Applied GeneralmiddotEquilibrlum MOdels of Taxalion and
International Tr8de An Introduction and Survey Journal of Economic literature 23t pp
1007middot1051
Slemrod J 1980 A General Equilibrium Model of Capital Income Taxation Ph D bullDissertation Harvard Univnrsity
Slemrod J 1983 A General Equilibrium Model of Taxation with Endogenous Financial
Behavior in M Feldstein edbull Behavioral Simulation Methods In Tax Policy Analysis Chicago
UniVerSity of Chicago Press
Stone J 1985 Sequential Optimization and Complementarily Techniques for computing
Economic equilibrium in A Manne Ed
Summers L 1981 Taxation and Corporate Investment a q-Theory Approach BroQkinas
Summers L 1985 Taxation and the Size and ComposHion 01 the Capital Stock An Asset Price
Approach Mlmeo Harvard University
Whalley J 1975 A Geneal Equiloibrium Assessment of the 1973 United Kingdom tax
reform Economica 42 pp 139middot61
1JLTIllOS JIQll~)ltG PAPERS PUBLICADOS bull
I
I I
j
112 70 - LUCENA iJiCIIJo ~ IIEnvironrilent Honitoring and Orqanization Structure T (Setellibro 1987)
nI 71 - L~ENA Diogo A Note on the Representation of Information Stxueture (SeteJObro 1987)
nQ 72 - LUCENA D1CgO Envuorment 111)nitoring and Organization
Structure II (Setelilhro 1987)
ril 73 - AlmmE3 Antonio Pais e CAsPAR Yitor bull Tributacao Inoentivos e InvesUlllentos Analise Qualitativa noYellibro 19S7)
n2 74 - liAlltOSA rose Pedro Opt1lllal lIage R1gidity A SUggestedtlthodology to Teet tb Theory and an Application (Outubro 1987)
nI 75 - CABRAL L 11 B Three lIotes on SYJIIlletric Gemes witt itsYJIIlletric Equilibri (NoveJObro 1987)
n2 76 - LUCENlt D10gOTo Search or Not to Search (Fevereiro1967)
nQ 17 - SA Jorge asconcelos euroI A Theory of SynergyB (Fevereiro1987)
n2 76 - VILARES tanuei Jose Os Be IntenaMios IlIIportados ColIIO Factor de Produao (Julho 1987)
nQ 79 - SA J~rge VasCQncel03 e HoY to Copete and Co~unicate in tature Industrial Products_ (aio 1988)
n2 80 - ROIl Rafael Learning and Capacity Expansion in a New larket Under Uncertainty (Fevereiro 1988)
n2 81 PEREIRA Alfredo llarvilo Survey of Dynamic Computational General EquilibriUlll Hodel for Tax Policy Evaluation (Outubro 1987)
Qualquer in ormacao sabre o 1IIorlltlng Papers ja publlcados sera prestada pelo Secretariado de Apoio aos Docentes podendo os ~e3OSf
ser adquirido5 ne Secyo de Vlndas da Faculdade de EconollLia UNL na Travessa Esteviio Pinto Cltpolide - 1000 LISllOA
44
bull
I I
I
I
1
Erlich $ V Ginsburgh and L van der Heyden 1987 Wher~ do Real Wage Policies Lead
Belgium bull A General Equilibrium Analysis European Economic Reyiew (forthcoming)
air R and J Taylor 1983 Solution and Maximum Likelihood Estimation of Dynamic
Nonlinear Ratioant Expeclations Models Ecooometrica 51 pp 1169middot1186
Feltenstein A 1984 Money and Bonds In a Disaggregated Ec~nonYin Searl 1-1 J Shoven
Eds Aoolied ~eoeral EqJiliDCiulD tioalYsis Cambridge University Press Cambridge
Feltenstein A 19B6 A Dynamic General Equllibrlu Analysis of Financial Crowding Out
Theory with an Application to Australia Journal of Public Economics 31 No1
Fullerlon D R Gorden 1983 A Reexamination of Tax Distorllons in General Equilibrium
Models in M Feldstein Ed Behayora Simulation Methods in Tax pOlicy Analysis Chicago
University of Chipgo Press
Fvllerlon D J Shaven and J Whalley 1983 Replacing the US Income Tax with a
Progressive Consumption Tax JoulOal of Public ECQnomics 20 pp3middot23
Fullerton D Y Henderson J Shaven 1984 A COmparision of Methodologies in Empirical
General Equilibrium Models of Taxation in Searl H J Shoven Eds ~pQlied Geoaral Jguilibrium
lIoalyss Cambridge University Press Cambridge
Fullerton 0 A King J Shaven and J Whaney 1981 Corporate Tax Integration in the United
Statesmiddot a General Equiiibdufi1 Approach American Economic Review 71 pp~ 677~691
Galper H R Lucke E Tader 1985 Taxation portlor Choice and the Allocation of Capital A
General Equilibrium Approach Mimeo
Gill p W Murray M Saunders M Wright 1986 Users Guide for NPSOL (Version 40) A
Fortran Package for Nonlinear Programming T echnicsl Report SOL 86middot2 Deparlment of
Operations Research Stanford University
Glnsbur~h V and L van der Heyden 1985 General Equilibrium with Wage Rigidities An
45
AppJication to Belgium in A Manne ed t Economic EQumprit1m~ Made Formulation and $oJioo
Mathematical Programming Study 23 Amsterdam NorthmiddotHolland ~
Goulet J 1968 1 Adjustrent CO$1S in the Theory o(the Firm Reyiew of EconomiC Studies 35 bull
pp47middot55
Goulder L 1985 Tax-Financed versus BondmiddotFinancedGhanges in Governent Spending
bull Mimeo Harva(d UniversilYmiddot
Goulder l J Shaven and J Whalley 1983 Domestic Tax Policy and the Foreign Sector The
importance of altemative foreign policy formulations to results from a general equilibrium tax
analysis model In ~n methods in Ibe antico of income from caoia1 ed Martin
Feldstein Chicago University of Chicago riSS
Goulder l L Summers 1987 Tax Policy Asset Prices and 3rowtly ~
A General Equilibrium Analysis NBER Working Pper No 212B
Grandmont J 1982 Temporary General Equilibrium Theory Ch 20 in Handbook 01
Mathemalical Eccnornics Vol II NorthmiddotHoliand
Harberger A 1959 The Corporate Income Tax An empirical Appraisal In lax Revision
Comoendjum Voll House Cornmitee on Ways and Means Washington DC Govemment Printing
Office
Harberger A 1962 The Incidence of the Corporate Income Tax JOULn1 of Political
Economy 70 pp 215middot240
Harberger A 1964 Taxation Resource Aliocalrlon and Welfare fn lb Role of Dir bullbulll and
Indjresl Taxes io tbe Fede[al Reserve s~I~Ill Princeton Princeton Unlveresity Press
James J 1985 New Developments in the Applications of General Equilibrium models to
Economic History in Piggott aand Whalley Eds
Jorgenson D and K Yun 1984 Tax Policy and Capital Allocation Harvard Inslitute of
bull
I
I f
Economic Research WP 1107
Judd K 1965 ShonmiddotRun Analysis of Fiscal Policy in a Simple Perfect Foresight Model
Journal of poJiHcal 5CCJiW1l pp 298middot319 r
Upton D J Potbc J Sachsa nd L Summers 1982 Multiple Shooting in Rational
Expectations Models fuoometrlc 50 1329middot1333
Manne A ed 1985 fconQrnjc Eoullibrium Model FOrIDujaljQo and Solution Mathematical
Pogrammlng Study 23 Amsterdam NorthmiddotHolland
Manne A 1985 Or the Formulation and Solution 01 Economic Equilibrium Models In A
Manne ed ECQOQlli-Jr~ibmiddotum ~tQdel EQlJulailon and Solution Malhematical Programming
Study 23 Amsterdam NorthmiddotHolland
Mansur A and J Whalley 1984 Numerical Specification of Applied General Equilibrium
Models in Scarf H J Shoven Eds Aoplied General Eouillbrium AnaiysectIsect Cambridge Universily
Press Cambridge
Merrill 0 1972 Applications and Extensions of an Algorithm thaI Computes FixedmiddotPoints of
Cenain Upper Semimiddotcominuou Point to Set Mappings PhD Disserlation Unlvel$ity 01 Michigan
Negishi T 1960 Welfar Economics and Exislence I an Equnlbrlum for a Competitive
Economy MelroeCOOQ11ka 12 92middot97
Pereira A 1985 On the Existence and Uniqueness of Optimal Ou1put Path lor CRTS Firms in
Adjustment Costs Technologies Mimeo Stanford University
Pereira A 1985a Consumer Expectations and Ihe COSI 01 UllUty In an Intertemporat
Framework Mlmeo Stanford University
Pereira A 1985b Corporale Tax Integrallon and Government Deficits A Dynamic General
Equilibrium Anaiysls Mlmeo Stanford University
Pereir A 1987 On Ihe Computation of the Users Costs of Stocks Ecooomi (forthcoming)
47
Pereira A bull 1987b Equal Yield Alternatives and Government Oeficlts Mimeof Stanford
University
Pereira A 1987c A Dynamic Applied General ~uHlbrium Model for Tax Policy Evaluation
Comlletamp Docurnent2tion Mimeo University of California San Diego (in progress)
Pereira A 1987d Should the investment Tax Credit be Rei(ltroduced Mlmeo University
of California San Diego (ir progreSs)
Piggott J and J Whalley eds 1985 tJew QeyeQpments in Applied General EQuilibrium
bullenalysjs Cambridge Cambridge University Press
Preckel Pbull 1985 Alternative algorithms for computing economic equilibria In A Manne Ed
Radnor R 19amp3 Equllibrium under Uncertainty In K Arrow M Intrilligaor ods
I ibull
Robinson S 1 g86 Mullisectorai Models of Oeveloping CountrieS a Survey Oivision of
Agricullure and Resources UC Borkeley WP 401
Scarf H 1967 The approximation of fixed points of a continuous mapping ~MM Jouwal of
epplid Mathornallcs 15 pp 1328middot43
Scarf H with T Hansen 1973 00 the Computation of EconomiC Eguilibria New Haven Yale
University Press
Scarf H J Shoven Eds 1984 61l1llied Genera Elujlibrium 6nalyssect Cambridge University
~ Press Cambridge
Shoven J 1976 The Incidence and Efficiency Effects of Taxe on Income from Capital
Jurnal of ramie1 Economy 84 pp 1261-83
Shoven J 1983 Applied General Equilibrium Tax Modemng IMF Slaff PaoerS 30
Shoven J and LB Simon 1986 Share Repurchases and Acquisitions An Analysis of which
Firms Participate NBER Working Paper No 2243
bull
48
IShaven J and J Whalley 1972 ft General Equmbrium Calculation of lhe Effects of
Differential Taxation of Incoma from Capital in the US Jauroe of public Econgmics 1
pp281middot321 (
Shoven J bull and J Whalley 1973 General Equilibrium with Tax$s a computation procedure
and an existence pr00f ReviEhi of Economic Studies 40 pp 475middot490
Shoven J J Whalley 1977 Equal Yield Tax Alternatives A General Equilibrium
Computa~ui0nal Tfchnique J~HJrnal of PUblic EconomIcs 8 pp 211~224 bull
Shaven J J Whalley 1984 Applied GeneralmiddotEquilibrlum MOdels of Taxalion and
International Tr8de An Introduction and Survey Journal of Economic literature 23t pp
1007middot1051
Slemrod J 1980 A General Equilibrium Model of Capital Income Taxation Ph D bullDissertation Harvard Univnrsity
Slemrod J 1983 A General Equilibrium Model of Taxation with Endogenous Financial
Behavior in M Feldstein edbull Behavioral Simulation Methods In Tax Policy Analysis Chicago
UniVerSity of Chicago Press
Stone J 1985 Sequential Optimization and Complementarily Techniques for computing
Economic equilibrium in A Manne Ed
Summers L 1981 Taxation and Corporate Investment a q-Theory Approach BroQkinas
Summers L 1985 Taxation and the Size and ComposHion 01 the Capital Stock An Asset Price
Approach Mlmeo Harvard University
Whalley J 1975 A Geneal Equiloibrium Assessment of the 1973 United Kingdom tax
reform Economica 42 pp 139middot61
1JLTIllOS JIQll~)ltG PAPERS PUBLICADOS bull
I
I I
j
112 70 - LUCENA iJiCIIJo ~ IIEnvironrilent Honitoring and Orqanization Structure T (Setellibro 1987)
nI 71 - L~ENA Diogo A Note on the Representation of Information Stxueture (SeteJObro 1987)
nQ 72 - LUCENA D1CgO Envuorment 111)nitoring and Organization
Structure II (Setelilhro 1987)
ril 73 - AlmmE3 Antonio Pais e CAsPAR Yitor bull Tributacao Inoentivos e InvesUlllentos Analise Qualitativa noYellibro 19S7)
n2 74 - liAlltOSA rose Pedro Opt1lllal lIage R1gidity A SUggestedtlthodology to Teet tb Theory and an Application (Outubro 1987)
nI 75 - CABRAL L 11 B Three lIotes on SYJIIlletric Gemes witt itsYJIIlletric Equilibri (NoveJObro 1987)
n2 76 - LUCENlt D10gOTo Search or Not to Search (Fevereiro1967)
nQ 17 - SA Jorge asconcelos euroI A Theory of SynergyB (Fevereiro1987)
n2 76 - VILARES tanuei Jose Os Be IntenaMios IlIIportados ColIIO Factor de Produao (Julho 1987)
nQ 79 - SA J~rge VasCQncel03 e HoY to Copete and Co~unicate in tature Industrial Products_ (aio 1988)
n2 80 - ROIl Rafael Learning and Capacity Expansion in a New larket Under Uncertainty (Fevereiro 1988)
n2 81 PEREIRA Alfredo llarvilo Survey of Dynamic Computational General EquilibriUlll Hodel for Tax Policy Evaluation (Outubro 1987)
Qualquer in ormacao sabre o 1IIorlltlng Papers ja publlcados sera prestada pelo Secretariado de Apoio aos Docentes podendo os ~e3OSf
ser adquirido5 ne Secyo de Vlndas da Faculdade de EconollLia UNL na Travessa Esteviio Pinto Cltpolide - 1000 LISllOA
45
AppJication to Belgium in A Manne ed t Economic EQumprit1m~ Made Formulation and $oJioo
Mathematical Programming Study 23 Amsterdam NorthmiddotHolland ~
Goulet J 1968 1 Adjustrent CO$1S in the Theory o(the Firm Reyiew of EconomiC Studies 35 bull
pp47middot55
Goulder L 1985 Tax-Financed versus BondmiddotFinancedGhanges in Governent Spending
bull Mimeo Harva(d UniversilYmiddot
Goulder l J Shaven and J Whalley 1983 Domestic Tax Policy and the Foreign Sector The
importance of altemative foreign policy formulations to results from a general equilibrium tax
analysis model In ~n methods in Ibe antico of income from caoia1 ed Martin
Feldstein Chicago University of Chicago riSS
Goulder l L Summers 1987 Tax Policy Asset Prices and 3rowtly ~
A General Equilibrium Analysis NBER Working Pper No 212B
Grandmont J 1982 Temporary General Equilibrium Theory Ch 20 in Handbook 01
Mathemalical Eccnornics Vol II NorthmiddotHoliand
Harberger A 1959 The Corporate Income Tax An empirical Appraisal In lax Revision
Comoendjum Voll House Cornmitee on Ways and Means Washington DC Govemment Printing
Office
Harberger A 1962 The Incidence of the Corporate Income Tax JOULn1 of Political
Economy 70 pp 215middot240
Harberger A 1964 Taxation Resource Aliocalrlon and Welfare fn lb Role of Dir bullbulll and
Indjresl Taxes io tbe Fede[al Reserve s~I~Ill Princeton Princeton Unlveresity Press
James J 1985 New Developments in the Applications of General Equilibrium models to
Economic History in Piggott aand Whalley Eds
Jorgenson D and K Yun 1984 Tax Policy and Capital Allocation Harvard Inslitute of
bull
I
I f
Economic Research WP 1107
Judd K 1965 ShonmiddotRun Analysis of Fiscal Policy in a Simple Perfect Foresight Model
Journal of poJiHcal 5CCJiW1l pp 298middot319 r
Upton D J Potbc J Sachsa nd L Summers 1982 Multiple Shooting in Rational
Expectations Models fuoometrlc 50 1329middot1333
Manne A ed 1985 fconQrnjc Eoullibrium Model FOrIDujaljQo and Solution Mathematical
Pogrammlng Study 23 Amsterdam NorthmiddotHolland
Manne A 1985 Or the Formulation and Solution 01 Economic Equilibrium Models In A
Manne ed ECQOQlli-Jr~ibmiddotum ~tQdel EQlJulailon and Solution Malhematical Programming
Study 23 Amsterdam NorthmiddotHolland
Mansur A and J Whalley 1984 Numerical Specification of Applied General Equilibrium
Models in Scarf H J Shoven Eds Aoplied General Eouillbrium AnaiysectIsect Cambridge Universily
Press Cambridge
Merrill 0 1972 Applications and Extensions of an Algorithm thaI Computes FixedmiddotPoints of
Cenain Upper Semimiddotcominuou Point to Set Mappings PhD Disserlation Unlvel$ity 01 Michigan
Negishi T 1960 Welfar Economics and Exislence I an Equnlbrlum for a Competitive
Economy MelroeCOOQ11ka 12 92middot97
Pereira A 1985 On the Existence and Uniqueness of Optimal Ou1put Path lor CRTS Firms in
Adjustment Costs Technologies Mimeo Stanford University
Pereira A 1985a Consumer Expectations and Ihe COSI 01 UllUty In an Intertemporat
Framework Mlmeo Stanford University
Pereira A 1985b Corporale Tax Integrallon and Government Deficits A Dynamic General
Equilibrium Anaiysls Mlmeo Stanford University
Pereir A 1987 On Ihe Computation of the Users Costs of Stocks Ecooomi (forthcoming)
47
Pereira A bull 1987b Equal Yield Alternatives and Government Oeficlts Mimeof Stanford
University
Pereira A 1987c A Dynamic Applied General ~uHlbrium Model for Tax Policy Evaluation
Comlletamp Docurnent2tion Mimeo University of California San Diego (in progress)
Pereira A 1987d Should the investment Tax Credit be Rei(ltroduced Mlmeo University
of California San Diego (ir progreSs)
Piggott J and J Whalley eds 1985 tJew QeyeQpments in Applied General EQuilibrium
bullenalysjs Cambridge Cambridge University Press
Preckel Pbull 1985 Alternative algorithms for computing economic equilibria In A Manne Ed
Radnor R 19amp3 Equllibrium under Uncertainty In K Arrow M Intrilligaor ods
I ibull
Robinson S 1 g86 Mullisectorai Models of Oeveloping CountrieS a Survey Oivision of
Agricullure and Resources UC Borkeley WP 401
Scarf H 1967 The approximation of fixed points of a continuous mapping ~MM Jouwal of
epplid Mathornallcs 15 pp 1328middot43
Scarf H with T Hansen 1973 00 the Computation of EconomiC Eguilibria New Haven Yale
University Press
Scarf H J Shoven Eds 1984 61l1llied Genera Elujlibrium 6nalyssect Cambridge University
~ Press Cambridge
Shoven J 1976 The Incidence and Efficiency Effects of Taxe on Income from Capital
Jurnal of ramie1 Economy 84 pp 1261-83
Shoven J 1983 Applied General Equilibrium Tax Modemng IMF Slaff PaoerS 30
Shoven J and LB Simon 1986 Share Repurchases and Acquisitions An Analysis of which
Firms Participate NBER Working Paper No 2243
bull
48
IShaven J and J Whalley 1972 ft General Equmbrium Calculation of lhe Effects of
Differential Taxation of Incoma from Capital in the US Jauroe of public Econgmics 1
pp281middot321 (
Shoven J bull and J Whalley 1973 General Equilibrium with Tax$s a computation procedure
and an existence pr00f ReviEhi of Economic Studies 40 pp 475middot490
Shoven J J Whalley 1977 Equal Yield Tax Alternatives A General Equilibrium
Computa~ui0nal Tfchnique J~HJrnal of PUblic EconomIcs 8 pp 211~224 bull
Shaven J J Whalley 1984 Applied GeneralmiddotEquilibrlum MOdels of Taxalion and
International Tr8de An Introduction and Survey Journal of Economic literature 23t pp
1007middot1051
Slemrod J 1980 A General Equilibrium Model of Capital Income Taxation Ph D bullDissertation Harvard Univnrsity
Slemrod J 1983 A General Equilibrium Model of Taxation with Endogenous Financial
Behavior in M Feldstein edbull Behavioral Simulation Methods In Tax Policy Analysis Chicago
UniVerSity of Chicago Press
Stone J 1985 Sequential Optimization and Complementarily Techniques for computing
Economic equilibrium in A Manne Ed
Summers L 1981 Taxation and Corporate Investment a q-Theory Approach BroQkinas
Summers L 1985 Taxation and the Size and ComposHion 01 the Capital Stock An Asset Price
Approach Mlmeo Harvard University
Whalley J 1975 A Geneal Equiloibrium Assessment of the 1973 United Kingdom tax
reform Economica 42 pp 139middot61
1JLTIllOS JIQll~)ltG PAPERS PUBLICADOS bull
I
I I
j
112 70 - LUCENA iJiCIIJo ~ IIEnvironrilent Honitoring and Orqanization Structure T (Setellibro 1987)
nI 71 - L~ENA Diogo A Note on the Representation of Information Stxueture (SeteJObro 1987)
nQ 72 - LUCENA D1CgO Envuorment 111)nitoring and Organization
Structure II (Setelilhro 1987)
ril 73 - AlmmE3 Antonio Pais e CAsPAR Yitor bull Tributacao Inoentivos e InvesUlllentos Analise Qualitativa noYellibro 19S7)
n2 74 - liAlltOSA rose Pedro Opt1lllal lIage R1gidity A SUggestedtlthodology to Teet tb Theory and an Application (Outubro 1987)
nI 75 - CABRAL L 11 B Three lIotes on SYJIIlletric Gemes witt itsYJIIlletric Equilibri (NoveJObro 1987)
n2 76 - LUCENlt D10gOTo Search or Not to Search (Fevereiro1967)
nQ 17 - SA Jorge asconcelos euroI A Theory of SynergyB (Fevereiro1987)
n2 76 - VILARES tanuei Jose Os Be IntenaMios IlIIportados ColIIO Factor de Produao (Julho 1987)
nQ 79 - SA J~rge VasCQncel03 e HoY to Copete and Co~unicate in tature Industrial Products_ (aio 1988)
n2 80 - ROIl Rafael Learning and Capacity Expansion in a New larket Under Uncertainty (Fevereiro 1988)
n2 81 PEREIRA Alfredo llarvilo Survey of Dynamic Computational General EquilibriUlll Hodel for Tax Policy Evaluation (Outubro 1987)
Qualquer in ormacao sabre o 1IIorlltlng Papers ja publlcados sera prestada pelo Secretariado de Apoio aos Docentes podendo os ~e3OSf
ser adquirido5 ne Secyo de Vlndas da Faculdade de EconollLia UNL na Travessa Esteviio Pinto Cltpolide - 1000 LISllOA
bull
I
I f
Economic Research WP 1107
Judd K 1965 ShonmiddotRun Analysis of Fiscal Policy in a Simple Perfect Foresight Model
Journal of poJiHcal 5CCJiW1l pp 298middot319 r
Upton D J Potbc J Sachsa nd L Summers 1982 Multiple Shooting in Rational
Expectations Models fuoometrlc 50 1329middot1333
Manne A ed 1985 fconQrnjc Eoullibrium Model FOrIDujaljQo and Solution Mathematical
Pogrammlng Study 23 Amsterdam NorthmiddotHolland
Manne A 1985 Or the Formulation and Solution 01 Economic Equilibrium Models In A
Manne ed ECQOQlli-Jr~ibmiddotum ~tQdel EQlJulailon and Solution Malhematical Programming
Study 23 Amsterdam NorthmiddotHolland
Mansur A and J Whalley 1984 Numerical Specification of Applied General Equilibrium
Models in Scarf H J Shoven Eds Aoplied General Eouillbrium AnaiysectIsect Cambridge Universily
Press Cambridge
Merrill 0 1972 Applications and Extensions of an Algorithm thaI Computes FixedmiddotPoints of
Cenain Upper Semimiddotcominuou Point to Set Mappings PhD Disserlation Unlvel$ity 01 Michigan
Negishi T 1960 Welfar Economics and Exislence I an Equnlbrlum for a Competitive
Economy MelroeCOOQ11ka 12 92middot97
Pereira A 1985 On the Existence and Uniqueness of Optimal Ou1put Path lor CRTS Firms in
Adjustment Costs Technologies Mimeo Stanford University
Pereira A 1985a Consumer Expectations and Ihe COSI 01 UllUty In an Intertemporat
Framework Mlmeo Stanford University
Pereira A 1985b Corporale Tax Integrallon and Government Deficits A Dynamic General
Equilibrium Anaiysls Mlmeo Stanford University
Pereir A 1987 On Ihe Computation of the Users Costs of Stocks Ecooomi (forthcoming)
47
Pereira A bull 1987b Equal Yield Alternatives and Government Oeficlts Mimeof Stanford
University
Pereira A 1987c A Dynamic Applied General ~uHlbrium Model for Tax Policy Evaluation
Comlletamp Docurnent2tion Mimeo University of California San Diego (in progress)
Pereira A 1987d Should the investment Tax Credit be Rei(ltroduced Mlmeo University
of California San Diego (ir progreSs)
Piggott J and J Whalley eds 1985 tJew QeyeQpments in Applied General EQuilibrium
bullenalysjs Cambridge Cambridge University Press
Preckel Pbull 1985 Alternative algorithms for computing economic equilibria In A Manne Ed
Radnor R 19amp3 Equllibrium under Uncertainty In K Arrow M Intrilligaor ods
I ibull
Robinson S 1 g86 Mullisectorai Models of Oeveloping CountrieS a Survey Oivision of
Agricullure and Resources UC Borkeley WP 401
Scarf H 1967 The approximation of fixed points of a continuous mapping ~MM Jouwal of
epplid Mathornallcs 15 pp 1328middot43
Scarf H with T Hansen 1973 00 the Computation of EconomiC Eguilibria New Haven Yale
University Press
Scarf H J Shoven Eds 1984 61l1llied Genera Elujlibrium 6nalyssect Cambridge University
~ Press Cambridge
Shoven J 1976 The Incidence and Efficiency Effects of Taxe on Income from Capital
Jurnal of ramie1 Economy 84 pp 1261-83
Shoven J 1983 Applied General Equilibrium Tax Modemng IMF Slaff PaoerS 30
Shoven J and LB Simon 1986 Share Repurchases and Acquisitions An Analysis of which
Firms Participate NBER Working Paper No 2243
bull
48
IShaven J and J Whalley 1972 ft General Equmbrium Calculation of lhe Effects of
Differential Taxation of Incoma from Capital in the US Jauroe of public Econgmics 1
pp281middot321 (
Shoven J bull and J Whalley 1973 General Equilibrium with Tax$s a computation procedure
and an existence pr00f ReviEhi of Economic Studies 40 pp 475middot490
Shoven J J Whalley 1977 Equal Yield Tax Alternatives A General Equilibrium
Computa~ui0nal Tfchnique J~HJrnal of PUblic EconomIcs 8 pp 211~224 bull
Shaven J J Whalley 1984 Applied GeneralmiddotEquilibrlum MOdels of Taxalion and
International Tr8de An Introduction and Survey Journal of Economic literature 23t pp
1007middot1051
Slemrod J 1980 A General Equilibrium Model of Capital Income Taxation Ph D bullDissertation Harvard Univnrsity
Slemrod J 1983 A General Equilibrium Model of Taxation with Endogenous Financial
Behavior in M Feldstein edbull Behavioral Simulation Methods In Tax Policy Analysis Chicago
UniVerSity of Chicago Press
Stone J 1985 Sequential Optimization and Complementarily Techniques for computing
Economic equilibrium in A Manne Ed
Summers L 1981 Taxation and Corporate Investment a q-Theory Approach BroQkinas
Summers L 1985 Taxation and the Size and ComposHion 01 the Capital Stock An Asset Price
Approach Mlmeo Harvard University
Whalley J 1975 A Geneal Equiloibrium Assessment of the 1973 United Kingdom tax
reform Economica 42 pp 139middot61
1JLTIllOS JIQll~)ltG PAPERS PUBLICADOS bull
I
I I
j
112 70 - LUCENA iJiCIIJo ~ IIEnvironrilent Honitoring and Orqanization Structure T (Setellibro 1987)
nI 71 - L~ENA Diogo A Note on the Representation of Information Stxueture (SeteJObro 1987)
nQ 72 - LUCENA D1CgO Envuorment 111)nitoring and Organization
Structure II (Setelilhro 1987)
ril 73 - AlmmE3 Antonio Pais e CAsPAR Yitor bull Tributacao Inoentivos e InvesUlllentos Analise Qualitativa noYellibro 19S7)
n2 74 - liAlltOSA rose Pedro Opt1lllal lIage R1gidity A SUggestedtlthodology to Teet tb Theory and an Application (Outubro 1987)
nI 75 - CABRAL L 11 B Three lIotes on SYJIIlletric Gemes witt itsYJIIlletric Equilibri (NoveJObro 1987)
n2 76 - LUCENlt D10gOTo Search or Not to Search (Fevereiro1967)
nQ 17 - SA Jorge asconcelos euroI A Theory of SynergyB (Fevereiro1987)
n2 76 - VILARES tanuei Jose Os Be IntenaMios IlIIportados ColIIO Factor de Produao (Julho 1987)
nQ 79 - SA J~rge VasCQncel03 e HoY to Copete and Co~unicate in tature Industrial Products_ (aio 1988)
n2 80 - ROIl Rafael Learning and Capacity Expansion in a New larket Under Uncertainty (Fevereiro 1988)
n2 81 PEREIRA Alfredo llarvilo Survey of Dynamic Computational General EquilibriUlll Hodel for Tax Policy Evaluation (Outubro 1987)
Qualquer in ormacao sabre o 1IIorlltlng Papers ja publlcados sera prestada pelo Secretariado de Apoio aos Docentes podendo os ~e3OSf
ser adquirido5 ne Secyo de Vlndas da Faculdade de EconollLia UNL na Travessa Esteviio Pinto Cltpolide - 1000 LISllOA
47
Pereira A bull 1987b Equal Yield Alternatives and Government Oeficlts Mimeof Stanford
University
Pereira A 1987c A Dynamic Applied General ~uHlbrium Model for Tax Policy Evaluation
Comlletamp Docurnent2tion Mimeo University of California San Diego (in progress)
Pereira A 1987d Should the investment Tax Credit be Rei(ltroduced Mlmeo University
of California San Diego (ir progreSs)
Piggott J and J Whalley eds 1985 tJew QeyeQpments in Applied General EQuilibrium
bullenalysjs Cambridge Cambridge University Press
Preckel Pbull 1985 Alternative algorithms for computing economic equilibria In A Manne Ed
Radnor R 19amp3 Equllibrium under Uncertainty In K Arrow M Intrilligaor ods
I ibull
Robinson S 1 g86 Mullisectorai Models of Oeveloping CountrieS a Survey Oivision of
Agricullure and Resources UC Borkeley WP 401
Scarf H 1967 The approximation of fixed points of a continuous mapping ~MM Jouwal of
epplid Mathornallcs 15 pp 1328middot43
Scarf H with T Hansen 1973 00 the Computation of EconomiC Eguilibria New Haven Yale
University Press
Scarf H J Shoven Eds 1984 61l1llied Genera Elujlibrium 6nalyssect Cambridge University
~ Press Cambridge
Shoven J 1976 The Incidence and Efficiency Effects of Taxe on Income from Capital
Jurnal of ramie1 Economy 84 pp 1261-83
Shoven J 1983 Applied General Equilibrium Tax Modemng IMF Slaff PaoerS 30
Shoven J and LB Simon 1986 Share Repurchases and Acquisitions An Analysis of which
Firms Participate NBER Working Paper No 2243
bull
48
IShaven J and J Whalley 1972 ft General Equmbrium Calculation of lhe Effects of
Differential Taxation of Incoma from Capital in the US Jauroe of public Econgmics 1
pp281middot321 (
Shoven J bull and J Whalley 1973 General Equilibrium with Tax$s a computation procedure
and an existence pr00f ReviEhi of Economic Studies 40 pp 475middot490
Shoven J J Whalley 1977 Equal Yield Tax Alternatives A General Equilibrium
Computa~ui0nal Tfchnique J~HJrnal of PUblic EconomIcs 8 pp 211~224 bull
Shaven J J Whalley 1984 Applied GeneralmiddotEquilibrlum MOdels of Taxalion and
International Tr8de An Introduction and Survey Journal of Economic literature 23t pp
1007middot1051
Slemrod J 1980 A General Equilibrium Model of Capital Income Taxation Ph D bullDissertation Harvard Univnrsity
Slemrod J 1983 A General Equilibrium Model of Taxation with Endogenous Financial
Behavior in M Feldstein edbull Behavioral Simulation Methods In Tax Policy Analysis Chicago
UniVerSity of Chicago Press
Stone J 1985 Sequential Optimization and Complementarily Techniques for computing
Economic equilibrium in A Manne Ed
Summers L 1981 Taxation and Corporate Investment a q-Theory Approach BroQkinas
Summers L 1985 Taxation and the Size and ComposHion 01 the Capital Stock An Asset Price
Approach Mlmeo Harvard University
Whalley J 1975 A Geneal Equiloibrium Assessment of the 1973 United Kingdom tax
reform Economica 42 pp 139middot61
1JLTIllOS JIQll~)ltG PAPERS PUBLICADOS bull
I
I I
j
112 70 - LUCENA iJiCIIJo ~ IIEnvironrilent Honitoring and Orqanization Structure T (Setellibro 1987)
nI 71 - L~ENA Diogo A Note on the Representation of Information Stxueture (SeteJObro 1987)
nQ 72 - LUCENA D1CgO Envuorment 111)nitoring and Organization
Structure II (Setelilhro 1987)
ril 73 - AlmmE3 Antonio Pais e CAsPAR Yitor bull Tributacao Inoentivos e InvesUlllentos Analise Qualitativa noYellibro 19S7)
n2 74 - liAlltOSA rose Pedro Opt1lllal lIage R1gidity A SUggestedtlthodology to Teet tb Theory and an Application (Outubro 1987)
nI 75 - CABRAL L 11 B Three lIotes on SYJIIlletric Gemes witt itsYJIIlletric Equilibri (NoveJObro 1987)
n2 76 - LUCENlt D10gOTo Search or Not to Search (Fevereiro1967)
nQ 17 - SA Jorge asconcelos euroI A Theory of SynergyB (Fevereiro1987)
n2 76 - VILARES tanuei Jose Os Be IntenaMios IlIIportados ColIIO Factor de Produao (Julho 1987)
nQ 79 - SA J~rge VasCQncel03 e HoY to Copete and Co~unicate in tature Industrial Products_ (aio 1988)
n2 80 - ROIl Rafael Learning and Capacity Expansion in a New larket Under Uncertainty (Fevereiro 1988)
n2 81 PEREIRA Alfredo llarvilo Survey of Dynamic Computational General EquilibriUlll Hodel for Tax Policy Evaluation (Outubro 1987)
Qualquer in ormacao sabre o 1IIorlltlng Papers ja publlcados sera prestada pelo Secretariado de Apoio aos Docentes podendo os ~e3OSf
ser adquirido5 ne Secyo de Vlndas da Faculdade de EconollLia UNL na Travessa Esteviio Pinto Cltpolide - 1000 LISllOA
bull
48
IShaven J and J Whalley 1972 ft General Equmbrium Calculation of lhe Effects of
Differential Taxation of Incoma from Capital in the US Jauroe of public Econgmics 1
pp281middot321 (
Shoven J bull and J Whalley 1973 General Equilibrium with Tax$s a computation procedure
and an existence pr00f ReviEhi of Economic Studies 40 pp 475middot490
Shoven J J Whalley 1977 Equal Yield Tax Alternatives A General Equilibrium
Computa~ui0nal Tfchnique J~HJrnal of PUblic EconomIcs 8 pp 211~224 bull
Shaven J J Whalley 1984 Applied GeneralmiddotEquilibrlum MOdels of Taxalion and
International Tr8de An Introduction and Survey Journal of Economic literature 23t pp
1007middot1051
Slemrod J 1980 A General Equilibrium Model of Capital Income Taxation Ph D bullDissertation Harvard Univnrsity
Slemrod J 1983 A General Equilibrium Model of Taxation with Endogenous Financial
Behavior in M Feldstein edbull Behavioral Simulation Methods In Tax Policy Analysis Chicago
UniVerSity of Chicago Press
Stone J 1985 Sequential Optimization and Complementarily Techniques for computing
Economic equilibrium in A Manne Ed
Summers L 1981 Taxation and Corporate Investment a q-Theory Approach BroQkinas
Summers L 1985 Taxation and the Size and ComposHion 01 the Capital Stock An Asset Price
Approach Mlmeo Harvard University
Whalley J 1975 A Geneal Equiloibrium Assessment of the 1973 United Kingdom tax
reform Economica 42 pp 139middot61
1JLTIllOS JIQll~)ltG PAPERS PUBLICADOS bull
I
I I
j
112 70 - LUCENA iJiCIIJo ~ IIEnvironrilent Honitoring and Orqanization Structure T (Setellibro 1987)
nI 71 - L~ENA Diogo A Note on the Representation of Information Stxueture (SeteJObro 1987)
nQ 72 - LUCENA D1CgO Envuorment 111)nitoring and Organization
Structure II (Setelilhro 1987)
ril 73 - AlmmE3 Antonio Pais e CAsPAR Yitor bull Tributacao Inoentivos e InvesUlllentos Analise Qualitativa noYellibro 19S7)
n2 74 - liAlltOSA rose Pedro Opt1lllal lIage R1gidity A SUggestedtlthodology to Teet tb Theory and an Application (Outubro 1987)
nI 75 - CABRAL L 11 B Three lIotes on SYJIIlletric Gemes witt itsYJIIlletric Equilibri (NoveJObro 1987)
n2 76 - LUCENlt D10gOTo Search or Not to Search (Fevereiro1967)
nQ 17 - SA Jorge asconcelos euroI A Theory of SynergyB (Fevereiro1987)
n2 76 - VILARES tanuei Jose Os Be IntenaMios IlIIportados ColIIO Factor de Produao (Julho 1987)
nQ 79 - SA J~rge VasCQncel03 e HoY to Copete and Co~unicate in tature Industrial Products_ (aio 1988)
n2 80 - ROIl Rafael Learning and Capacity Expansion in a New larket Under Uncertainty (Fevereiro 1988)
n2 81 PEREIRA Alfredo llarvilo Survey of Dynamic Computational General EquilibriUlll Hodel for Tax Policy Evaluation (Outubro 1987)
Qualquer in ormacao sabre o 1IIorlltlng Papers ja publlcados sera prestada pelo Secretariado de Apoio aos Docentes podendo os ~e3OSf
ser adquirido5 ne Secyo de Vlndas da Faculdade de EconollLia UNL na Travessa Esteviio Pinto Cltpolide - 1000 LISllOA
1JLTIllOS JIQll~)ltG PAPERS PUBLICADOS bull
I
I I
j
112 70 - LUCENA iJiCIIJo ~ IIEnvironrilent Honitoring and Orqanization Structure T (Setellibro 1987)
nI 71 - L~ENA Diogo A Note on the Representation of Information Stxueture (SeteJObro 1987)
nQ 72 - LUCENA D1CgO Envuorment 111)nitoring and Organization
Structure II (Setelilhro 1987)
ril 73 - AlmmE3 Antonio Pais e CAsPAR Yitor bull Tributacao Inoentivos e InvesUlllentos Analise Qualitativa noYellibro 19S7)
n2 74 - liAlltOSA rose Pedro Opt1lllal lIage R1gidity A SUggestedtlthodology to Teet tb Theory and an Application (Outubro 1987)
nI 75 - CABRAL L 11 B Three lIotes on SYJIIlletric Gemes witt itsYJIIlletric Equilibri (NoveJObro 1987)
n2 76 - LUCENlt D10gOTo Search or Not to Search (Fevereiro1967)
nQ 17 - SA Jorge asconcelos euroI A Theory of SynergyB (Fevereiro1987)
n2 76 - VILARES tanuei Jose Os Be IntenaMios IlIIportados ColIIO Factor de Produao (Julho 1987)
nQ 79 - SA J~rge VasCQncel03 e HoY to Copete and Co~unicate in tature Industrial Products_ (aio 1988)
n2 80 - ROIl Rafael Learning and Capacity Expansion in a New larket Under Uncertainty (Fevereiro 1988)
n2 81 PEREIRA Alfredo llarvilo Survey of Dynamic Computational General EquilibriUlll Hodel for Tax Policy Evaluation (Outubro 1987)
Qualquer in ormacao sabre o 1IIorlltlng Papers ja publlcados sera prestada pelo Secretariado de Apoio aos Docentes podendo os ~e3OSf
ser adquirido5 ne Secyo de Vlndas da Faculdade de EconollLia UNL na Travessa Esteviio Pinto Cltpolide - 1000 LISllOA