Post on 14-Feb-2017
NATURAL RESOURCE MODELINGVolume 19, Number 1, Spring 2006
TIMBER MARKETS AND FUEL TREATMENTSIN THE WESTERN U.S.
KAREN L. ABTForestry Sciences Laboratory
USDA Forest ServiceP.O. Box 12254
Research Triangle Park, NC 27709E-mail: kabt@fs.fed.us
JEFFREY P. PRESTEMONForestry Sciences Laboratory
USDA Forest ServiceP.O. Box 12254
Research Triangle Park, NC 27709
ABSTRACT. We developed a model of interrelated timbermarkets in the U.S. West to assess the impacts of large-scalefuel reduction programs on these markets, and concomitanteffects of the market on the fuel reduction programs. Thelinear programming spatial equilibrium model allows inter-state and international trade with western Canada and therest of the world, while accounting for price effects of intro-ducing softwood logs to the market. The model maximizesarea treated, given fire regime-condition class priorities, max-imum increases in softwood processing capacity, maximumrates of annual treatments, prohibitions on exports of U.S. andCanadian softwood logs from public lands and a fixed annualtreatment budget. Results show that the loss to U.S. privatetimber producers is less than the gains for timber consumers(mills). Effects on Canadian private producers and Canadiantimber consumers depend on wildland-urban interface con-straints and condition class treatment priorities. States re-ceiving the greatest amount of treatments include California,Oregon, Washington and New Mexico, due to their concen-tration of stands with relatively low treatment costs.
KEY WORDS: Wildfire, mechanical treatments, spatialequilibrium, welfare.
Introduction. Recent increases in wildfire suppression expendituresand financial losses resulting from large fire events have increased theinterest in developing policies to address forest health conditions inwestern U.S. forests. One particular concern is that previous firesuppression and prevention, combined with climate change, has led
Received by the editors on June 1, 2004, and in revised form on Nov. 23, 2004.
Copyright c©2006 Rocky Mountain Mathematics Consortium
15
16 K.L. ABT AND J.P. PRESTEMON
to hazardous fuel accumulations in many forested areas (Parsons andDeBenedetti [1979], Bonnicksen and Stone [1982], Parker [1984], Chang[1996], Harrod et al. [1999], Arno et al. [1997], Skinner and Chang[1996]). A second concern is encroachment of human populationsinto these forests, often referred to as the wildland-urban interface(WUI), which creates the potential for increased loss of private propertyduring wildfire events. One proposed solution addressing both of theseconcerns is to reduce fuel accumulations in the WUI and surroundinglands using mechanical treatments.
The increase in fuel treatments has been promoted in the NationalFire Plan1 and the Healthy Forests Initiative (HFI)2 which Congressmodified and passed, as well as the Healthy Forests Restoration Act3.The USDA Forest Service adopted the Comprehensive Strategy andImplementation Plan4, partly in response to criticisms from the GAOregarding its lack of cohesion in developing a fuels reduction strategy(U.S. General Accounting Office [1999]). Current funding for fueltreatment programs on both public and private lands exceeds $1 billionper year, although the funding for mechanical treatments with biomassproduct removals is considerably smaller.
One effect of implementing these treatment programs is the de factoremoval of limits on federal harvests, in particular federal harvests thatwould result in below-cost timber sales. It is possible that areas mostin need of treatment are also the areas with the smallest proportionof high quality wood and the most expensive treatment costs. Thus,these treatments would require large subsidies in many areas becausethe treatments would not provide adequate products for sale whilesimultaneously reducing fuels.
Large-scale biomass removals programs done to lower wildfire risksand associated potential damages on National Forests, other govern-ment forests and private lands may have economic impacts across arange of temporal, spatial and sectoral scales, including short- andlong-run impacts on economies, impacts on local, regional and nationalforest product markets, and effects on other economic sectors in andaround the removal zone (Mercer et al. [2000]).
The treatments will affect the markets for privately supplied timberby increasing the supply of some products, thus lowering the avail-able prices and hence the quantity offered by private producers. In
TIMBER MARKETS AND FUEL TREATMENTS 17
a treatment program that maximizes treatment acreage subject to asubsidy computed as the costs of treatments minus revenues accruedfrom materials removed and sold, the market price effect could reducethe ultimate acreage treated. This paper presents a model wherein wemaximize acres treated given a subsidy, not allowing for offsetting treat-ment costs with treatment revenues. Instead, we assume that treatmentrevenues go to the Treasury and hence do not offset treatment costs.
The magnitude of the timber market effects will depend on thecurrent state of the market, including excess wood processing capacity.Some of these treatments would require a complete subsidy. That is,there would be no revenues from products removed. Others wouldpay at least partially for themselves through sales of timber products.Thus the amount of fuel treatment that could be accomplished woulddepend on the proportion of the treatments receiving subsidies, andthe subsidies would influence timber markets, and thus timber prices.
Two recent studies have addressed large scale fuel treatment pro-grams without directly addressing issues of market welfare and prices.A Strategic Assessment of Forest Biomass and Fuel Reduction Treat-ments in western States (Western Biomass Assessment) (Rummer etal. [2003]) quantified the biomass available for treatment in the west-ern U.S. by state. This study examined the potential biomass availableby using current prices, costs derived from a treatment model and in-ventory data from USDA Forest Service Forest Inventory and Analysis.A second study (Fried et al. [2003]) evaluated the potential location ofnew biomass processing facilities in Southern Oregon. Fried et al. [2003]evaluated conditions and treatable volumes in the region, as measuredon Forest Inventory and Analysis plots. Although their study did findlarge treatable volumes on the ground, some of which could be eco-nomically treated, investment potential hinged on the assumption thatprices were not endogenous to treatment quantities.
Our research extends the above analyses by directly addressing themarket impacts of alternative program scales. To account for marketprice effects, we use a spatial equilibrium market model of twelvewestern states and other trading partners, and assess the outcome of a$1 billion fuel treatment program. States included in the analysis areArizona, California, Colorado, Idaho, Montana, Nevada, New Mexico,Oregon, western South Dakota (the Black Hills), Utah, Washington andWyoming. The model objective is to maximize treated acres in federal
18 K.L. ABT AND J.P. PRESTEMON
forests, allowing for shifts in private timber harvest. After treatedacres (given price effects) are maximized, social welfare is maximizedas a secondary objective, allowing consuming mills to purchase woodand private producers to sell wood across state and national borders.The treatments by state, forest type and condition class, and thewelfare implications of alternative subsidy programs, are described anddiscussed. We also analyze how a national forest lands mechanical fuelreduction treatment program limited to the wildland-urban interfacewould affect timber markets and treatment applications.
Literature and methods. Timber markets of the western U.S. havelong been dominated by harvests from public lands, but the situationhas changed in the last decade (Haynes et al. [2001]). National forests ofthe West provided nearly half of all timber harvested during the period1950 1985. Since the late 1980s, however, harvests in the West aredominated by industrial and nonindustrial forestland owners. Marketsin the region have contracted overall, and timber outputs in majorproducing states have declined. Large-scale biomass removals couldresult in a near doubling of timber output in many parts of the fire-prone West.
Large changes in the amount of wood on local markets can affect thewelfare of timber producers and timber consumers by shifting supply(Holmes [1991]). Outward shifts in supply decrease prices and increaseoverall welfare. However, such shifts can have differential impactson various producer and timber consumer groups, especially whenthe outward shift in supply results in a contraction in demand foralternative products. Of particular interest is the effect of such shifts onprivate timber producers, whose outputs compete directly with outputsof national forests (Adams et al. [1996]).
Outward shifts in supply such as those associated with salvage re-moval programs are often short-run, affecting local markets for only afew months to a couple of years (Holmes [1991], Prestemon and Holmes[2000]). If such outward shifts in supply are perceived by demandersto be temporary, then demand will not increase and market responsesto supply increases are limited by local mill production capacities andmarkets for mill outputs. In the case of a short-run supply shift, timberconsumers gain but producers, especially producers not participatingin a biomass reduction program, may be harmed. This effect is similar
TIMBER MARKETS AND FUEL TREATMENTS 19
to programs such as tree planting incentives which are intended to in-crease future timber supplies (Boyd and Hyde [1989]). If supply shiftsare perceived as permanent, then arbitrage can lead to outward de-mand shifts that return prices to normal and increase economic welfaregenerally within the sector.
Timber supply in areas of significant federal timber harvests can bedescribed as consisting of two components: price-responsive privatesupply and policy-driven public supply. Increased public land timberharvests benefit mills but harm private timber suppliers by drivingdown the price of timber. In the calculation of timber market effects oftreatments, economic surplus is estimated using techniques describedby Just et al. [1982]. These techniques have been applied frequently inforestry, e.g., Holmes [1991], Wear and Lee [1993].
From the perspective of public timber supply, there are at leasttwo possible outcomes for the timber market with a biomass removalsprogram in place, and these two outcomes depend on whether biomassremovals substitute for regular removals from public lands or merelyadd to them. There would be no immediate market effects on producersor timber consumers in the short-run if (1) product removals frombiomass treatments substitute for regular harvests and (2) the costsper unit to get the product to markets are the same and (3) the mix ofproducts going to markets remains the same. Over the long run, theprograms would affect residual stands, including growth rates, finalproducts obtained from a different inventory structure and wildfiresalvage volumes obtained from the forests under an altered wildfirerisk.
In this analysis, biomass removals are assumed to supplement regularpublic harvests. The price effects of these programs would dependon how fast timber demand can adjust to a larger supply obtainedfrom the treatments. Large, brief programs of biomass removals willdepress prices, harming private producers and benefiting wood productmanufacturers (the timber consumers).
From the perspective of timber demand, the impact of biomassremovals programs on markets will depend on the demand response,either (1) affecting trade across regions and the process of spatialarbitrage and/or (2) altering capacity utilization rates at existing millsand/or (3) influencing the creation of new capacity in the vicinity of
20 K.L. ABT AND J.P. PRESTEMON
the treatment zones. The current study evaluates only the short-runsituation where local demand stays constant, thus market responses(1) trade and (2) capacity utilization will be incorporated. Local priceeffects may be moderated if the products from treatment programscan be profitably moved outside the region. Such movement, however,would have economic effects outside the region that should be accountedfor (Murray and Wear [1998]). Recent research (Prestemon and Holmes[2000], Nagubadi et al. [2001], Bingham et al. [2003]) has shown thatlow product values and high shipping costs or market inefficiencies canlimit the transmission of local market changes to more distant markets.
Figure 1 illustrates the welfare, price and quantity effects arising froma fuels treatment program in a market with both public and privateharvests but no trade. Private timber supply is price-responsive, in-creasing in quantity with increasing price. Public supply is symbolizedby a vertical line, SG
0 . Price in the market for timber is set where thecurve representing the sum of private supply and public supply, S0,intersects the timber derived demand curve, D, at point a, resulting inthe equilibrium price P0 and quantity Q0 without a fire-related biomassremovals program. Producer economic surplus is the area above theprivate supply curve and below price, area P0ac. Economic surplusaccruing to timber consumers is the area above price and below thedemand curve, area daP0.
Where the harvests from a large scale program of fire-related biomasstreatments on national forests add to regular removals, thus not sub-stituting for regular harvests, the government supply curve shifts toSG
1 , so that total supply shifts outward, to S1. The total quantityoffered with the shifted supply curve, including private supply, is Q1,and price drops to P1. Producer economic surplus for private timberproducers is area feb, while government revenues from treatments andregular harvests amount to area P1fe. Surplus accruing to timber con-sumers is larger than without the biomass removals and is now areadbP1. Note that the producer surplus accruing to private producersis reduced unambiguously by such a treatment program, as the mar-ket price declines and their volume sold shrinks5. Also, these changesmight apply to producers and consumers locally, but the opportuni-ties to ship product outside a region of fuel treatments can reduce thewelfare effects of such a program. Emphases on different parts of atreatment region can also have anomalous effects on consumers and
TIMBER MARKETS AND FUEL TREATMENTS 21
d
P0
P1
QG0 Q
G1 Q0 Q1
a
c e
b
S0
S1
SG0 SG
1
D
f
FIGURE 1. Timber market structure in areas with public timber harvests,with a representation of the effects of a biomass removals program.
producers. If treatment programs effectively substitute less valuabletreated material for more valuable untreated material, provided thatthe government subsidizes such treatments, consumers may be left onlymarginally better off and producers much worse off.
If supply shifts past the point where demand intersects the horizontalaxis, then producers lose all timber market surplus they sell no timber,as the market price for timber is zero and public landowners get novalue. In fact, the public would have to pay timber harvesters to removethe biomass. Timber consumers would still gain, however, as wood isprovided to them at no cost or they would need to be paid to cut it6.
A similar story could be told about benefits accruing to producersand consumers if, instead of new capacity developing locally, demandshifts outward as a result of expanded wood product exports out ofthe region to other parts of the United States or abroad. The timberconsumers benefited in that case would reside outside of the treatmentregion.
Another component of market analysis is identification of intermarketrelationships. If markets are integrated, that is, if local market shocksare transmitted, see Ravallion [1986], and biomass harvests are large,
22 K.L. ABT AND J.P. PRESTEMON
0
80,000
160,000
240,000
320,000
400,000
AZ
CA
CO ID MT
NV
NM OR
SD UT
WA
WY
Are
a T
reat
ed (
Acr
es)
CC 3 Only Available
CC 2 & CC 3 Available
All CC Available
A
B
0
80,000
160,000
240,000
320,000
400,000
AZ
CA
CO ID MT
NV
NM OR
SD UT
WA
WY
Are
a T
reat
ed (
Acr
es)
CC 3 Only Available
CC 2 & CC 3 Available
All CC Available
FIGURE 2. Treated by state under alternative condition class constraints,without (A) and with (B) state-level treatment proportional constraints, undera treated acreage maximization primary objective, on national forests of theUnited States.
TIMBER MARKETS AND FUEL TREATMENTS 23
then the price effects of these harvests can be transmitted across re-gions, resulting in economic effects of these harvests that go beyond thetreatment zone. Unless these intermarket relationships are understoodand quantified, single market analyses would reach erroneous conclu-sions. One method of evaluating the degree of market linkages is byexamining the costs of transfer of various forest products among pro-ducing and consuming markets. These are the methods outlined bySamuelson [1952] and Takayama and Judge [1964].
Data. The 12 states included in the model incorporate 13 foresttypes and are modeled as producing 4 softwood products which aregrouped here as (1) ponderosa pine, including ponderosa pine (Pinusponderosa) and sugar pine (Pinus lambertiana), (2) lodgepole pine(Pinus contorta), (3) other softwoods and (4) softwood chips. In thecurrent model, the chip market uses a fixed price of $0.35 per cubicmeter and is assumed to be not influenced by the treatment program7.
The model allows for trade with other regions by adding westernCanada and the rest of the world as two additional markets. West-ern Canada includes British Columbia and Alberta. The remainingCanadian provinces, the eastern and southern U.S., and all other coun-tries are referred to as the rest of the world. Northern Alberta andNorthern BC are modeled as lodgepole pine, southern Alberta as con-taining mainly ponderosa pine and BC as containing other softwood.Trade with the rest of the world is allowed only by the coastal statesand easternmost states, while trade with Canada is allowed only bythe northern border states. Exports of logs from public lands in boththe western U.S. and from Canada are restricted by law and thus areconstrained to zero in the model. Exports of logs from private landsare allowed from both countries. Import taxes on logs are set to zero,consistent with the North American Free Trade Agreement and the ma-jority of U.S. and Canadian log export destinations and import sources.
Inventory and acres. Inventory volume and acreage data are fromthe Western Biomass Assessment (Rummer at al. [2003]) and consistof volumes and acres by major species group by state, Table 1. Thesevolumes were then proportionately assigned to the wildland urban in-terface (WUI) and to fire regime-condition classes (CC) by the propor-tion of the total area in each state in these classes. Wildland-urban
TA
BLE
1.
Nati
onalfo
rest
tim
ber
land
acr
es,pro
port
ion
trea
table
,tr
eata
ble
acr
esby
fire
regim
e-co
nditio
ncl
ass
,and
pro
port
ion
inth
ew
ildla
nd
urb
an
inte
rface
.
Sta
teTota
lP
roport
ion
Tre
ata
ble
acr
esby
fire
Pro
port
ion
tim
ber
land
trea
table
regim
e-co
ndit
ion
class
ofacr
es
acr
esC
C1
CC
2C
C3
inW
UI
Ari
zona
18,3
64,4
81
0.5
11,1
35,3
20
5,5
48,0
57
2,6
91,0
24
0.0
5
Califo
rnia
38,4
59,6
10
0.4
02,8
01,0
71
7,7
07,6
96
4,8
08,2
16
0.1
8
Colo
rado
20,5
67,5
47
0.4
53,4
20,4
35
3,2
42,3
24
2,6
56,3
55
0.0
6
Idaho
21,9
38,0
49
0.5
11,9
24,5
02
5,8
25,4
47
3,3
94,1
00
0.0
1
Monta
na
21,2
72,3
89
0.5
12,6
46,0
15
4,6
35,8
77
3,5
36,8
36
0.0
1
Nev
ada
8,9
23,4
68
0.5
8783,7
11
2,8
85,9
08
1,4
76,2
79
0.0
3
New
Mex
ico
19,3
20,4
42
0.4
01,6
18,8
55
3,3
40,9
44
2,7
25,0
91
0.0
2
Ore
gon
27,7
38,7
34
0.4
01,9
10,8
62
4,6
92,7
01
4,5
68,9
23
0.0
4
South
Dakota
1,6
40,5
14
0.4
745,0
48
243,3
08
489,5
31
0.0
1
Uta
h15,2
44,3
99
0.4
83,8
52,1
92
2,9
58,8
95
502,6
21
0.0
4
Wash
ingto
n19,8
09,1
21
0.3
62,6
76,2
79
3,3
25,4
12
1,2
22,3
34
0.1
2
Wyom
ing
10,8
28,3
93
0.4
33,3
95,1
83
967,8
92
338,6
05
0.0
1
TIMBER MARKETS AND FUEL TREATMENTS 25
interface area was calculated using wildland-urban interface maps pro-vided directly by Radeloff (V. Radeloff, personal communication [May2003]). Fire regime-condition class represents a broad scale determina-tion of fire risk, where condition class represents the degree of departurefrom historical fire regimes, with CC 1 closest to historical and CC 3furthest away from the historical fire regime.
Treatments are applied only to national forest timberland and lessthan 60 percent of the timberland is assumed harvestable, consistentwith the Western Biomass Assessment (Rummer et al. [2003]). Ourestimates of potentially treatable acres by state by condition class arein Table 2. This table also shows the proportion of treatable timberlandin each state that is assumed to be in the wildland-urban interface.
Current production. Current harvest data were developed using the1997 Forest Statistics of the United States removals and product data(Smith et al. [2001]). Hardwood removals account for only 4 percentof all timber removals in the 12 states and were not further addressedin this study. The removals used as base data for the model are shownin Table 2. Removals are reported by national forest ownership and bynon-national forest ownership and by major species group.
Base level mill capacities and timber production levels in the U.S. andCanada derive from Spelter and Alderman [2003] and were adjusted foreach state to reflect non-included mills. Outside the U.S. and Canada,softwood log production and processing capacities were set at 100%and 110% of production, respectively, as reported by FAO [2004] for2002.
Treatments. The fuel reduction treatment, applied to all 13 foresttypes and 3 condition classes, removes volume on a stand to reducestand density to 30 percent of maximum stand density index for thatforest type and ecoregion. This treatment removes trees from all sizeclasses, concentrating most removals in the small to mid-sized diameterclasses. The prescription had no upper diameter limit on removals.Fiedler et al. [2001] found that removals from all diameter classes hadmore effect on reducing fire risk measures than did removals from onlythe smaller diameter classes. The SDI prescription is an ‘average’treatment that allows analysis of all western forests at an aggregate
TA
BLE
2.
Rem
ova
lsfr
om
nati
onalfo
rest
sand
oth
erow
ner
ship
s
by
majo
rsp
ecie
sgro
up
by
state
.
Nati
onalfo
rest
sN
on-n
ati
onalfo
rest
s
Sta
tePonder
osa
Lodgep
ole
Oth
erPonder
osa
Lodgep
ole
Oth
er
pin
epin
eso
ftw
ood
pin
epin
eso
ftw
ood
Thousa
nd
board
feet
Ari
zona
29,2
04
08,1
60
15,7
25
04,3
94
Califo
rnia
411,4
47
37,7
02
1,5
99,8
91
96,5
12
8,8
44
375,2
83
Colo
rado
3,5
89
5,4
33
22,0
27
1,9
33
2,9
25
11,8
61
Idaho
64,3
07
60,1
89
473,4
51
34,6
27
32,4
10
254,9
35
Monta
na
41,6
93
89,0
15
282,5
45
22,4
50
47,9
31
152,1
40
Nev
ada
1,0
91
433
1,7
25
587
233
929
New
Mex
ico
16,5
78
012,3
10
8,9
27
06,6
28
Ore
gon
272,1
49
44,5
63
2,4
84,7
29
79,0
11
12,9
38
721,3
73
South
Dakota
13,0
02
06,7
70
25,2
40
013,1
42
Uta
h1,7
01
1,5
01
14,8
35
916
808
7,9
88
Wash
ingto
n151,4
03
42,7
57
2,5
15,4
21
43,9
56
12,4
13
730,2
84
Wyom
ing
5,4
79
9,3
55
19,4
73
2,9
50
5,0
37
10,4
86
TIMBER MARKETS AND FUEL TREATMENTS 27
level. Actual treatments for stands are not expected to correspondto these prescriptions and should be designed for each stand usinginformation that is not available in the Forest Inventory and Analysisdatabase.
Although many mechanical fuel treatments will not remove anyproducts, these treatments are not discussed in this analysis as theywill not have any market effects. It is also possible that treatments willbe chosen that will remove some products and leave non-merchantablebiomass on site. The treatment applied in this model does not allow forpartial product removal from the stand. Only sawlogs were modeledin this analysis, and products were merchandized as in the WesternBiomass Assessment (Rummer et al. [2003]).
Costs and prices. Product prices were derived from National ForestSystem Cut and Sold reports for the second quarter of 20038 and areshown in Table 3. Regional prices were adjusted by the percentageof harvest from each species group to provide species prices. Pricesdiffer by state and major species group, ranging from a low of $39 perthousand board feet (mbf) in Arizona and New Mexico for lodgepolepine to a high of $528/mbf in Oregon and Washington for ponderosapine. Treatment costs are determined using the methodology of theWestern Biomass Assessment (Rummer et al. [2003]), Table 3. TheSTHarvest model (Hartsough et al. [2001]) was developed to assessthe costs of removing small diameter timber from stands in the PacificNorthwest and was applied to volumes and acres in all western states.The resulting costs range from a low of $288 per acre in Utah’s othersoftwood types to a high of $3,044 per acre in Utah’s lodgepole pinetype. Variations are due to slope and distances to mill differences acrossstate and differences in treating various species groups in each state.
Trade between states and regions will occur when the net cost to animporting region is less than the cost of procuring logs locally. Thustransportation costs will be essential to development of trade patternsin the model. Following Fight (personal communication [May 2003])and Rummer et al. [2003], we assume the cost of transporting woodbetween states is $0.35/bone dry ton (bdt) per mile or $1/mbf/mile.Distance from stump to mill is proxied by the distance to mill average ineach state. Distances between states for trade purposes are determinedby using the distance between spatial-center of forestland in each state.
TA
BLE
3.
Del
iver
edlo
gpri
ces
and
trea
tmen
tco
sts
by
majo
rsp
ecie
sgro
up
by
state
.
Log
pri
ceTre
atm
ent
cost
Sta
tePonder
osa
Lodgep
ole
Oth
erPonder
osa
Lodgep
ole
Oth
er
pin
epin
eso
ftw
ood
pin
epin
eso
ftw
ood
$/th
ousa
nd
board
feet
$/acr
e
Ari
zona
43
39
200
1,7
96
1,6
32
383
Califo
rnia
300
280
400
1,1
39
1,2
08
2,3
98
Colo
rado
219
204
200
1,1
29
2,0
35
1,3
63
Idaho
493
460
400
816
2,2
86
1,3
14
Monta
na
493
460
400
816
2,2
86
1,3
14
Nev
ada
98
91
200
680
3,0
24
397
New
Mex
ico
43
39
200
1,3
22
1,8
91
504
Ore
gon
528
493
400
1,0
20
1,8
47
1,7
28
South
Dakota
219
204
200
994
2,0
35
1,3
68
Uta
h98
91
200
680
3,0
24
288
Wash
ingto
n528
493
400
1,0
20
1,8
47
1,7
28
Wyom
ing
98
91
200
994
2,0
35
1,3
68
TIMBER MARKETS AND FUEL TREATMENTS 29
Western U.S. spatial equilibrium model. The model is currentlysolved for a single representative period. We assess a five-year programbut do not grow the stands, treated or untreated. The model maximizeswestern U.S. national forest acres treated in each year of a 5-yearprogram subject to constraints. The acreage maximization occurs overall available acres in each forest type and condition class within eachstate. The objective function and constraints are shown in Table 4,the elasticities assumed are in Table 5 and a summary of the modelsimulations is in Table 6.
The model is solved first as a base case with the objective of max-imizing social net welfare across all producers and timber consumers(summing across western U.S., western Canada and the rest of theworld), not allowing mechanical fuel treatments and thus requiring nosubsidy. Consistent with Samuelson [1952], social net welfare (SNW) isthe sum of producer surplus plus consumer surplus minus transporta-tion costs. In our case, SNW also includes government harvest andtreatment revenues less government harvest costs and less the subsidy.National forest timber harvests are held constant, and a market equi-librium of private plus public production, consumption, and trade arefound. For national forest timber, regardless of WUI, two sets of simu-lations were run, with each set including three condition class-restrictedmodels. The first set modeled a $1 billion subsidy without a limit onexpenditures by state. The second set limited expenditures by stateto a maximum equal to the proportion of timberland that state hadin each condition class. The condition class-restricted models allowedtreatments on (1) CC 3 only, (2) CC 2 + CC 3, and (3) all three condi-tion classes. A similar set of simulations was done restricting treatmentto only the designated WUI acres.
Products analyzed include softwood sawtimber and chips, and weallow for a maximum of a 30 percent outward shift in demand, i.e.,demand capacity within a region can expand by 30 percent. Demandand supply elasticities with respect to price are inelastic and inventorysupply elasticity is assumed to be one, Table 5. These elasticities areconsistent with the published literature (Adams and Haynes [1980],Majerus [1980], Regional Forester [1984], Adams et al. [1986], Wear[1989], Adams et al. [1991], Newman [1987], Abt et al. [2000], Hayneset al. [2001]).
TA
BLE
4.
Spati
aleq
uilib
rium
model
obje
ctiv
efu
nct
ions
and
const
rain
ts.
Base
Model
Obje
ctiv
efu
nct
ion
Maxim
ize
soci
alnet
wel
fare
=su
mofpro
duce
rsu
rplu
s
+su
mofco
nsu
mer
surp
lus
+gover
nm
ent
reven
ues
from
regula
rharv
est−
gover
nm
ent
harv
est
cost
s
Const
rain
tsP
ublic
log
export
=0
Harv
est
cannot
exce
ed1.3
tim
escu
rren
tharv
est
Capaci
tyuti
liza
tion
islim
ited
toa
maxim
um
of
1.3
tim
escu
rren
tpro
duct
ion
and
a
min
imum
of.4
5tim
escu
rren
tpro
duct
ion.
Tre
atm
ents
occ
ur
only
on
feder
alla
nds.
Non-t
reatm
ent
gover
nm
ent
harv
est
isco
nst
ant.
Acr
esTre
ate
dM
odel
s
Obje
ctiv
efu
nct
ion
Maxim
ize
acr
estr
eate
d
Const
rain
tsA
bove
const
rain
tsplu
slim
itati
ons
on
trea
ted
acr
esby
wildla
nd
urb
an
inte
rface
class
,co
ndit
ion
class
and
state
alloca
tion
TA
BLE
5.
Dem
and
and
supply
elast
icitie
suse
dto
dev
elop
linea
rdem
and
and
supply
funct
ions
at
curr
ent
price
sand
quanti
ties
. Oth
erst
ate
s:
Califo
rnia
,O
regon
&W
ash
ingto
n
Soft
wood
saw
tim
ber
Supply
with
resp
ect
topri
ce0.3
0.4
3
Supply
with
resp
ect
toin
ven
tory
volu
me
11
Dem
and
wit
hre
spec
tto
pri
ce-0
.5-0
.5
TA
BLE
6.
Sum
mary
ofsi
mula
tions
for
afive
yea
rm
echanic
altr
eatm
ent
pro
gra
mof$1
billion
per
yea
r.Sim
ula
tions
wer
eru
nfo
rall
nati
onalfo
rest
lands
and
for
WU
I-only
lands.
Label
Sim
ula
tion
des
crip
tion
Base
Maxim
ize
net
soci
alw
elfa
re,no
trea
tmen
ts
No
alloca
tion
by
state
Subsi
dy
alloca
tion
tost
ate
sis
unco
nst
rain
ed
CC
3M
axim
ize
trea
ted
acr
eson
condit
ion
class
3acr
esonly
CC
2+
3M
axim
ize
trea
ted
acr
eson
condit
ion
class
2and
3acr
esonly
CC
1+
2+
3M
axim
ize
trea
ted
acr
eson
all
condit
ion
class
es
Alloca
tion
lim
ited
by
state
Subsi
dy
alloca
tion
tost
ate
sis
lim
ited
by
pro
port
ion
ofst
ate
inea
chco
ndit
ion
class
CC
3M
axim
ize
trea
ted
acr
eson
condit
ion
class
3acr
esonly
CC
2+
3M
axim
ize
trea
ted
acr
eson
condit
ion
class
2and
3acr
esonly
CC
1+
2+
3M
axim
ize
trea
ted
acr
eson
all
condit
ion
class
es
32 K.L. ABT AND J.P. PRESTEMON
The model has specific requirements for U.S. federal ownerships,recalling that treatments are allowed only on national forests. Publictimber supply (nontreatment harvest) is assumed non-price responsive,and is held constant for the simulations. Export of logs from federallands in both Canada and the U.S. is prohibited. Calculation of surplusfrom national forest harvests and treatments represents governmentrevenues from sales less costs of treatment or harvest. We assume thecurrent level of national forest harvest generates no surplus (revenues =costs), and use current cost/mbf as the cost of national forests harvestin all simulations.
Simulation results.
No wildland-urban interface constraints. The effect of the treatmentprogram on private timber markets is broadly negative, Table 7, reduc-ing SNW by about 18 percent in the western U.S., as an average acrosstreatment scenarios. This effect is net of gains accruing to western U.S.timber consumers due to cheaper and larger aggregate log volumes en-tering mills, losses for western U.S. private timber producers due tolower prices and lower private timber harvests in response, the costsof wood transport, losses in the value of regular timber harvested fromnational forests also due to lower prices, and the cost to taxpayers ofthe subsidy. For example, under the scenario that limits treatmentsto stands classed as condition class 3 (CC 3), losses in SNW amountto $1.05B in the U.S. West resulting from treatment of about 833,000acres. Relaxing constraints on condition classes to allow any conditionclass stand to be treated leads to similar losses. Timber consumers inthe U.S. West are the primary beneficiaries of a treatment program, asprices drop and consumption increases, with gains ranging from $0.55to about $0.82B, or 12 18 percent. Private producers in the U.S. Westlose no more than $0.19B in economic surplus, losses ranging from 5to 10 percent. Their losses are due to price and output reductions.Similarly, Canadian timber consumers gain, by 2 9 percent, becausedomestic Canadian prices drop and consumption increases in westernCanada, even while total output in that region declines. For exam-ple, if treatments are restricted to CC 3 lands, then Canadian tim-ber consumers gain $0.01B. If treatments are allowed on all conditionclasses, then these effects are accentuated, with gains to Canadian log
TA
BLE
7.
Sim
ula
tion
outc
om
esunder
abase
case
with
aso
cialnet
wel
fare
obje
ctiv
eand
six
acr
eage
maxim
izati
on
obje
ctiv
esw
her
eboth
WU
Iand
non-W
UI
acr
esare
trea
table
.
Sim
ula
tion
WU
I+
non-W
UI
acr
es
Base
case
No
alloca
tion
by
state
Alloca
tion
lim
ited
by
state
CC
3C
C2+
3C
C1+
2+
3C
C3
CC
2+
3C
C1+
2+
3
Billion
dollars
Soci
alnet
wel
fare
,W
.U
.S.
6.1
95.1
44.9
84.7
35.2
45.1
55.0
9
Subsi
dy
use
d0.0
01.0
01.0
01.0
00.9
81.0
01.0
0
Govt.
trea
tmen
tre
ven
ues
0.0
00.2
70.2
30.1
90.3
00.2
60.2
2
Reg
ula
rgovt.
harv
est
reven
ues
1.0
90.9
20.9
50.9
60.9
20.9
30.8
9R
egula
rgovt.
harv
est
cost
s1.0
11.0
11.0
11.0
11.0
11.0
11.0
1
Consu
mer
surp
lus,
W.U
.S.
4.5
75.2
75.1
35.1
25.2
75.2
35.3
9
Pri
vate
pro
duce
rsu
rplu
s,W
.U
.S.
1.7
91.6
41.6
81.7
01.6
41.6
51.6
0Tra
nsp
ort
cost
s,W
.U
.S.
0.2
50.9
61.0
11.2
40.9
00.9
01.0
1
Consu
mer
surp
lus,
W.C
anada
0.6
10.6
20.6
50.6
60.6
20.6
40.6
6P
riva
tepro
duce
rsu
rplu
s,W
.C
anada
0.8
90.8
70.8
50.8
40.8
70.8
60.8
4C
onsu
mer
surp
lus,
rest
ofth
ew
orl
d3.8
93.9
93.9
83.9
84.0
03.9
93.9
7P
roduce
rsu
rplu
s,re
stofth
ew
orl
d15.4
215.3
315.3
315.3
315.3
115.3
315.3
4
Acr
es
Tota
lare
atr
eate
d0
832,6
41
1,0
73,7
22
1,2
21,1
64
814,9
34
1,0
11,4
66
1,0
62,1
41
Ponder
osa
pin
e0
423,7
93
734,7
78
776,8
81
394,9
01
627,3
06
605,1
60
Lodgep
ole
pin
e0
137,8
13
5,7
21
39,1
32
127,7
72
25,8
75
5,7
17
Oth
erso
ftw
ood
0271,0
35
333,2
22
405,1
51
292,2
61
358,2
85
451,2
63
34 K.L. ABT AND J.P. PRESTEMON
consumers as high as $0.06B. U.S. western exports to the eastern U.S.and elsewhere outside the West help to alleviate downward price pres-sure in the western U.S. Note that these effects ignore somewhat com-pensatory effects on secondary product markets, e.g., lumber, thatcould be quantified but were not explicitly modeled here.
When state-level restrictions on treatment subsidies are imposed, allof the effects of a treatment program are changed. Timber consumers inthe western U.S. are benefited in a manner similar to an unconstrainedprogram, because little new log volumes are produced and prices do notdrop much. Also, the state-level restrictions appear to work in favor ofconsumers, forcing more volume to be treated in places with higher millcapacities and higher value product. Private producers in the regionare harmed more, in contrast, as public production uses available millcapacity and replaces more valuable private timber to accommodate thetreatment program. The amount of forest treated is also smaller whena state-level spending constraint is imposed. On average, imposingsuch a constraint reduces the area treated by about 3 percent, from833,000 to 815,000 acres when only CC 3 lands are treatable; when allcondition classes can receive mechanical fuel treatments, the effect ofthe constraint is to reduce the area treated by 15 percent.
Under this set of scenarios, not limited to wildland-urban interfaceforests, the area treated differs greatly across forest types, Table 7.In most cases, ponderosa pine types receive the greatest amount oftreatment because these stands comprise much of the riskiest andtreatable area, Table 1. When less risky forest stands are includedin the available treatable forests, lodgepole pine stands go largelyuntreated, while resources are devoted substantially to other softwoodforest types and to ponderosa pine types. Despite this heterogeneityin the area treated across forest types, the share of land receivingtreatment differs less radically across the West than in the simulationswithout a state constraint. This rather even distribution is reflectiveof the fact that ponderosa and other softwood types are common inthe region, especially compared to lodgepole, a type most common innorthern and high elevation locations.
As constraints on the condition class are relaxed to include all classes,treatments shift across states. States such as New Mexico, Idaho,Montana and Oregon experience smaller changes or even negativechanges in treated areas as constraints on condition class are relaxed.
TA
BLE
8.
Sim
ula
tion
outc
om
esunder
abase
case
with
aso
cialnet
wel
fare
obje
ctiv
e
and
six
acr
eage
maxim
izati
on
obje
ctiv
esw
her
eonly
WU
Iacr
esare
trea
table
.
Sim
ula
tion
WU
I+
non-W
UI
acr
es
Base
case
No
alloca
tion
by
state
Alloca
tion
lim
ited
by
state
CC
3C
C2+
3C
C1+
2+
3C
C3
CC
2+
3C
C1+
2+
3
Billion
dollars
Soci
alnet
wel
fare
,W
.U
.S.
6.1
96.1
46.0
76.0
36.1
56.0
86.0
6
Subsi
dy
use
d0.0
00.0
70.1
70.2
60.0
60.1
40.2
0
Govt.
trea
tmen
tre
ven
ues
0.0
00.0
70.1
70.2
50.0
70.1
40.2
1
Reg
ula
rgovt.
harv
est
reven
ues
1.0
91.0
71.0
51.0
41.0
71.0
41.0
2R
egula
rgovt.
harv
est
cost
s1.0
11.0
11.0
11.0
11.0
11.0
11.0
1
Consu
mer
surp
lus,
W.U
.S.
4.5
74.6
54.7
24.8
14.6
44.7
94.8
7
Pri
vate
pro
duce
rsu
rplu
s,W
.U
.S.
1.7
91.7
61.7
31.6
91.7
71.7
21.6
8Tra
nsp
ort
cost
s,W
.U
.S.
0.2
50.3
40.4
20.4
80.3
30.4
60.5
0
Consu
mer
surp
lus,
W.C
anada
0.6
10.6
10.6
10.6
10.6
10.6
10.6
1P
riva
tepro
duce
rsu
rplu
s,W
.C
anada
0.8
90.8
90.8
90.8
90.8
90.8
90.8
9C
onsu
mer
surp
lus,
rest
ofth
ew
orl
d3.8
93.8
93.9
03.9
03.8
93.8
93.8
9P
roduce
rsu
rplu
s,re
stofth
ew
orl
d15.4
215.4
215.4
115.4
115.4
215.4
215.4
2
Acr
es
Tota
lare
atr
eate
d0
52,6
45
115,7
38
171,0
76
47,6
34
99,0
84
148,2
72
Ponder
osa
pin
e0
33,2
20
61,3
74
67,9
08
32,0
66
58,8
11
65,2
11
Lodgep
ole
pin
e0
4,6
52
9,7
51
18,3
29
3,5
71
7,5
08
14,1
54
Oth
erso
ftw
ood
04,7
73
44,6
12
84,8
39
11,9
96
32,7
66
68,9
07
36 K.L. ABT AND J.P. PRESTEMON
The area added to the total mainly comes from Utah and California,which have an abundance of CC 1 land that can be cheaply treated andfor which capacity constraints do not bind. When state-level treatmentspending constraints are enforced, added acres in CC 1 come principallyfrom Utah, Washington and Wyoming. Utah and Washington haveavailable demand capacity, while Wyoming has the ability to cheaplyexport more material to the consuming mills of the eastern U.S. andMontana.
Wildland-urban interface constraints imposed. When treatment pro-grams are limited to treating national forest lands in the West includedin the wildland-urban interface, the market impacts and area treatedare much smaller, Table 8. In this set of scenarios, the $1B subsidyavailable is never exhausted, i.e., it never binds in simulations. Thiskeeps the annual cost of such a treatment program low and moderatesthe negative overall effects on welfare. For example, SNW, includinggovernment harvest revenues and costs, decreases by an average of onlyabout 2 percent across all scenarios, declining from about $6.19B to anaverage of $6.09B. The beneficial effects on western U.S. consumers aresimilarly small, rising 1 percent under scenarios limiting treatments toCC 3 WUI lands and 7 percent when all condition class lands in theWUI are treatable. Western U.S. private timber producers lose from 1to 6 percent in economic surplus under alternative scenarios.
Forest types receiving treatments reflect which kinds of treatableacres exist close to urban areas. In this case, the focus on California,Oregon and Washington means that more non-ponderosa pine forestsare treated. In fact, when all condition class constraints are relaxed,non-ponderosa forests get most of the treatment dollars allocated. Theeffects of the wildlandurban interface constraint on treatment leadsto obvious biases in states receiving the bulk of the mechanical fueltreatments, Figure 3. These biases result directly from the amountof wildlandurban interface and the available capacity; large availablecapacities tend to limit the negative impacts of the treatment program.In this set of scenarios, states receiving the lion’s share of the programare California, Washington and Oregon, with intermediate amountsgoing to Colorado, New Mexico and Arizona. The program thereforeleaves out the low population and low-capacity states of Nevada, SouthDakota, Wyoming and Utah. Although Montana, Idaho and even
TIMBER MARKETS AND FUEL TREATMENTS 37
A
0
20,000
40,000
60,000
80,000
100,000
AZ
CA
CO ID MT
NV
NM OR
SD UT
WA
WY
Are
a T
reat
ed (
Acr
es)
All CC Available
CC 2 & CC 3 Available
CC 3 Only Available
B
0
20,000
40,000
60,000
80,000
100,000
AZ
CA
CO ID MT
NV
NM OR
SD UT
WA
WY
Are
a T
reat
ed (
Acr
es)
All CC Available
CC 2 & CC 3 Available
CC 3 Only Available
FIGURE 3. WUI-only area treated by state, without (A) and with (B) state-level treatment proportional constraints, under a treated acreage maximizationprimary objective, on National Forests of the United States.
38 K.L. ABT AND J.P. PRESTEMON
Arizona have at least some capacity, their low amounts of nationalforest land with significant urban interface pressure means that they,too, get little of the subsidy under this scenario.
Conclusions. While mechanical treatments may not be the closestsurrogate to wildfire in fire prone areas, mechanical methods may bepreferred to prescribed fire, for several reasons. Fire-related mechanicaltreatments are not constrained by the logistical difficulties of weathertiming or by risk of damage to adjacent properties and do not createsmoke, with its attendant air quality consequences. In addition, somehave viewed broad-scale fuel treatments as potentially beneficial to tim-ber markets, reinvigorating communities, which in places of substan-tial federal forest holdings have been negatively affected economicallyby federal harvest reductions (Spelter et al. [1996], USDA Forest Ser-vice [2000], Paun and Jackson [2000]). However, treatment programadministrators and timber market analysts should understand them asmultifaceted, affecting fire risk, ecological values and the timber mar-ket. Further, timber market analysts should place the program’s mar-ket impacts into the context of trade in forest products. Finally, giventhat one focus of the Healthy Forest Restoration Act is to reduce fuelsin the vicinity of communities at risk of catastrophic wildfires, such ananalysis should evaluate the market impacts of imposing WUI-relatedconstraints on program spending allocations. The research reportedhere is a first attempt to understand these issues.
Our analysis has shown that large scale programs of mechanical fueltreatments have the potential to influence the timber market signifi-cantly. With large fuel reduction programs, we find that timber con-sumer welfare in the western U.S. could be enhanced by up to 18 per-cent if managers were successful in implementing the program broadlyacross the region. On the other hand, private timber producers wouldface declines of up to 11 percent in welfare, due to lower productionand prices. The revenues accruing to the U.S. Treasury through na-tional forest harvests would increase, providing more payments in lieuof taxes for counties where treatments occur. However, the revenueswould not rise by as much as the timber volume, since mill capacitylimits would be quickly reached in certain locations and transportingsome of the material out of the treatment zone is relatively expen-sive and therefore would not be economically feasible. Also, decision
TIMBER MARKETS AND FUEL TREATMENTS 39
makers and land managers need to be cognizant of potential geographi-cal inequities in such treatment programs across counties. For example,one county could receive much of the treatment and payments in lieuof taxes, due to its large federal holdings, while the neighboring countywith mainly private lands could be negatively impacted by lower tim-ber prices and lower output. From a broader geographical perspective,a program focusing on treating as many acres as possible under a fixedbudget would result in certain states receiving most of the subsidy.Coastal states and states with ponderosa pine forests would get muchtreatment, while high-elevation and northerly states such as Montana,Wyoming and Colorado would get little.
From the standpoint of treatment efficacy, a treatment program thatreduced timber market welfare by these magnitudes should have fireeffects and ecosystem restoration benefits that are at least equal to themarket losses. Inclusion of the expected net benefits of the treatmentsin reducing aggregate losses and costs of wildfire may provide a differentperspective on the value of a mechanical treatment program.
Large scale programs would have uneven or scale-dependent effectson our trading partners and other parts of the U.S. Western Canadiantimber consumers were found to be beneficiaries of a U.S. mechanicalfuel treatment program through lower prices and greater volumesconsumed, while producers there appear to lose 1 6 percent in producersurplus (due to lower prices and lower volume produced). Prohibitionsagainst the export of softwood logs from public lands in the U.S. West,however, mean that the effects of a treatment program on Canadawould be muted and constrained. This conclusion, however, ignores thebeneficial impacts of greater treatment volumes in the lumber marketin the U.S.: domestic lumber manufacturers would find themselves ina more favorable position relative to competing Canadian softwoodlumber producers, compared to no treatment program.
A program exclusively focused on the wildland-urban interface wouldhave small impacts on western and non-western timber markets in theU.S., Canada and abroad. If a program were implemented so thatonly WUI lands would receive treatments on national forest land, thisprogram would be most active in California, Washington and Col-orado. States such as South Dakota, Wyoming and Nevada wouldbe essentially ignored. Our analysis did not address the potential
40 K.L. ABT AND J.P. PRESTEMON
benefits of a WUI-only program in terms of reducing the costs ofwildfires.
On the other hand, given that the budget for treatments in a WUI-only program never binds, it seems logical that the program evaluatedin our research is not the kind that would be implemented. Instead,government would most likely create a program that imposes a WUI-first constraint.
This research leaves substantial room for further analysis. First, weconfined our analysis to evaluating trade impacts with western Canadaand the rest of the world. A more disaggregated analysis, at leastquantifying effects in eastern Canada and the eastern U.S., could pro-vide more relevant information for decision makers in government andforest industry. Second, this study assumed that treatment programswould add to regular timber harvest on federal lands. Further analy-sis would quantify how differing rates of substitution between regularharvest and treatment removals would affect treatment revenues, reg-ular harvest revenues and the amount of treatments ultimately com-pleted under a fixed budget. Managerial limits on federal lands lead tothe expectation of at least some substitution. Third, our analysis wasshort-run and single-year, ignoring allocations over time and changesin demand capacity. The market effects of a treatment program couldbe smaller if the treatments were optimally spread out over time tomaintain a desired maximum fire risk or fire fuel level. This model alsoignored potential adaptations in timber demand, particularly capacityadjustments.
Fourth, a model that accounted for the potential differential costs oftreatment in the wildland-urban interface compared to merely wildlandforests would quantify the tradeoffs among alternative WUI-focusedstrategies. Fifth, our analysis was limited to federal lands of the West,which is naive in at least one respect: much of the forests especially inneed of risk reduction treatments are privately owned. An economicframework that could correctly describe a subsidy program for mechan-ical fuel treatments on private lands would need to account for exter-nalities including free-rider behavior, adoption behavior and methodsof encouraging optimal treatment patterns on the landscape. Sixth, ouranalysis used fire regime-condition classes as the yardsticks for prior-itizing treatments across the U.S. West. Alternative approaches include
TIMBER MARKETS AND FUEL TREATMENTS 41
treating stands to maximize fire risk reduction, or to maximize productvolumes removed, or to maximize long-run net benefits.
Acknowledgments. The authors thank Robert C. Abt, David T.Butry, Roger Fight, R. James Barbour, Robert Rummer and Peter Incefor their advice in various stages of this paper’s development.
ENDNOTES
1. The National Fire Plan is available at: http://www.fireplan.gov.
2. The Healthy Forests Initiative is available at: http://www.fs.fed.us/projects/documents/HealthyForests Pres Policy%20A6 v2.pdf.
3. Healthy Forests Restoration Act, HR 1904, December 3, 2003. Available at:http://frwebgate.access.gpo.gov/cgi-bin/getdoc.cgi?dbname=108 cong bills&docid=f:h1904enr.txt.pdf.
4. A collaborative approach for reducing wildland fire risks to communities andthe environment: 10 year comprehensive strategy (8/2001) (http://www.fireplan/gov/reports/7-19-en.pdf) and Implementation Plan (5/2002) (http://www.fireplan.gov/reports/11-23-en.pdf).
5. We note here that integrated forest products firms, with both timberlandsand mills, would be ambiguously affected by a treatment program simultaneouslyreducing the value of their timber and also the price of log inputs to their mills.
6. In areas of high wildfire risk, biomass removals that are done at negative pricesfor the biomass may still yield positive net economic benefits for society, if treatmentcosts are less than benefits derived from reducing expected losses from wildfire. Thepayoffs from these kinds of biomass removals programs therefore require an analysisof the trade-offs of the program costs and expected wildfire net value changes.
7. Chip values from treatment harvests are small relative to log values. Tothe extent that treatments introduce these chips into markets, welfare of chipconsumers, e.g., panel and pulp mills, and chip producers will be affected. Wehave found in initial analysis that chip market volume effects are second order tolog market effects.
8. Cut and sold reports, National Forest System, Second Quarter 2003, availableat: http://www.fs.fed.us/forestmanagement/reports/sold-harvest/2003q1harv.shtml. Accessed on October 6, 2004.
REFERENCES
R.C. Abt, F.W. Cubbage and G. Pacheco [2000], Southern Forest ResourceAssessment Using the Subregional Timber Supply (SRTS) Model, Forest Prod. J.50, 25 33.
D.M. Adams, R.J. Alig, B.A. McCarl, J.M. Callaway and S.M. Winnett [1996],An Analysis of the Impacts of Public Timber Harvest Policies on Private ForestManagement in the United States, Forest Sci. 42, 343 358.
42 K.L. ABT AND J.P. PRESTEMON
D.M. Adams, C.S. Binkley and P.A. Cardellichio [1991], Is the Level of NationalForest Timber Harvest Sensitive to Price? Land Econ. 67, 74 84.
D.M. Adams and R.W. Haynes [1980], The 1980 Softwood Timber AssessmentMarket Model : Structure, Projections and Policy Simulations, Forest Sci. Monogr.22, 64 pp.
D.M. Adams, B.A. McCarl and L. Homayounfarrokh [1986], The Role of ExchangeRates in Canadian-United States Lumber Trade, Forest Sci. 32, 973 988.
S.F. Arno, H.Y. Smith and M.A. Krebs [1997], Old Growth Ponderosa Pine andWestern Larch Stand Structures: Influences of Pre-1900 Fires and Fire Exclusion,USDA Forest Serv. Res. Paper INT-RP-495, 20 pp.
M. Bingham, J.P. Prestemon, D.J. MacNair and R.C. Abt [2003], Market Struc-ture in U.S. Southern Pine Roundwood, J. Forest Econ. 9, 97 117.
T.M. Bonnicksen and E.P. Stone [1982], Reconstruction of a Presettlement Gi-ant Sequoia-Mixed Conifer Forest Community Using the Aggregation Approach,Ecology 63, 1134 1148.
R.G. Boyd and W.F. Hyde [1989], Comparing the Effectiveness of Cost Sharingand Technical Assistance Programs, in Forestry Sector Intervention: The Impactsof Public Regulation on Social Welfare, Iowa State Univ. Press, Ames, IA, 48 89.
C. Chang [1996], Ecosystem Responses to Fire and Variations in Fire Regimes, inSierra Nevada Ecosystem Project: Final Report to Congress, Vol. II Assessmentsand Scientific Basis for Management Options, Wildland Resources Center Rep.No. 37, Centers for Water and Wildland Resources, Univ. of California, Davis,1071 1099.
FAO [2004], FAOSTAT Forestry Data, 2002. Data available on the Internetat http://faostat.fao.org/faostat/collections?subset=forestry. Accessed onMay 10, 2004.
C. Fiedler, C. Keegan, C. Woodall, T. Morgan, S. Robertson and J. Chmelik[2001], A Strategic Assessment of Fire Hazard in Montana, Report submitted tothe Joint Fire Science Program. http://jfsp.nifc.gov/documents/NMreport.pdf.Accessed on November 21, 2004.
J.S. Fried, J. Barbour and R. Fight [2003], FIA BioSum: Applying a Multi-ScaleEvaluation Tool in Southwest Oregon, J. Forest. 101, 8.
R.J. Harrod, B.H. McRae and W.E. Hartl [1999], Historical Stand Reconstruc-tion in Ponderosa Pine Forests to Guide Silvicultural Prescriptions, Forest Ecol.Manage. 114 (2-3), 433 446.
B.R. Hartsough, X. Zhang and R.D. Fight [2001], Harvesting Cost Model forSmall Trees in Natural Stands in the Interior Northwest, Forest Prod. J. 51, 54 61.
R.W. Haynes, D.M. Adams, R. Alig, D. Brooks, I. Durbak, J. Howard, P. Ince,D. McKeever, J. Mills, K. Skog and X. Zhou [2001], The 2000 RPA TimberAssessment: An Analysis of the Timber Situation in the United States, 1996 to 2050,http://www.fs.fed.us/pnw/sev/rpa/rpa2000.htm. Accessed on March 1, 2001.
T.P. Holmes [1991], Price and Welfare Effects of Catastrophic Forest Damagefrom Southern Pine Beetle Epidemics, Forest Sci. 37, 500 516.
B. Hartsough, X. Zhang and R. Fight [2001], Harvesting Cost for Small Trees inNatural Stands in the Interior Northwest, Forest Prod. J. 51, 54 61.
TIMBER MARKETS AND FUEL TREATMENTS 43
R.E. Just, D.L. Hueth and A.S. Schmitz [1982], Applied Welfare Economics andPublic Policy, Prentice-Hall, Inc., Englewood Cliffs, NJ, 491 pp.
G.A. Majerus [1980], Econometric Estimation of Demand and Supply Curves forTimber in Montana, 1962 1980, unpublished M.S. Thesis, University of Montana.
D.E. Mercer, J.M. Pye, J.P. Prestemon, D.T. Butry and T.P. Holmes [2000],Economic Effects of Catastrophic Wildfires: Assessing the Effectiveness of Fuel Re-duction Programs for Reducing the Economic Impacts of Catastrophic Forest FireEvents, Final Report, 68 pp. [Topic 8 of the research grant “Ecological and EconomicConsequences of the 1998 Florida Fires,” funded by the Joint Fire Science Program].http://flame.doacs.state.fl.us/joint fire sciences/economic.pdf. AccessedNovember 21, 2004.
B.C. Murray and D.N. Wear [1998], Federal Timber Restrictions and InterregionalArbitrage in U.S. Lumber, Land Econ. 74, 76 91.
V. Nagubadi, I.A. Munn and A. Tahai [2001], Integration of Hardwood StumpageMarkets in the Southcentral United States, J. Forest Econ. 7, 69 98.
D.H. Newman [1987], An Econometric Analysis of the Southern Softwood StumpageMarket: 1950 1980, Forest Sci. 33, 932 945.
A.J. Parker [1984], A Comparison of Structural Properties and CompositionalTrends in Conifer Forests of Yosemite and Glacier National Parks, USA, North-west Sci. 58, 131 141.
D.J. Parsons and S.H. DeBenedetti [1979], Impact of Fire Suppression on aMixed-Conifer Forest, Forest Ecol. Manage. 2, 21 33.
D. Paun and G. Jackson [2000], Potential for Expanding Small-Diameter TimberMarket: Assessing Use of Wood Posts in Highway Applications, USDA Forest Serv.Gen. Tech. Rep. FPL-GTR-120, 28 pp.
J.P. Prestemon and T.P. Holmes [2000], Timber Price Dynamics Following aNatural Catastrophe, Amer. J. Agr. Econ. 82, 145 160.
M. Ravallion [1986], Testing Market Integration, Amer. J. Agr. Econ. 68, 102 109.
Regional Forester [1984], Montana Timber Demand and Supply: A Short-RunModel of the Timber Market and an Assessment of the Economy, Employmentand Fuel Wood Use, unpublished report to the Montana State Forester from theNorthern Region, USDA Forest Serv., Missoula, MT.
B. Rummer, J.P. Prestemon, D. May, P. Miles, J.S. Vissage, R.E. McRoberts,G. Liknes, W.D. Shepperd, D. Ferguson, W. Elliot, S. Miller, S.E. Reutebuch, J.Barbour, J. Fried, B. Stokes, E. Bilek, K. Skog and B. Hartsough [2003], A StrategicAssessment of Forest Biomass and Fuel Reduction Treatments in Western States,http://www.fs.fed.us/research/pdf/Western final.pdf. Accessed on November21, 2004.
P. Samuelson [1952], Spatial Price Equilibrium and Linear Programming, Amer.Econ. Rev. 42, 283 303.
C.N. Skinner and C. Chang [1996], Fire Regimes, Past and Present, in SierraNevada Ecosystem Project: Final Report to Congress, Vol. II Assessments andScientific Basis for Management Options, Wildland Resources Center Rep. No. 37,Centers for Water and Wildland Resources, Univ. of California, Davis, 1041 1069.
W.B. Smith, J.S. Vissage, D.R. Darr and R.M. Sheffield [2001], Forest Resourcesof the United States, 1997. USDA Forest Serv. Gen. Tech. Rep. NC-219, 190 pp.
44 K.L. ABT AND J.P. PRESTEMON
H. Spelter and M. Alderman [2003], Profile 2003: Softwood Sawmills in the UnitedStates and Canada, USDA Forest Serv. Resource Paper FPL-RP-608.
H. Spelter, R. Wang and P. Ince [1996], Economic Feasibility of Products fromInland West Small-Diameter Timber, USDA Forest Serv. Gen. Tech. Rep. FPL-GTR-92, 17 pp.
T. Takayama and G.G. Judge [1964], Equilibrium among Spatially SeparatedMarkets: A Reformulation, Econometrica 32, 510 523.
U.S. General Accounting Office [1999], Western National Forests: A CohesiveStrategy Is Needed to Address Catastrophic Wildfire Threats, GAO/RCED 99-65,60 pp.
D.N. Wear [1989], Structural Change and Factor Demand in Montana’s SolidWood Products Industries, Canad. J. Forest. Research 19, 645 650.
D.N. Wear and K.J. Lee [1993], US Policy and Canadian Lumber : Effects of the1986 Memorandum of Understanding, Forest Sci. 39, 799 815.