Could Imports be Beneficial for Economic Growth? Some Evidence from Republic of Korea

8/22/2019 Could Imports be Beneficial for Economic Growth? Some Evidence from Republic of Korea

1/33

Economics and REsEaRch dEpaRtmEnt

cul ir be Beefl

fr E Grw:

se Evee fr

Reubl f Kre

Sangho Kim, Hyunjoon Lim,and Donghyun Park

October 2007

RD WoRking PaPER SERiES no. 103


2/33


3/33

ERD Wrkin Paper N. 103

Could Importsbe benefICIalfor eConomIC Growth?

some evIdenCefrom republICof Korea

sanGho KIm, hyunjoon lIm, and donGhyun parK

oCtober 2007

Sangho Kim is Proessor, College o Business, Honam University, Republic o Korea; Hyunjoon Lim is a graduate

student at the Department o Economics, University o Rochester, New York; and Donghyun PARK is Senior Economist,

Macroeconomics and Finance Research Division, Economics and Research Department, Asian Development Bank.


4/33

Asian Development Bank6 ADB Avenue, Mandaluyong City

1550 Metro Manila, Philippineswww.adb.org/economics

2007 by Asian Development BankOctober 2007

ISSN 1655-5252

The views expressed in this paperare those o the author(s) and do notnecessarily reect the views or policies

o the Asian Development Bank.


5/33

FoREWoRD

The ERD Working Paper Series is a orum or ongoing and recently completedresearch and policy studies undertaken in the Asian Development Bank or onits behal. The Series is a quick-disseminating, inormal publication meant tostimulate discussion and elicit eedback. Papers published under this Series

could subsequently be revised or publication as articles in proessional journalsor chapters in books.


6/33


7/33

CoNtENts

Abstract vii

I. Introduction 1

II. Literature Review 2

III. Empirical Analysis and Results 4

A. ariablesA. ariables 4 B. Causality between Trade and Productivity 8 C. Eects o Dierent Import Components on Productivity 12

I. Concluding Remarks 17

Appendix 19

Reerences 20


8/33


9/33

AbstRACt

The Republic o Korea is widely seen as a classical example o East Asiasexport-driven economic growth. The ocus in the literature on exports in theeconomys growth has led to an almost complete neglect o the role o imports.This study investigates the relationship between exports, imports, and economic

growth using quarterly data rom 1980 to 2003. Results indicate that imports havea signifcant positive eect on productivity growth but exports do not. Furthermore,the evidence reveals that the productivity-enhancing impact o imports is due

to competitive pressures arising rom consumer good imports and technologicaltransers embodied in capital good imports rom developed countries. Most othe studys results still hold using gross domestic product growth rather thanproductivity growth as the measure o economic growth. The evidence implies

that under certain circumstances, import liberalization can make a positive andsignifcant contribution to growth and development.


10/33


11/33

I. INtRoDuCtIoN

Many economists view the export-driven growth strategy as the cornerstone o East Asiasremarkable economic resurgence during the second hal o the 20th century. Exports contributed togrowth by acilitating labor mobilization and capital accumulation. Just as importantly, exposureto the fercely competitive global marketplace orced the regions frms to become more efcient.

In The East Asian Miracle, the World Bank (1993) suggested that exports and export-promotingpolicies had been instrumental in East Asias adoption o rontier technologies, which enhanced theproductivity o exporting frms and the economy in general, thus accelerating economic growth. In

addition, many studies provide empirical support or the export-led growth hypothesis by showingthat exports had a signifcant positive eect on Asias productivity and economic growth.1

An integral part o East Asias export-driven growth strategy was to protect domestic markets with

trade barriers so as to enable domestic frms to grow rom inancy to international competitiveness.This has let a legacy o viewing exports as benefcial and imports as harmul among the regionspolicymakers as well as general public. However, as Lawrence and Weinstein (1999) point out, WorldBank (1993) and the supportive empirical literature ocus only on the exportgrowth relationship,

ignoring the role o imports in promoting productivity and growth. In the case o Japan, Lawrenceand Weinstein ound that protectionism was actually harmul to productivity, and exports did notenhance productivity whereas imports did. This suggests that learning, innovation, and competitivepressures resulting rom oreign imports may be important channels or enhancing productivity at

the frm level as well as the economy as a whole.

This study uses data rom the Republic o Korea (henceorth Korea), a classic example o EastAsias export-driven economic growth, to empirically investigate the relationship between imports and

productivity growth. First, the paper investigates the direction o causality between trade variablesand productivity growth using a vector error correction model (ECM). Second, in order to identiythe specifc nature o the importproductivity relationship, equations are empirically estimated

to explain productivity in relation to exports, imports, and other variables. Results rom the twointerrelated empirical exercises indicate that imports, not exports, are a signifcant determinanto productivity. Furthermore, the benefcial impact o imports stems not only rom the competitivepressures arising rom imports o consumer goods, but also rom technological transers embodied

in imports o capital goods rom developed countries.

Previous empirical studies on the relationship between trade and economic growth in Koreaand other East Asian countries almost invariably defne trade as exports. As such, those studies

ocus exclusively on the relationship between exports and growth, and ignore the role o importsin growth. This is not surprising in light o the central role o export-oriented industrialization inthe East Asian miracle. However, the almost complete neglect o imports is still rather surprising

since economic intuition and theory suggest that imports may serve as an important channel or1 For surveys o the debates on productivity growth and trade in East Asia, see Chen (1997) and Edwards (1993),For surveys o the debates on productivity growth and trade in East Asia, see Chen (1997) and Edwards (1993),

respectively.


12/33

technological transer, productivity growth, and economic growth. Such neglect is also unortunatebecause it helps to reinorce the deep-seated regional bias against imports. This study attempts topromote a more balanced and comprehensive approach to the analysis o trade and economic growthin Korea and East Asia which, in turn will help to better understand the potential contribution o

imports to the regions growth and development.

II. LItERAtuRE REvIEW

In theory, there is a two-way causal relationship between trade and productivity, althoughadvocates o export-led growth generally contend that exports enhance productivity growth.2 These

economists argue that frms tend to learn advanced technologies through exports and must adoptthem to compete in the oreign marketplace.3 Firms also learn by doing, and emulate oreignrivals through trial and error inherent in the production and sale o export goods.4 Furthermore,the expansion in production resulting rom exports reduces unit production prices, thus increasing

productivity.5 In addition to these eects, exports also provide a country with oreign exchange,which is oten scarce in the early stages o economic development, enabling a country to import

capital and intermediate goods. Thus, or a variety o reasons, exports increase productivity growth. 6The reverse causation rom productivity growth to exports is also intuitively straightorward.

Productivity growth improves a countrys international competitiveness in price and quality, andthereby boosts its exports.

An extensive empirical literature exists on the relationship between exports and growth, largely

because o its bidirectionality. In act, much o the empirical literature on trade and productivitydefnes trade as exports rather than imports. Empirical studies have tried to determine whetherexports cause productivity to increase.7 However, results in this regard seem to depend on both

the sample periods and the countries examined. Some studies fnd unidirectional causality runningrom exports to productivity while others fnd reverse causality between the two variables.8 Clerideset al. (1998) argue that only relatively efcient frms engage in exports, and that exports do not

bring down unit production costs. The basic thrust o all these works is a unidirectional causalityrom productivity growth to exports.9 In their studies o frms in the United States (US), Bernardand Jensen (1999a and 1999b) also fnd that frms with high productivity usually export theirproducts, and exporting frms do not experience productivity and wage increases greater than thoseo nonexporting frms. On the other hand, a number o studies fnd either bidirectional causality

or absence o causality between exports and productivity.10 For Korea, some studies support theexport-led growth hypothesis, whereas other studies either ail to fnd causality rom growth to

2 See, or example, Baldwin (2003), Bonelli (1992), Haddad, De Melo and Horton (1996), Weinhold and Rauch (1997),See, or example, Baldwin (2003), Bonelli (1992), Haddad, De Melo and Horton (1996), Weinhold and Rauch (1997),

Yean (1997), and Sjoeholm (1999).3 Please reer to Balassa (1978), Krueger (1980) and Nishimizu and Robinson (1982).4 See rossman and Helpman (1991).See rossman and Helpman (1991).5 See Helpman and Krugman (1985).See Helpman and Krugman (1985).6 See McKinnon (1964).See McKinnon (1964).7 For a literature survey, see reenaway and Sapsord (1994).For a literature survey, see reenaway and Sapsord (1994).8 Haddad et al. (1996) is an example o the ormer while Pavcnik (2000) is an example o the latter.Haddad et al. (1996) is an example o the ormer while Pavcnik (2000) is an example o the latter.9 Shan and Sun (1998), or example, ail to fnd unidirectional causality rom exports to output and thus reject theShan and Sun (1998), or example, ail to fnd unidirectional causality rom exports to output and thus reject the

export-led growth hypothesis or the Peoples Republic o China.10 These include Hsiao (1987), Kunst and Marin (1989), and Jin and Yu (1996).These include Hsiao (1987), Kunst and Marin (1989), and Jin and Yu (1996).

OctOber 2007

cOuld ImpOrtsbebenefIcIalfOr ecOnOmIc GrOwth? sOme evIdencefrOm republIcOf KOrea

sanGhO KIm, hyunjOOn lImand dOnGhyun parK


13/33

sectIOn II

lIterature revIew

exports, or fnd bidirectional causality.11 The evidence on the direction o the causality betweenexports and growth in Korea is thus ambiguous at best, despite the widespread presumption o thevalidity o the export-led growth hypothesis.

The theoretical relationship between imports and productivity tends to be more complicated

than that between exports and productivity. Increased imports o consumer products encouragedomestic import-substituting frms to innovate and restructure themselves in order to competewith oreign rivals; thereore, imports enhance productive efciency. Under perect competition in

the neoclassical model, an industry reduces actor usage in the short run when trade barriers areremoved and the market is opened up to imports. In the long run, however, the industry becomesmore productive and competitive, and expands its investments in new technology, resulting in a

rightward shit o the industry supply curve.12 In general, the eect on productivity o opening themarket depends on both market structure and institutional actors. Under imperect competition,an import-substituting domestic market shrinks as imports increase, causing investment to all andthereby productivity to eventually all.13 Furthermore, higher uture expected profts lead to more

active research and development (R&D) investment and innovation eorts, and such R&D may begreater or exporting frms than or import-substituting frms in light o the large impact o marketopening. Imports o capital goods and intermediate goods that cannot be produced domesticallyenable domestic frms to diversiy and specialize, urther enhancing their productivity.14 Finally,

there are also theoretical grounds or both positive and negative causality rom productivity toimports.15

Relative to the empirical literature on exports and productivity, the number o empiricalstudies on the relationship between imports and productivity is quite limited. In particular, aspointed out earlier, the only study to empirically examine the relationship or KoreaLawrenceand Weinstein (1999)does so only very briey and ocuses on Japan. Their main fnding is that

imports contributed to total actor productivity (TFP) growth or a panel data set o Japanesemanuacturing industries, mainly through competition eects. A preliminary analysis in the samestudy ails to yield any systematic evidence that greater levels o protection improved productivityin Korea and the US. Lawrence (1999) shows empirically that import competition brought about

TFP growth in US industries. Muendler (2004) fnds that in the Brazilian manuacturing sector, thecompetitive eects o imports on competition are large even though the eect o intermediateimports on labor productivity is small.16

To summarize, the above review o the existing literature reveals two key trends in the ongoingdebate about the tradegrowth relationship. First, empirical tests o the export-led growth hypothesis

11 Jung and Marshall (1985), u (1996), and Choi (2002) support the export-led growth hypothesis. Darrat (1986),Jung and Marshall (1985), u (1996), and Choi (2002) support the export-led growth hypothesis. Darrat (1986),Hsiao (1987), Dodaro (1993), and Dutt and hosh (1996) ail to fnd causality rom growth to exports. Finally, Hsiao

(1987), Chow (1987), Bahmani-Oskooee and Shabsigh (1991), Bahmani-Oskooe and Alse (1993), and Jin (1995) fndbidirectional causality.

12 See Haddad et al. (1996) or a more comprehensive discussion. The basic argument is that market competition awakensSee Haddad et al. (1996) or a more comprehensive discussion. The basic argument is that market competition awakens

frms rom the laziness and comort o a monopoly market and provides incentives or innovation.13 See Tybout (2000) or an extended discussion. Schumpeter, however, suggested that a certain level o monopoly inSee Tybout (2000) or an extended discussion. Schumpeter, however, suggested that a certain level o monopoly in

the market provides frms with excess profts with which to make R&D investments, thus promoting productivity.14 See rossman and Helpman (1991), Sjoeholm (1999), and Tybout (2000).See rossman and Helpman (1991), Sjoeholm (1999), and Tybout (2000).15 Productivity growth triggers economic growth and increases income, which in turn stimulates imports. On the otherProductivity growth triggers economic growth and increases income, which in turn stimulates imports. On the other

hand, increased productivity in an import-substituting industry crowds out imports rom the domestic market and thushas a negative impact.

16 Additional studies on imports and productivity include okcekus (1997), who fnds that protectionism reduces technicalAdditional studies on imports and productivity include okcekus (1997), who fnds that protectionism reduces technicalprogress.

erd wOrKInGpaper serIesnO. 103 3


14/33

have produced mixed results. Second, with respect to the role o imports in growth, the empiricalliterature has lagged behind the theoretical literature. This study extends previous research on therelationship between trade and productivity in several directions. First, it hopes to contribute to thelimited empirical literature on the importsgrowth nexus by taking an in-depth look at the impact o

imports on Korean TFP. It is a particularly interesting to explore the role o imports in Korea sincethe countrys economic success has become a byword or export-led growth. Second, imports aredisaggregated into various components in order to more clearly understand the channels through

which imports aect productivity in Korea. Third, the cyclical eects that co-move with businesscycles are eliminated rom the productivity measures in order to control or spurious relation dueto cyclical biases. Finally, the data set is updated up to 2003 and a TFP equation is estimated toinvestigate the macroeconomic relationship between trade and productivity.

III. EmPIRICAL ANALysIs AND REsuLts

This section discusses the empirical analysis o the relationship between trade and TFP inKorea. It also reports the main empirical results and their implications about the role o trade in

Korean productivity growth.

A. variale

Many existing studies on the tradeproductivity nexus use labor productivity as the productivity

measure, but this partial measure does not allow or consideration o the eect o actor substitutionbetween capital and labor. This eect is especially important or the Korean economy, which hascontinuously experienced capital deepening and adoption o new production technologies. Measures

o labor productivity generally include the eects o capital deepening, along with technologicalprogress and structural efciency changes that determine TFP. It has recently been argued thatKorean economic growth was driven mostly by actor accumulation rather than by productivitygrowth. Thereore, TFP, rather than capital deepening or labor productivity growth resulting rom

trade-induced economies-o-scale, is used as the measure o productivity in order to measure theeects o trade on structural and technological changes.

Data on TFP were sourced rom various sources in the Bank o Korea database. Quarterly data

estimate Solow residuals or the period 1985Q12002Q4. The capital stock is taken to be the realamount o tangible fxed assets, adjusted or the capital utilization rate. In addition, labor input isproxied by the number o work hours, with gross domestic product (DP) as the measure o output.17All variables are converted into constant 1995 prices. The measured Solow residual is generally not

a good measure o productivity growth in the absence o perect competition, constant returns toscale, and ull employment o labor and capital. This implies that the measured Solow residual maybe aected by demand-side variables.18 In the case o Korea, Kim and Lim (2004) fnd that the

Solow residual is not a strictly exogenous variable but instead co-moves with demand shocks.

I measured productivities are indeed inuenced by cyclical movements, an empirical correlationbetween trade and productivity may be spurious in the sense that it is driven by a correlation

between trade and business cycles. For this reason, it is desirable to control or cyclical bias inthe productivity measure. To address this, the study ollows the method suggested by Basu and

17 For capital, the paper used the perpetual inventory method to expand the capital estimated by Pyo (2003).18 See Mankiw (1989) and Hall (1989) or more comprehensive discussions.See Mankiw (1989) and Hall (1989) or more comprehensive discussions.

OctOber 2007




15/33

sectIOn III

empIrIcal analysIsandresults

Kimball (1997) and Ball and Moftt (2001). This frst step in this method is to regress the logdierence o the measured Solow residual on the log dierence o the capital utilization rate, whichis a proxy or business cycles. The next step is to adjust the average o the regression error termso that it equals the original productivity measure when the productivity measure is adjusted or

cyclical actors.The estimation results indicate that the eect o the business cycle on the measured Solow

residual is signifcant.19 The estimation results are shown in (1) below, where CU denotes the

capital utilization rate and where t-statistics are inside the parentheses. 20 Results show thatremoving cyclical eects rom the measured Solow residual does not aect its overall movementbut considerably reduces its variation.21

log( ) . . log( )TFP CU t t= + 0 02 0 26 1 ,

( . ) ( . )8 11 3 89 R D W2

0 16 2 20= =. . . . (1)

Figure 1 shows the growth rates o the measured Solow residual, and the TFP estimates obtained

ater eliminating the cyclical eects rom the residual. TFP increased steeply ater the mid-1980sbut slowed somewhat in the 1990s, then ell sharply during the fnancial crisis o 19971998. TFPrecovered shortly ater the crisis but then ell again ater 2000.

fIGure 1Growthofthe solow resIdualand tfp wIth CyClICal adjustment

forthe Korean eConomy, 19802003

1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002

Solow Residuals

Fitted TFP Growth Rate

8

6

4

2

0

-2

-4

-6

-8

Percent

19 Labor productivity not being aected by the capital utilization rate is a well-established result in real business cycleLabor productivity not being aected by the capital utilization rate is a well-established result in real business cycle

theory. Other proxies or business cycles such as military spending, oil shocks, and political dummies have been

suggested. However, a complete treatment is beyond the scope o this paper.20 While the business cycle can aect productivity, productivity can also aect business cycles. To eliminate this endogeneityWhile the business cycle can aect productivity, productivity can also aect business cycles. To eliminate this endogeneity

problem, only lagged values o the capital utilization rate are included as explanatory variables in the regression.21 The purpose o adjusting TFP is to eliminate any error that may exist in the Solow residual as a productivity measure,The purpose o adjusting TFP is to eliminate any error that may exist in the Solow residual as a productivity measure,

i.e., to identiy the part o the Solow residual that represents pure productivity. While the cyclical movement o the

adjusted TFP is still smaller than the residual, the adjustment is not intended to completely eliminate the correlationbetween TFP and business cycles. The causality rom TFP to business cycles is well-established in real business cycle

theory, while the reverse eect rom business cycles to productivity should be eliminated to prevent a spuriousrelationship. Thereore, it is quite natural to have a high correlation ater adjustment.

erd wOrKInGpaper serIesnO. 103


16/33

Figure 2 illustrates recent trends in Koreas imports and exports. International trade accountsor a substantial share o the Korean economy. The ratio o exports (EP) to DP ell below 30%during the late 1980s and mid-1990s but was about 40% or more during other periods. Imports(IMP) showed a great deal o cyclical uctuation, but their share in DP ranged between 30% and

40%. Overall, the share o imports declined in the early 1980s but bounced back since the early1990s.

fIGure 2trendsIn Koreas exportsand Imports, 19802003

1982

1992

Imports/GDP Imports Imports Growth Rates

Exports/GDP Exports Exports Growth Rates

Imports

Percent60

50

40

30

20

10

0

-10

-20

-30

RatiosofI

mportstoGDP,

ImportsGrowthRates

(seasonallyadjustedgrowthratesoverperviousquarter)

ExportsPercent

70

60

50

40

30

20

10

0

-10

RatiosofExportstoGDP,

Ex

portsGrowthRates

(seasonallyadjustedgrowthratesoverperviousquarters)

AmountofImports

(seasonallyadjusted,at1995constantprices)

AmountofExpo

rts

(seasonallyadjusted,at1995

constantprices)

60,000

50,000

40,000

30,000

20,000

10,000

0

80,000

70,000

60,000

50,000

40,000

30,000

20,000

10,000

0

1980 1982 1984 1986 1988 1990 1994 1996 1998 2000 2002

(billion Korean won)

(billion Korean won)

19921984 1986 1988 1990 1994 1996 1998 2000 20021980

OctOber 2007




17/33

Beore perorming the empirical analysis, augmented Dickey-Fuller (ADF); Phillips-Peron (PP);and Kwiatkowski, Phillips, Schmidt and Shin (KPSS 1992) unit root tests were carried out to examinewhether the time-series o the variables ollow stochastic trends. Table 1 reports the test results orboth levels and frst dierences. The tests unambiguously suggest the existence o one unit root

or every variable, indicating that the time-series are integrated o order 1, I(1).table 1

unItroottestsofthe varIablesforthe relatIonshIpbetween tradeand produCtIvIty Growthfor Korea, 1980Q12003Q3

ADF PP KPss

I(0) I(1) I(0) I(1) I(0) I(1)

LIMP 2.18 5.59* 2.52 8.36* 0.16** 0.06LEP 1.96 4.53* 2.15 9.58* 0.18** 0.04

LTFP 0.15 4.53* 0.32 12.19* 0.28* 0.13****, ** and *** denote rejection o the null hypothesis at the 1, 5, and 10% signifcance level, respectively.LIMP means log o imports, LEP means log o exports, LTFP means log o total actor productivity.

Note: Test regressions contain a constant and a linear time trend, and lags o the dependent variable are chosen by Akaike InormationCriteria. The null hypothesis is the existence o unit root or augmented Dickey-Fuller and Phillips-Peron tests, and thenonexistence o unit root or Kwiatkowski, Phillips, Schmidt and Shin test.

To address this problem, the existence o long-run relationships among the variables is checked.It is possible to derive a long-run equilibrium that does not suer rom spurious regression.

Table 2 presents the maximum-likelihood ratio statistics, which indicate the number o long-runrelationships and thus the number o cointegration vectors in the parameter matrix. Johansenscointegration test is perormed on the log values o three sets o variables, namely (1) exports

(LEP) and TFP (LTFP); (2) imports (LIMP) and TFP (LTFP); and (3) exports, imports, and TFP. Thetest results indicate that a restricted constant, which allows a non-zero drit in the unit root process,is included in the multivariate system o equations. The lag values o the ECMs are set equal to

two. The null hypothesis o r=0 is rejected at the one-percent level but the null hypothesis o r1cannot be rejected.22 Consequently, the estimated likelihood ratio tests indicate the presence o acointegration vector and a long-run relationship in the underlying data-generating process o thetime-series variables.

22 See Osterwald-Lenum (1992) or critical values.See Osterwald-Lenum (1992) or critical values.

sectIOn III




18/33

table 2johansens loG lIKelIhood testfor CoInteGratIonofthe varIables forthe relatIonshIpbetween

tradeand produCtIvIty Growthfor Korea, 1980Q12003Q3

h0 : ranK=r eIGenvalue max-eIGenstat5%CrItICal traCestat. 5%CrItICal

LTFP, LEPNone 0.134

0.041

13.31

3.877**

14.26

3.841

17.19**

3.877**

15.49

3.841R1LTFP, LEP

None 0.298

0.023

32.63**

2.223

15.899.164

34.85**

2.223

20.269.164R1

LTFP, LIMP, LEP

None 0.3540.139

40.30**13.86

22.2915.89

56.35**16.04

35.1920.26R1

**, and *** denote rejection o the hypothesis at the 5 and 10% signifcance level, respectively.LTFP means log o total actor productivity, LEP means log o exports, LIMP means log o imports.

Note: Test regression includes a constant and a linear deterministic trend in the data. The test indicates 1 cointegrating equation atthe 5% signifcance level or every set o variables.

b. Caali eween trade and Prdcii

Based on the test results o the previous section, a vector error correction model (ECM) isestimated and tested or ranger causality on the basis o the coefcient estimates o the model.A ECM model consisting o the variables o IMP, EP, and TFP may be written as ollows:

y yt i tt

p

t= +=

11

, , y y y Z wt t i t t t j

p

= + + + =

1 11

1* ,

where i jj i

p*

=

= +

1

, = = + + + + ( )1 1 2Ik p , (2)

w Nt~ ( , )0 , and is a nondiagonal symmetric matrix.

In a ECM, all variables included in yt must satisy I(1), and residuals rom a long-runcointegrating relationship are used as lagged error correction terms in a AR. I =y ct representsa long-run cointegrating relationship and there is a deviation rom long-run equilibrium, the error

y ct is removed to restore equilibrium at adjustment speed .

The ECM models consist o two variables (IMP and TFP or EP and TFP) and three variables(IMP, EP, and TFP). The chosen ordering o variables is EP, TFP and IMP, TFP or the bivariate

models; and EP, IMP, and TFP or the trivariate model. This ordering reects the degree o exogeneityo the variables. However, changes in the order o these variables do not signifcantly aect theestimation results. Thus, yt = [LEP (LIMP), LTFP] or yt = [LEP, LIMP, LTFP], depending on the

number o variables considered. To consider the eects o the Asian crisis and to eliminate anyspurious correlation between trade and productivity growth, a dummy or the period 1998Q11998Q3is included as the exogenous variable (Zt) in (2) above. ECM systems with a lag length o two areestimated, and these lags are chosen to minimize the Akaike Inormation Criteria (AIC); however,changes in the lag length do not aect the results.

Table 3 below reports the ECM parameter estimates o the three variables, along with AR

OctOber 2007




19/33

model estimates. Although an autoregressive unit root characterizes every variable in the system,the AR model is also estimated to avoid any possible loss o valuable inormation owing todierencing.23 In particular, it can be expected that AR will elucidate the long-run relationshipamong the variables. Qualitatively, there are ew dierences between the ECM and AR coefcient

estimates. The most salient result o both models is that imports have a signifcant positive impacton TFP but exports do not.

table 3CoeffICIent estImatesofthe veCm forthe relatIonshIp

between tradeand produCtIvIty Growthfor Korea, 1980Q12003Q3

IndependentvarIables

veCtor error CorreCtIon model veCtor auto reGressIon modelLtFP LImP LEXP LtFP LImP LEXP

Error correction(-1)

0.001 0.493 0.028

(0.038) (3.985) (0.254)LTFP(-1) 0.330 0.627 0.664 0.602 0.090 0.389

(2.987) (1.500) (1.737) (6.139) (0.237) (1.119)LTFP(-2) 0.113 0.482 0.354 0.413 0.360 0.393

(1.085) (1.215) (0.976) (4.201) (0.941) (1.127)LIMP(-1) 0.064 0.068 0.102 0.066 0.825 0.073

(2.253) (0.635) (1.044) (2.338) (7.426) (0.730)

LIMP(-2) 0.016 0.005 0.209 0.074 0.052 0.046(0.526) (0.044) (1.933) (2.696) (0.491) (0.483)

LEP(-1) 0.047 0.120 0.063 0.043 0.086 0.955(1.400) (0.940) (0.542) (1.322) (0.679) (8.225)

LEP(-2) 0.022 0.205 0.014 0.047 0.111 0.018(0.663) (1.606) (0.121) (1.451) (0.868) (0.161)

Constant 0.016 0.036 0.038 0.060 0.524 0.004

(5.701) (3.299) (3.815) (1.523) (3.400) (0.030)

Dummy 0.015 0.114 0.013 0.013 0.103 0.004(2.223) (4.433) (0.564) (2.326) (4.491) (0.210)

R2 0.276 0.313 0.066 0.999 0.996 0.997

LTFP means log o total actor productivity, LIMP means log o imports, LEP means log o exports.Note: For ECM, all variables are frst dierenced or estimation. T-statistics are inside parentheses.

Figure 3 reports the impulse response unctions in the AR model, which are the simulatedresponses o TFP to the trade variables. The responses allow or investigation into the long-runrelationships between TFP and the trade variables. The impulse response unction extends over 10

quarters and is measured in terms o standard deviations. The eect o a one-standard-deviationshock to imports on TFP is initially positive and signifcant, and subsequently diminishes to zero.The eect o a shock to exports on TFP is positive but insignifcant over the whole period. Theresponses o TFP to import and export shocks imply that TFP is correlated with imports but not

with exports.

23 See Sims (1980) or a uller discussion.See Sims (1980) or a uller discussion.

sectIOn III




20/33

fIGure 3Impulse response funCtIonsIna var modelof

Imports, exports, and total faCtor produCtIvItyfor Korea, 19802003

1 2 3 4 5 6 7 8 9 10

1 2 3 4 5 6 7 8 9 10

1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10

1 2 3 4 5 6 7 8 9 10

1 2 3 4 5 6 7 8 9 10

.05

.04

.03

.02

.01

.00

-0.1

-0.2

.05

.04

.03

.02

.01

.00

-0.1

-0.2

.05

.04

.03

.02

.01

.00

-0.1

-0.2

.05

.04

.03

.02

.01

.00

-0.1

-0.2

.016

.012

.008

.004

.000

-0.04

-0.08

.016

.012

.008

.004

.000

-0.04

-0.08

Response of LIMP to LEXP Response of LIMP to LTFP

Response of LTFP to LEXP Response of LTFP to LIMP

Response of LEXP to LIMP Response of LEXP to LTFP

LEP means log o exports, LIMP means log o imports.

10 OctOber 2007




21/33

Table 4 reports ranger-causality tests based on the coefcient estimates o three dierentmodels. The estimated models consist o two-variable ECMs, a three-variable ECM, and a three-variable AR model. All the results indicate that there is no causality between exports and TFPgrowth. On the other hand, all the results indicate that imports ranger-cause TFP growth. In

addition, the AR model also indicates reverse causation rom TFP growth to imports. The fndingo no correlation between exports and productivity growth in Korea is consistent with the earlierfndings o Darrat (1986), Hsiao (1987), Dodaro (1993), and Dutt and hosh (1996). The most

striking eature o the results is the correlation between imports and productivity. The specifcmechanism underlying the importproductivity nexus can be investigated by decomposing importsinto their various components, such as consumer goods versus capital goods, an issue addressedin the next section.

table 4GranGer CausalIty testsforthe relatIonshIp

between tradeand total faCtor produCtIvIty Growthfor Korea, 1980Q12003Q3

null hypothesIs (h0) teststatIstICs (2) probabIlIty results

Bi-variate (ECM)LIMP LTFP 18.04* 0.0001 Reject

LTFP LIMP 1.176 0.555 Do not rejectLEP LTFP 4.088 0.129 Do not reject

LTFP LEP 1.366 0.505 Do not rejectTri-variate (ECM)

LIMP LTFP 5.987* 0.050 RejectLEP LTFP 2.595 0.273 Do not rejectLTFP LIMP 2.765 0.250 Do not reject

LEP LIMP 3.213 0.200 Do not rejectLTFP LEP 3.154 0.206 Do not reject

LIMP LEP 4.235 0.120 Do not rejectTri-variate (AR)

LIMP LTFP 7.282* 0.026 RejectLEP LTFP 2.228 0.328 Do not reject

LTFP LIMP 12.84* 0.001 RejectLEP LIMP 1.199 0.548 Do not rejectLTFP LEP 1.299 0.522 Do not reject

LIMP LEP 0.595 0.742 Do not reject

LIMP means log o imports. LTFP means log o total actor productivity, LEP means log o exports.Note: Test statistics are Wald statistics, and test results reer to the rejection o the null hypothesis at the 1% signifcance level.

Finally, as discussed in the literature, productivity growth has opposing eects on imports andmay cancel each other out. Productivity growth raises imports by raising income but also reduces

imports by increasing domestically produced import substitutes. The fnding that ranger causalityrom productivity to imports is signifcant only in the AR model may reect this ambiguity.

To check or the sensitivity o the results to the specifcation o economic growth, TFP growth

is substituted with DP growth in the ECM and AR models, and the same tests are perormed.Table A1 in the Appendix reports the ranger-causality tests based on the coefcient estimates

sectIOn III




22/33

o these models.24 The test results or DP are qualitatively identical to the results or TFP, i.e.,no causality exists between exports and DP, and imports ranger-cause DP growth in everymodel. Similarly, reverse causation rom DP growth to imports is also ound in the AR model.The empirical results are thus robust in the sense that they are not sensitive to the specifcation

o economic growth. This evidence again ails to directly support the export-led growth hypothesisor Korea during 19802003.

C. Effec f Differen Ipr Cpnen n Prdcii

The empirical results o the previous section suggest that causality runs rom productivity

growth to ination. Based on these results, imports are disaggregated into various components andincluded in a productivity determination equation or the purpose o investigating the importTFPrelationship in more detail. To determine the short-run dynamics o productivity determination, TFPgrowth is regressed against import components, exports, and other variables such as government

size and R&D investments. The TFP equation may be written as:

LTFP LIMP LEXP LGOV t i t i i t i i

m

i t i

i

n

= + + + + =

=

0 1 20

3

0

4ii t i t

i

p

i

l

LRD ==

+ 00

(3)

As additional explanatory variables, government size (O) and R&D investments (R&D), which

have been widely considered in the productivity literature, are used to represent institutional andtechnological actors, respectively. overnment consumption expenditure is used as a proxy or Oto capture the inefciencies arising rom government ailure. The number o patents registered inthe US is used as the measure o R&D. O and R&D are included in addition to the trade variables

in estimating the dynamic impact o trade variables on productivity. Imports according to country-o-origin are disaggregated into imports rom developed 7 countries (DIMP) and imports romother countries (OIMP). Developed countries in general and the 7 countries in particular are the

global technological leaders. As such, imports rom the 7 countries are more likely to embodyadvanced technology than imports rom elsewhere. Imports according to processing stages are alsodisaggregated into imports o raw materials (RIMP), capital goods (KIMP), and consumer products(CIMP). In particular, the critical distinction is between imported consumer goods and imported

capital goods. In contrast to consumer goods, capital goods such as machines are used to produceother goods. Thereore, while the main eect o consumer good imports is to intensiy competition inthe market or consumer goods, the main eect o capital good imports is to import the technologyembodied in the good, thus bringing about a more efcient production o other goods. The data

on the fve import components just defned are available rom Korea Trade Inormation Servicesonly ater 1988Q1.

Seasonality is eliminated rom the variables by means o an 12-ARIMA, and unit root tests

are perormed on the variables. Since every variable is integrated o order 1, I(1), frst dierencesare used in the actual estimation. Lags o the explanatory variables are chosen by means o thegeneral-to-specifc method, in which the most insignifcant lagged variable is eliminated iterativelyrom a set o lagged variables. In addition, the AIC criteria are applied in selecting the appropriate

number o lags. To incorporate the eects o the Asian crisis, the period 1998Q1Q3 is representedwith a dummy and included in the estimation.

24 The coefcient estimates are not reported here but available rom the authors upon request.The coefcient estimates are not reported here but available rom the authors upon request.

1 OctOber 2007




23/33

Table 5 presents the coefcient estimates or equation (3) or each o the models. Prior toestimation, an examination o the correlation among the import variables is made. The dierent

import components are ound to be strongly correlated with each other. The correlation between

(RIMP, KIMP); (RIMP, CIMP); (KIMP, CIMP); and (DIMP, OIMP) is 0.965, 0.888, 0.916 and 0.966.The high correlation between the import components causes multicollinearity, and the signifcanceo all import variables thus disappears when they are regressed together. Thereore, the dierent

import variables are not simultaneously included in the estimation.

table 5CoeffICIent estImatesof total faCtor produCtIvIty Growth eQuatIonfor Korea, 1988Q12003Q3


model 1 model 2 model 3 model 4 model 5 model 6

Intercept0.013

(6.603)*

0.013

(6.113)*

0.013

(6.161)*

0.013

(6.092)*

0.013

(6.645)*

0.013

(6.337)*

LIMP 0.079(3.420)*

LRIMP0.012

(0.562)

LKIMP0.042

(1.855)***

LCIMP0.042

(2.045)**

LDIMP0.058

(1.983)***

LOIMP0.045

(1.644)

LEP0.012

(0.398)

0.020

(0.617)

0.002

(0.954)

0.004

(0.136)

0.012

(0.352)

0.006

(0.184)

LO0.130(2.509)**

0.085(1.502)

0.074(1.352)

0.079(1.505)

0.075(1.395)

0.085(1.617)

LR&D0.036

(1.956)***0.039

(1.870)***0.040

(1.993)***0.036

(1.787)***0.037

(1.858)***0.041

(2.033)**

Dummy0.009(2.216)*

0.015(3.564)*

0.013(3.098)*

0.013(3.051)*

0.013(2.891)*

0.014(3.059)*

R 2 0.381 0.254 0.283 0.287 0.286 0.275

D.W. 2.467 2.456 2.506 2.403 2.489 2.487

*, ** and *** are statistically signifcant at the 1, 5 and 10% signifcance level, respectively.DIMP means imports rom developed 7 countries; OIMP means imports rom other countries; RIMP means imports o raw materials; KIMP

means capital goods; CIMP means consumer products; O means government consumption expenditure.Note: Absolute values o t-statistics are in parentheses.

The empirical results indicate that exports do not have a signifcant eect on TFP growth.Furthermore, their coefcients are all negative except or Model 2. These results ail to directlysupport the export-led growth hypothesis with respect to TFP growth in Korea. However, these results

do not imply that exports are not benefcial or developing economies since exports may improve

sectIOn III




24/33

economic perormance through channels other than TFP growth. For example, exports allow orthe realization o economies o scale and provide oreign exchange, thereby speeding up economicgrowth through capital deepening. This issue is discussed in more detail later.

The coefcients on government size (O) are all negative but insignifcant, except in the

case o Model 1, implying that increases in government spending may slow down TFP growth. TFPis a measure that captures productivity, which is based partly on institutional actors. Increasedgovernment spending may create inefciency in the orm o new regulations and bureaucracies.

However, in light o Os overall insignifcance, one should be careul about drawing such conclusions.All o the coefcient estimates on R&D are positive and signifcant. Those fndings suggest thatR&D promotes innovation and technical progress, and thus promotes TFP growth.

All import component coefcients are positive, but their signifcance depends on both producttype and country o origin. Imports o capital goods (KIMP) and consumer goods (CIMP) have positiveand signifcant eects on TFP, but raw material imports (RIMP) do not have any insignifcant eect.

In addition, the coefcient estimates on imports rom developed 7 countries (DIMP) are positiveand signifcant, but those rom other countries (OIMP) are insignifcant. Among the various import

components, consumer imports (CIMP) are the most signifcant and imports rom developed countrieshave the largest coefcients. The signifcant coefcient magnitudes o import components range

rom 0.042 to 0.058, suggesting that imports have a strong impact on TFP growth.

The empirical fndings suggest that increased imports o consumer goods intensiy marketcompetition in Korea. reater competition rom imports orces import-substituting Korean frms

to become more competitive by improving quality, cutting costs, or both. Examples o specifccompetitiveness-enhancing activities include adopting more eicient production techniques,engaging in innovation, and pursuing cost-cutting restructuring. Our fndings also suggest that

imports o capital goods help Korean frms to improve their productivity. I certain capital goodsare domestically unavailable, their availability through imports may enable a frm to use them toimprove the quality or reduce the costs o their products. Those imports can also enable a frm

to produce a wider range o products. Capital goods oten embody advanced technology and thusserve as powerul mechanisms or transerring technology to the importing frm. The results alsoimply that technology transer rom developed countries has a signifcant positive eect on theefciency o Korean frms. An important channel or importing superior technology is by importinggoods that embody superior technology. Developed countries in general and the large 7 economies

in particular are the global technological leaders. Thereore, imports rom those countries are muchmore likely to embody advanced technology, in particular technology unavailable to Korean frms,than imports rom other countries. In short, imports o consumer goods, capital goods, and goods

rom developed countries have all contributed to Korean productivity growth. The contribution oconsumer goods is largely through competitive eects while the contribution o capital goods andimports rom developed countries is largely through technology transer eects.

The intuition behind why competition should improve frms productivity is straightorwardand plausible. Competition pushes down prices and hence costs, reduces slack and misallocation oresources, provides incentives or organizing production more efciently, and may even potentiallypromote innovation. There are also more ormal theoretical explanations or a positive relationship

between competition and productivity.25 For example, an increase in the number o frms will

25 See Okada (2005) or an overview o the theoretical literature.See Okada (2005) or an overview o the theoretical literature.

1 OctOber 2007




25/33

lead to sharper managerial incentives and thus improve managerial eort.26 This is due to greateropportunities or comparing managerial perormance. Alternatively, greater competition may makeprofts more sensitive to managerial eort.27 A number o empirical studies provide support orthe notion that competition promotes productivity.28 Thereore, there are solid theoretical and

empirical oundations or why we observe a signifcant positive impact o consumer good importson Korean productivity.

riliches (1992) oers a theoretical basis or the papers fnding o signifcant positive TFP

eects o imports o capital goods and imports rom developed countries. riliches introduces thedistinction between embodied and disembodied technological spillovers. Embodied technologicalspillovers reer to knowledge and technology ows that arise directly rom ows o goods and

services between frms. riliches argues that such spillovers can arise rom either some kind olearningpure spilloversor the ability to reap the rents derived by the technological activitieso other frmsrent spillovers. More specifcally, rent spillovers are positive externalities that arisewhen the value o inputs exceeds the cost o inputs. The buying frm thus captures some o the rent

associated with the technology o the imported good. Pure spillovers are positive externalities thatarise rom the dissemination o various competencies as well as knowledge in general. Technologicalspillovers are by no means limited to a specifc country and can spread across borders. Indeed anumber o empirical studies confrm the importance o international technological spillovers as a

source o TFP growth.29 The fndings or Korea lend urther support to such evidence.

With regard to the tradegrowth nexus, the broad thrust o the results rom the TFP growthequations is that exports do not cause growth but imports have a signifcant positive impact on

growth. To check or the sensitivity o the empirical results to the specifcation o economic growth,TFP growth is replaced with DP growth. Although the results or the TFP growth regression indicatethat exports do not contribute signifcantly to TFP growth, this does not necessarily mean that

exports are not benefcial or a developing country, as pointed out earlier. In particular, exports allowan economy to realize economies o scale and provide it with oreign exchange, both o which actas catalysts or capital accumulation and thus more rapid economic growth. That is, exports maycontribute to growth through capital deepening in East Asia. This is in act the viewpoint o the

accumulationists, who assert that East Asian growth was mostly input-driven rather than productivity-driven. In short, exports may not cause TFP growth but nonetheless bring about economic growththrough capital deepening. Thereore, it is worthwhile to run DP growth regressions to examine

the role o exports in the Korean economy.

Table 6 reports the coefcient estimates or the DP growth regressions. The results o theDP and TFP growth regressions are broad. In particular, they both indicate that imports in general

and consumer imports in particular beneft growth. Imports as a whole (IMP) and consumer-goodsimports (CIMP) have a signifcant positive impact on DP growth. In addition, all import componentshave positive eects on DP growth, but their signifcances alls in two cases. Specifcally, importsrom developed countries (DIMP) and capital-goods imports (KIMP) become insignifcant. Thus, the

technological transer eects o imports are not as strong or DP growth as they are or TFP growth.The overall impact o exports seems to be greater or DP growth than or TFP growth since allexport coefcient signs become positive except in Model 10. However, the export coefcients are

26 See, or example, Nalebu and Stiglitz (1983).See, or example, Nalebu and Stiglitz (1983).27 See, or example, Willig (1987).See, or example, Willig (1987).28 See Okada (2005), Nickel (1996), and Nickel et al. (1997), among others.See Okada (2005), Nickel (1996), and Nickel et al. (1997), among others.29 Such studies include Coe and Helpman (1995), Keller (2000), runeld (2002), and Chuang and Hsu (2004).Such studies include Coe and Helpman (1995), Keller (2000), runeld (2002), and Chuang and Hsu (2004).

sectIOn III




26/33

still all insignifcant except in Model 8. The results thus still ail to directly support the export-ledgrowth hypothesis. All the R&D coefcients are insignifcant or DP growth, although they weresignifcant or TFP growth. Finally, all the government spending coefcients are insignifcant andnegative except in Model 9.

table 6CoeffICIent estImatesof Gdp Growth eQuatIonfor Korea, 1988Q12003Q3


model 7 model 8 model 9 model 10 model 11 model 12

Intercept0.010

(4.304)*0.012

(2.657)*0.012

(3.705)*0.014

(3.096)*0.013

(2.914)*0.013

(2.911)*

LIMP0.202

(7.216)*

LRIMP0.021

(0.884)

LKIMP0.041

(1.723)LCIMP

0.129

(3.212)*

LDIMP0.043

(1.342)

LOIMP0.041

(1.378)

LEP0.026

(0.684)0.109

(2.044) **0.077

(1.445)0.038(0.695)

0.085(1.533)

0.087(1.586)

LO0.098(1.559)

0.013(0.100)

0.039(0.467)

0.030(0.228)

0.012(0.095)

0.037(0.279)

LR&D0.021

(0.955)0.030

(0.917)0.024

(0.789)0.021

(0.670)0.022

(0.550)0.024

(0.597)

Dummy 0.015(3.041)*

0.032(4.677)*

0.029(4.367)*

0.024(3.351)*

0.030(4.209)*

0.030(4.416)*

R 2 0.674 0.376 0.399 0.413 0.384 0.385

D.W. 2.089 1.839 1.971 1.987 1.923 1.878

*, ** and *** are statistically signifcant at the 1, 5 and 10% signifcance level, respectively.DIMP means imports rom developed 7 countries; OIMP means imports rom other countries; RIMP means imports o raw materials; KIMP

means capital goods; CIMP means consumer products; O means government consumption expenditure.Note: Absolute values o t-statistics are in parentheses.

1 OctOber 2007




27/33

sectIOn Iv

cOncludInG remarKs

Iv. CoNCLuDINg REmARKs

The existing empirical literature on the relationship between trade and economic growth inKorea ocuses almost exclusively on the relationship between exports and growth. This is hardly

surprising in light o the act that Korea is very much part o the East Asian miracle, which wasdriven by export-oriented industrialization. The consequent neglect o the growth-promoting roleo imports in the literature on East Asian growth is unortunate because it reinorces a widespreadprejudice among policymakers and general public that exports are benefcial and imports are harmul.

Although such prejudice is global, it is perhaps stronger in the region, which has relied heavily onexport-oriented industrialization to power its rapid growth.

At a broader level, the central empirical result is that imports have a signifcant positive eect

on TFP. At a narrower level, the imports o consumer goods and capital goods have a signifcantpositive impact on productivity whereas raw material imports do not. Imports rom the developed7 countries have a signifcant positive eect on productivity but imports rom all other countries

do not. These fndings imply that the benefcial impact o imports stems not only rom competitivepressures arising rom the imports o consumer goods, but also rom technological transers embodied

in the imports o capital goods rom developed countries. The results signifcantly strengthen thecase or why empirical analysis o the relationship between trade and growth should incorporate

imports as well as exports, rather than just exports.

The papers ailure to fnd a signifcant positive eect o exports on productivity does notdisprove the export-led growth hypothesis. Nevertheless, the apparent absence o an eect is puzzling

given the central role o export-led industrialization in the East Asian miracle. To investigate thepossibility that exports contribute to economic growth primarily by mobilizing capital and laborrather than by promoting productivity, TFP growth is replaced with DP growth as the measure oeconomic growth, still, no eects were ound or exports. However, it is possible that the eect

o imports on productivity partly reects the eect o exports on productivity, i.e., competing inoreign markets orces frms to become more efcient. Two important means o becoming more

efcient are competing with oreign imports in domestic markets and absorbing advanced technologyembodied in imported capital goods. That is, imports may enhance international competitivenessand hence export capacity.

The empirical fndings have signifcant implications or policymakers. There is still a widespread

philosophical tendency in Korea and East Asia to view exports as benefcial and imports as harmul.However, the fndings clearly show that the notion o desirable exports and undesirable importsmay be misguided and counterproductive. In the case o Korea, imports have a signifcant positiveimpact on productivity growth and hence economic growth. Imports can promote productivity by

promoting both competition and adoption o advanced technology. The unmistakable implicationor policymakers is that import liberalization can stimulate institutional and technological progressconducive to productivity growth. In light o this evidence rom Korea, it is possible that import

liberalization will bring about substantial benefts in the orm o higher productivity and economicgrowth or other East Asian economies as well.



28/33

The two main limitations o the study stem rom data limitations. First, industry-specifc eectso trade on productivity could not be examined due to lack o Korean industry data. In act, suchdata limitations prevented Lawrence and Weinstein (1999) rom analyzing the importsproductivityrelationship in Korean industries. This is unortunate since imports inuence an economys productivity

through their impact on the productivity o frms and industries. Second, unavailability o pre-1980data limits the analysis to the post-1980 period. Koreas rapid growth and industrialization beganwell beore 1980 and thus the analysis does not cover Koreas transition rom a low-income to a

medium-income economy. This reduces somewhat the applicability o the policy implications orlow-income economies.30 An exciting area o uture research is to examine the impact o importson productivity in other East Asian countries so as to clariy the role o trade in the East Asianmiracle.31

30 On the other hand, intuitively there is no obvious reason why import-induced competition and technological progressshould be any less benefcial or productivity in low-income economies.

31 Thangavelu and Rajaguru (2004) fnd that imports have a signifcant positive eect on labor productivity in a numbero Asian economies.

1 OctOber 2007




29/33

appendIx

APPENDIX

GranGer CausalIty testsforthe relatIonshIpbetweentradeand Gdp Growthfor Korea, 1980Q12003Q3

null hypothesIs (h0) teststatIstICs (2) probabIlIty results

Bi-variate (ECM)LIMP LDP 13.36* 0.001 Reject

LDP LIMP 0.204 0.902 Do not rejectLEP LDP 0.224 0.893 Do not reject

LDP LEP 0.016 0.992 Do not rejectTri-variate (ECM)

LIMP LDP 11.68* 0.002 RejectLEP LDP 0.397 0.819 Do not rejectLDP LIMP 1.179 0.554 Do not reject

LEP LIMP 1.626 0.443 Do not reject

LDP LEP 0.866 0.648 Do not rejectLIMP LEP 3.330 0.189 Do not reject

Tri-variate (AR)

LIMP LDP 16.98* 0.000 RejectLEP LDP 0.056 0.972 Do not rejectLDP LIMP 7.611* 0.022 Reject

LEP LIMP 0.787 0.674 Do not rejectLDP LEP 0.123 0.940 Do not reject

LIMP LEP 0.865 0.648 Do not reject

LDP means log o DP, LEP means log o exports, LIMP means log o imports.Note: Test statistics are Wald statistics, and test results denote i the test rejects the null at the 5% signifcance level.



30/33

REFERENCEs

Bahmani-Oskooee, M., and J. Alse. 1993. Export rowth and Economic rowth: An Application o Cointegrationand Error Correction Modeling. Journal o Developing Areas 27:53532.

Bahmani-Oskooee, M., and . Shabsigh. 1991. Exports, rowth and Causality in LDCs: A Re-Examination.

Journal o Development Economics 36:40515.Balassa, B. 1978. Exports and Economic rowth. Journal o Development Economics 23:18189.

Baldwin, R. 2003. Openness and rowth: Whats the Empirical Relationship? NBER Working Paper No. 9578,National Bureau o Economic Research, Cambridge.

Ball, L., and R. Moftt. 2001. Productivity rowth and the Phillips Curve. NBER Working Paper No. 8421,National Bureau o Economic Research, Cambridge.

Basu, S., and M. Kimball. 1997. Cyclical Productivity with Unobserved Input ariation. NBER Working Paper

No. 5915, National Bureau o Economic Research, Cambridge.Bernard, A., and J. Jensen. 1999a. Exceptional Exporter Perormance: Cause, Eect or Both? Journal o

International Economics 47:126.________. 1999b. Exporting and Productivity. Yale University. Processed.

Bonelli, R. 1992. rowth and Productivity in Brazilian Industries: Impacts o Trade Orientation. Journal oDevelopment Economics 39:85109.

Chen, E. 1997. The Total Factor Productivity Debate: Determinants o Economic rowth in East Asia.Asian-Pacifc Economic Literature 10:1838.

Choi, Y. 2002. Exports and rowth: An Empirical Investigation in Korea. Quarterly Economic Analysis

8:375.Chow, P. 1987. Causality between Export rowth and Industrial Development: Empirical Evidence rom the

NICs.Journal o Development Economics 26:5563.Chuang, Y., and P. Hsu. 2004. FDI, Trade and Spillover Efciency: Evidence rom Chinas Manuacturing

Sector.Applied Economics 36:110315.

Clerides, S., S. Lach, and J. Tybout. 1998. Is Learning by Exporting Important? Micro-Dynamic Evidence romColombia, Mexico and Morocco. Quarterly Journal o Economics 103:90347.

Coe, D., and E. Helpman. 1995. International R&D Spillovers. European Economic Review39:85987.Darrat, A. 1986. Trade and Development: The Asian Experience. Cato Journal 6:69599.

Dodaro, S. 1993. Exports and rowth: A Reconsideration o Causality. Journal o Development Economics27:22744.Dutt, S., and D. hosh. 1996. The Export rowth-Economic rowth Nexus: A Causality Analysis. Journal o

Developing Areas 30:16781.Edwards, S. 1993. Openness, Trade Liberalization and rowth in Developing Countries. Journal o Economic

Literature 31:135893.okcekus, O. 1997. Trade Liberalization and Productivity rowth: New Evidence rom the Turkish Rubber

Industry. Applied Economics 29:63945.reenaway, D., and D. Sapsord. 1994. What Does Liberalization Do or Exports and rowth? Weltwirtschatliches

Archiv130:15274.

riliches, Z. 1992. The Search or R&D Spillovers. Scandinavian Journal o Economics 94s:29s47s.rossman, ., and E. Helpman. 1991. Innovation and Growth in the Global Economy. Cambridge: MIT Press.

runeld, L. 2002. International R&D Spillovers and the Eect o Absorptive Capacity: An Empirical Study.

NUPI Working Paper No.630, Norwegian Institute o International Aairs, Oslo.Haddad, M., J. De Melo, and B. Horton. 1996. Trade Liberalization, Exports and Industrial Perormance. In

M. Roberts and J. Tybout, eds.,Industrial Evolution in Developing Countries: Micro Patterns o Turnover,

Productivity and Market Structure. Washington, DC: Oxord University Press.

Hall, R. 1989. Temporal Agglomeration. NBER Working Paper No. 3143, National Bureau o Economic Research,Cambridge.

0 OctOber 2007




31/33

references

Helpman, E., and P. Krugman. 1985. Market Structure and Foreign Trade. Cambridge: MIT Press.

Hsiao, M. 1987. Tests o Causality and Exogeneity between Exports and Economic rowth: The Case o AsianNICs.Journal o Economic Development12:14359.

Jin, J. 1995. Export-Led rowth and the Four Little Dragons. Journal o International Trade and EconomicDevelopment4:20315.

Jin, J., and E. Yu. 1996. Export-led rowth and the U.S. Economy: Another Look.Applied Economic Letters3:34144.

Jung, W., and P. Marshall. 1985. Exports, rowth and Causality in Developing Countries. Journal o DevelopmentEconomics 18:112.

Keller, W. 2000. Do Trade Patterns and Technology Flows Aect Productivity rowth? World Bank EconomicReview14:1747.

Kim, S., and H. Lim. 2004. Does Solow Residual or Korea Reect Pure Technology Shocks? Paper presented

at the Far Eastern Meeting o the Econometric Society (FEMES), 1 July, Seoul.Krueger, A. 1980. Trade Policy as an Input to Development. American Economic Review70(2):28892.Kunst, R., and D. Marin. 1989. On Exports and Productivity: A Causal Analysis Study. Review o Economics

and Statistics 71:699703.Kwiatkowski, D., P. Philips, P. Schmidt, and Y. Shin, 1992. Testing the Null Hypothesis o Stationarity against

the Alternative o Unit Root: How Sure are We that Economic Time Series have a Unit Root? Journalo Econometrics 54(3):15978.

Lawrence, R. 1999. Does a Kick in the Pants et You oing or Does It just Hurt? The Impact o InternationalCompetition on Technological Change in US Manuacturing. In R. Feenstra, ed., Globalization and Wages.Chicago: University o Chicago Press.

Lawrence, R., and D. Weinstein. 1999. Trade and rowth: Import-led or Export-led? Evidence rom Japan andKorea. NBER Working PaperNo. 7264, National Bureau o Economic Research, Cambridge.

Mankiw, . 1989. Real Business Cycles: A New Keynesian Perspective. Journal o Economic Perspectives3:7990.

McKinnon, R. 1964. Foreign Exchange Constraint in Economic Development and Efcient Aid Allocation.Economic Journal 74:388409.

Muendler, M. A. 2004. Trade, Technology and Productivity: A Study o Brazilian Manuacturers 1986-98. CESIFOWorking PaperNo. 1148, Io Institute or Economic Research, Munich.

Nalebu, B., and J. Stiglitz. 1983. Prizes and Incentives: Towards a eneral Theory o Compensation andCompetition. Bell Journal o Economics 14:2143.Nickel, S. 1996. Competition and Corporate Perormance. Journal o Political Economy104:72446.

Nickel, S., D. Nicolitsas, and N. Dryden. 1997. What Makes Firms Perorm Well? European Economic Review41:78396.

Nishimizu, M., and S. Robinson. 1982. Trade Polices and Productivity Change in Semi-Industrialized Countries.Journal o Development Economics 16:177206.

Okada, Y. 2005. Competition and Productivity in Japanese Manuacturing Industries. NBER Working Paper No.

11540, National Bureau o Economic Research, Cambridge.Osterwald-Lenum, M. 1992. A Note with Quantiles o the Asymptotic Distribution o the Maximum Likelihood

Rank Test Statistics. Oxord Bulletin o Economics and Statistics 54:46171.Pavcnik, N. 2000. Trade Liberalization, Exit and Productivity Improvement: Evidence rom Chilean Plants. NBER

Working PaperNo. 7852, National Bureau o Economic Research, Cambridge.

Pyo, H. 2003. Estimates o Capital Stock by Industries and Types o oods in Korea (1953-2000). Panel orKorean Economic Analysis 9:20374.

Shan, J., and F. Sun. 1998. On the Export-Led rowth Hypothesis: The Econometric Evidence rom China.Applied Economics 30:105565.

Sims, C. A. 1980. Macroeconomics and Reality. Econometrica 48:148.Sjoeholm, F. 1999. Exports, Imports and Productivity: Results rom Indonesian Establishment Data. World

Development27:70515.



32/33

Thangavelu, S., and . Rajaguru. 2004. Is There an Export or Import-Led Productivity rowth in Rapidly

Developing Asian Countries? A Multivariate AR Analysis. Applied Economics 36:108393.Tybout, J. 2000. Manuacturing Firms in Developing Countries. Journal o Economic Literature 38:1144.

Weinhold, D., and J. Rauch. 1997. Openness, Specialization and Productivity rowth in Less DevelopedCountries. NBER Working PaperNo. 6131, National Bureau o Economic Research, Cambridge.

Willig, R. 1987. Corporate overnance and Market Structure. In A. Razin and E. Sadka, eds., Economic Policyin Theory and Practice. London: Macmillan.

World Bank. 1993. The East Asian Miracle, Economic Growth and Public Policy. New York: Oxord University

Press.u, Z. 1996. On the Causality between Export rowth and DP rowth: An Empirical Reinvestigation. Review

o International Economics 4:17284.Yean, T. 1997. Determinants o Productivity rowth in the Malaysian Manuacturing Sector.ASEAN Economic

Bulletin 13:33343.

OctOber 2007




33/33

abu e per

Sangho Kim, Hyunjoon Lim, and Donghyun Park write that while exports have longbeen recognized as an integral ingredient of the East Asian miracle, there has beena notable lack of research on the growth-promoting potential of imports. Empiricalevidence from Republic of Korea suggests that under certain circumstances importscan promote economic growth by fostering total factor productivity growth. Theanalysis implies that imports seem to have contributed to higher productivity byunleashing competitive pressures in domestic markets as well as by expanding theaccess of domestic producers to advanced technology from industrialized countries.

Asian Development Bank6 ADB Avenue, Mandaluyong City1550 Metro Manila, Philippineswww.adb.org/economicsISSN: 1655-5252

abu e a devele Bk

ADB aims to improve the welfare of the people in the Asia and Pacific region,particularly the nearly 1.9 billion who live on less than $2 a day. Despite manysuccess stories, the region remains home to two thirds of the worlds poor. ADB isa multilateral development finance institution owned by 67 members, 48 from theregion and 19 from other parts of the globe. ADBs vision is a region free of poverty.Its mission is to help its developing member countries reduce poverty and improvetheir quality of life.

ADBs main instruments for helping its developing member countries are policydialogue, loans, equity investments, guarantees, grants, and technical assistance.ADBs annual lending volume is typically about $6 billion, with technical assistanceusually totaling about $180 million a year.

ADBs headquarters is in Manila. It has 26 offices around the world and morethan 2,000 employees from over 50 countries.

< 0 1 0 0 6 0 7 0 >

Could Imports be Beneficial for Economic Growth? Some Evidence from Republic of Korea

Documents

Transcript of Could Imports be Beneficial for Economic Growth? Some Evidence from Republic of Korea