IMPACT OF FOREIGN DIRECT INVESTMENT AND

TRADE ON ECONOMIC GROWTH: EVIDENCE

FROM DEVELOPING COUNTRIES

SHIVA S. MAKKI AND AGAPI SOMWARU

Downloaded from https://academic.oup.com/ajae/article-abstract/86/3/795/63091 by guest on 04 February 2020

Foreign direct investment (FDI) and trade are results suggest that FDI, trade, human cap-
often seen as important catalysts for economic ital, and domestic investment are important
growth in the developing countries. FDI is an sources of economic growth for developing
important vehicle of technology transfer from countries. We find a strong positive interac-
developed countries to developing countries. tion between FDI and trade in advancing eco-
FDI also stimulates domestic investment and nomic growth. Our results also show that FDI
facilitates improvements in human capital and stimulates domestic investment. The contribu-
institutions in the host countries. International tion of FDI to economic growth is enhanced
trade is also known to be an instrument of by its positive interaction with human capital
economic growth (Frankel and Romer). Trade and sound macroeconomic policies and insti-
facilitates more efficient production of goods tutional stability.
and services by shifting production to countries
that have comparative advantage in producing
them. Review of Literature
Even though past studies show that FDI
and trade have a positive impact on economic Studies based on the neoclassical approach ar-
growth, the size of such impact may vary across gue that FDI affects only the level of income
countries depending on the level of human and leaves the long-run growth unchanged
capital, domestic investment, infrastructure, (Solow, De Mello). They argue that long-run
macroeconomic stability, and trade policies. growth can only arise because of technolog-
The literature continues to debate the role of ical progress and/or population growth, both
FDI and trade in economic growth as well as considered exogenous. Thus, according to neo-
the importance of economic and institutional classical models of economic growth, FDI will
developments in fostering FDI and trade. This only be growth advancing if it affects tech-
lack of consensus limits our understanding of nology positively and permanently. More re-
the role of FDI and trade policies in economic cent endogenous growth models, on the other
growth processes and restricts our ability to hand, imply that FDI can affect growth en-
develop policies to promote economic growth. dogenously if it generates increasing returns
This article analyzes the role of FDI and in production via externalities and spillover ef-
trade in promoting economic growth across fects. In these models, FDI is considered to be
selected developing countries and the interac- an important source of human capital and tech-
tion among FDI, trade, and economic growth. nological diffusion. FDI introduces new man-
We examine data from sixty-six developing agement practices and organizational arrange-
countries over the last three decades. Our ments in addition to providing labor training in
the host country production facilities. FDI en-
Shiva S. Makki is senior research specialist, Department of Agricul- courages the incorporation of new inputs and
tural, Environmental, and Developmental Economics, Ohio State technologies in the production systems of host
University; and Agapi Somwaru is senior economist, Economic countries.
Research Service, U.S. Department of Agriculture.
The authors gratefully acknowledge helpful comments received Applying endogenous growth theory to a
from Demcey Johnson, Mike Trueblood, Suchada Langley, Mary cross section of forty-six developing countries,
Bohman, and V.V. Chari.
The views expressed herein are those of the authors and not
Balasubramanyam, Salisu, and Sapsford show
necessarily those of the Economic Research Service or the U.S. that the growth-enhancing effects of FDI are
Department of Agriculture. stronger in countries that pursued a policy of
This article was presented at the ASSA winter meetings
(San Diego, CA, January 2004). Articles in these sessions are not export promotion rather than import substitu-
subjected to the journal’s standard refereeing process. tion. Their econometric analysis indicates that
Amer. J. Agr. Econ. 86(3) (August 2004): 795–801
Copyright 2004 American Agricultural Economics Association
796 August 2004 Amer. J. Agr. Econ.

the elasticity of output with respect to FDI important to understand the interrelationships
exceeds that of domestic capital investment, among FDI, trade, and economic growth. Since
which implies that FDI is the driving force in theory is unclear, this issue has been the sub-
the growth process. ject of empirical studies.
Borensztein, Gregorio, and Lee examine the
role of FDI in promoting economic growth us- Methodology and Data
ing an endogenous growth model. They ana-
lyzed FDI flows from industrial countries to Our econometric model is derived from a pro-
sixty-nine developing countries during 1970– duction function in which the level of a coun-
1989. Their results also show that FDI is try’s productivity depends on FDI, trade, do-

Downloaded from https://academic.oup.com/ajae/article-abstract/86/3/795/63091 by guest on 04 February 2020

an important vehicle of technology transfer, mestic investment, human capital, and initial
contributing more to economic growth than gross domestic product (GDP) per capita. The
domestic investment. They make a case for model is based on endogenous growth theory,
a minimum threshold stock of human capi- in the tradition of Balasubramanyam, Salisu,
tal necessary to absorb foreign technologies and Sapsford and Borensztein, Gregorio, and
efficiently. Lee, where FDI contributes to economic
Several other studies, including Feder, Ram, growth directly through new technologies and
and Salvatore and Hatcher, have analyzed other inputs as well as indirectly through im-
the export-led economic growth hypothesis. proving human capital, infrastructure, and in-
They argue that exports increase factor pro- stitutions (see Borensztein, Gregorio, and Lee
ductivity because of better utilization of ca- for the theoretical derivation of this model ap-
pacity and economies of scale. They also proach). To assess empirically the effects of
argue that exports are likely to alleviate FDI and trade on economic growth, we specify
foreign-exchange constraints and thereby fa- the following basic formulation:
cilitate importation of better technologies and (1) g = a + b1 FDI + b2 TRD + b3 HC
production methods. Grossman and Helpman
argue that open-trade regimes go hand-in- + b4 K + b5 G 0 + c1 FDI ∗ TRD
hand with good investment climates, technol-
+ c2 FDI ∗ HC + c3 FDI ∗ K
ogy externalities, and learning effects.
Empirical studies by Dollar, Sachs and + d1 IRT + d2 TX + d3 GC + e
Werner, and Lipsey generally support the view
that open economies grow faster. There are where g is the per capita GDP growth rate;
other studies, however, that question the wis- FDI, the foreign direct investment; TRD, the
dom of trade openness. Rodriguez and Rodrik, trade (exports plus imports) of goods and ser-
for example, present a critical view of the vices; HC, the stock of human capital; K, the
link between open-trade policy and economic domestic capital investment; G0 , the initial
growth. They argue that past studies fail to GDP (initial stock); IRT, the inflation rate; TX,
account for institutional differences among the tax on income, profits, and capital gains in
countries resulting in an upwardly biased es- the host country expressed as percentage of
timate of trade and other policy restrictions. current revenue; and GC is government con-
Their analysis shows that the relationship sumption. The variables FDI, TRD, K, GC are
between average tariff rates and economic measured as ratios to GDP. Our model extends
growth is only slightly negative and nowhere the work of Borensztein, Gregorio, and Lee to
near statistical significance. include the decade of the 1990s when FDI and
The question of whether FDI and trade trade grew rapidly in the developing countries.
trigger economic growth or the economic We also account for interaction of FDI with
development brings FDI and trade is an trade and domestic investment, in addition to
unresolved issue. Past studies either ana- human capital.
lyzed the impact of trade and FDI on eco- Past empirical studies have indicated that
nomic growth (Borensztein, Gregorio, and FDI, trade, human capital, and domestic in-
Lee; Balasubramanyam, Salisu, and Sapsford) vestment have a positive impact on economic
or analyzed the effects of economic growth on growth in developing countries. We expect the
FDI (Barrel and Pain, Lipsey). A positive ef- estimated coefficients for these variables to be
fect of FDI and trade on economic growth may positive. We also expect positive interactions
simply reflect the fact that FDI is attracted between FDI and trade and FDI and domes-
to countries that are expected to grow faster tic capital investment in promoting economic
and follow open-trade policies. It is, therefore, growth.
Makki and Somwaru Foreign Direct Investment and Trade 797

The stock of human capital in a host coun- Data for our analysis are obtained from
try is critical for absorbing foreign knowledge the World Development Indicators (WDI)
and an important determinant of whether po- database. The WDI database, published by
tential spillovers will be realized. We postulate the World Bank and International Monetary
not only a positive relationship between FDI Fund, includes variables such as GDP, per
and the GDP growth rate but also a positive capita income, GDP growth rates, FDI, trade
interaction between FDI and human capital in in goods and services, domestic capital invest-
advancing economic growth. The application ment, human capital, market openness, infla-
of advanced technologies embodied in FDI re- tion rate, tax income, and government con-
quires a sufficient level of human capital in host sumption. The data cover sixty-six countries

Downloaded from https://academic.oup.com/ajae/article-abstract/86/3/795/63091 by guest on 04 February 2020

countries. That is, the higher the level of human for the years 1960 through 2000. However, we
capital in a host country, the higher the effect limit our analysis to 1971 through 2000 because
of FDI on the country’s economic growth. the flow of FDI to most developing countries
One of the key questions regarding FDI began in 1970s. All variables represent the av-
and economic growth is: “What is the interac- erage over the following decades: 1971–1980,
tion between FDI and domestic investment”? 1981–1990, and 1991–2000.
The answer to this question seems to be less We estimate a system of three equations,
controversial in theory than in practice. As where the dependent variables are the mean
argued before, FDI is an important vehicle values of per capita GDP growth rates in
for the transfer of capital, technology, and each decade. We estimate the system of equa-
knowledge to host countries, thereby gener- tions using the seemingly unrelated regression
ating high-growth opportunities. In practice, (SUR) method as well as instrumental variable
however, the growth-enhancing impact of FDI (three-stage least squares, TSLS) approach.
depends critically on the absorptive capacity The SUR estimation allows for different er-
of a host country and whether FDI “crowds ror variances in each equation and for correla-
out” its domestic investment. Thus, an impor- tion of these errors across equations (Greene),
tant question to be addressed is: “What is the while the instrumental variable technique al-
extent to which FDI substitutes for or comple- lows us to overcome potential biases induced
ments domestic investment”? In our empirical by endogeneity problems between FDI and
model, we include FDI and domestic invest- economic growth.
ment separately as well as an interaction term
between FDI and domestic investment (FDI ∗
K). The interaction term estimates the com- Empirical Results
bined impact of FDI and domestic investment
on growth and indicates the nature of the re- The purpose of our empirical investigation
lationship between the two. A positive coef- is to analyze the effects of FDI and trade
ficient for the interaction term would suggest on economic growth and to examine how
that FDI and domestic investment (K) rein- FDI interacts with trade, human capital, and
force (complement) each other in advancing domestic investment in advancing economic
economic growth. growth in developing countries. We control
The initial GDP, measured in terms of con- for preexisting economic conditions by in-
stant U.S. dollars, controls for preexisting eco- cluding initial GDP as one of the explana-
nomic and institutional conditions in the host tory variables. We also account for differences
economy. We expect the initial GDP (ex- in macroeconomic policies and institutions in
pressed in logarithms) to be negatively re- the host countries by including variables, such
lated with GDP growth rates. The inflation as inflation rate, tax burden, and government
rate is a key indicator of fiscal and monetary consumption.
policies of a country. A lower inflation rate We test the effects of FDI and trade on eco-
should mean a better climate for investment, nomic growth in a framework of cross-country
trade, and, therefore, economic growth (Fisher equations utilizing data from sixty-six devel-
and Modigliani, Froot and Stein). Government oping countries over the last three decades—
consumption and tax on income, profits, and 1971–1980, 1981–1990, and 1991–2000. The
capital gains are proxies for institutions and system has three equations, where the depen-
infrastructure in the host countries. Since our dent variables are the per capita GDP growth
objective is to quantify the effects of FDI and rates (mean value) in each decade. We con-
trade on economic growth, we focus on devel- strained the model such that all three equa-
oping countries. tions yield the same coefficients in the
798 August 2004 Amer. J. Agr. Econ.

Table 1. Interrelationship between GDP Growth and FDI and Trade: Econometric Results
SUR Estimates
Dependent Variable:
Dependent Variable:
Per Capita GDP Growth Rate
Domestic Investment
Independent Variable 1.1 1.2 1.3 1.4
Intercept 0.5444 0.6232 1.0360a 0.7005a
(0.4844) (0.4654) (0.4293) (0.1757)
FDI 0.1856a 0.0475 0.0792 0.0716a
(0.0697) (0.0942) (0.0793) (0.0238)

Downloaded from https://academic.oup.com/ajae/article-abstract/86/3/795/63091 by guest on 04 February 2020

Trade 0.0126 0.0111 0.0059 0.0031
(0.0115) (0.0111) (0.0096) (0.0038)
Human capital 0.0089c 0.0086c 0.0095b −0.0020
(0.0055) (0.0053) (0.0047) (0.0023)
Domestic investments −0.0364 −0.0937 0.0419b
(0.0938) (0.0975) (0.2013)
Initial GDP per capita 0.0187 −0.0005 −0.0184 0.0087
(0.0213) (0.0195) (0.0167) (0.0081)
FDI ∗ Trade 0.0050a 0.0045a 0.0011c
(0.0017) (0.0013) (0.0007)
FDI ∗ Human capital −0.0035 0.0058 −0.0016
(0.0109) (0.0106) (0.0034)
FDI ∗ Domestic investment 0.0455c 0.0036
(0.0258) (0.0225)
Inflation rate −0.0047a −0.0001
(0.0009) (0.0003)
Tax rate −0.0040b −0.0001
(0.0021) (0.0007)
Government consumption −0.0574a 0.9018a
(0.1932) (0.0299)
System R2 0.0698 0.1175 0.2560 0.8903
a 99% confidence level.
b 95% confidence level.
c 90% confidence level.

three time periods with the exception of the Regression 1.1 reveals that FDI and trade
intercepts. have a positive impact on economic growth
Table 1 presents the econometric results after controlling for human capital, domestic
and compares alternative specifications. Re- investment, and initial income (table 1). The
gressions 1.1, 1.2, and 1.3, different variants estimated coefficient for FDI is positive and
of equation (1) above, are estimated using statistically significant while the estimated co-
the SUR method. Regression 1.1 is our ba- efficient for trade is not statistically signifi-
sic specification with explanatory variables of cant. Since the coefficient of FDI is larger than
FDI, trade, human capital, domestic invest- the coefficient of trade, it indicates the dif-
ment, and initial GDP. Regression 1.2 extends ferential impact of FDI in the host country’s
1.1 to include interaction of FDI with trade, economic growth. The coefficient for human
human capital, and domestic investment. Re- capital is positive, implying that human capi-
gression 1.3 (final specification) builds on re- tal contributes positively to economic growth
gression 1.2 by controlling for inflation rate, tax (significant only at a confidence level of 88%).
burden, and government consumption. Our re- The coefficients for domestic investment and
sults show that most coefficients have the ex- initial income are not statistically significant.
pected signs, particularly in specification 1.3. Including interactions between FDI and
Note that signs change for some coefficients trade, FDI and human capital, and FDI and
across specifications. The estimated R2 are domestic investment not only improves the
generally low but reasonable given the cross- overall performance of the estimation but also
sectional nature of the data used. allows us to capture their interaction effects
Makki and Somwaru Foreign Direct Investment and Trade 799

on economic growth (table 1). In regression icy variables—inflation rate, government con-
1.2, the interaction of FDI and trade yields a sumption, and tax on income, profits, and capi-
positive and statistically significant coefficient tal gains—are negative and statistically signifi-
while the effects of FDI and trade, by them- cant. This implies that lowering the inflation
selves, are positive but not statistically signifi- rate, tax burden, and government consump-
cant. Regression 1.2 also reveals that the FDI tion would promote economic growth. Lower
interacts positively with domestic investment inflation rates would indicate that the host
in advancing economic growth. The estimated country’s macroeconomic policies are stable
coefficient for domestic investment is posi- and disciplined. Lower tax burden would make
tive and statistically significant at a confidence the investments, both foreign and domestic,

Downloaded from https://academic.oup.com/ajae/article-abstract/86/3/795/63091 by guest on 04 February 2020

level of 90%. The estimated coefficients indi- more profitable. Decreasing the government
cate that host countries benefit positively both consumption would leave more money for
from FDI, itself, and through FDI’s positive in- investments.
teraction with trade and domestic investment.
The interaction between FDI and human FDI “Crowds-In” Domestic Investment
capital, although positive, is not statistically
significant. One of the important questions raised in the lit-
Regression 1.3 includes additional variables erature is whether FDI augments a host coun-
to control for macroeconomic policies and in- try’s capital investment or crowds out domestic
stitutional stability that could have a signifi- investment. Even though not statistically sig-
cant impact on FDI and trade and, thus, on nificant, the positive interaction between FDI
economic growth. Recent literature indicates and domestic investment in regression 1.3 im-
that FDI is greatly influenced by host coun- plies that domestic investment is unlikely to be
try policies, such as monetary, fiscal, and open- crowded out in developing countries.
market policies. We include inflation rates, tax To further strengthen our argument, we
income, and government consumption to be estimate the contribution of FDI to domes-
proxies for monetary and fiscal policies, as well tic investment after controlling for trade, hu-
as institutions in the host nations. Inclusion man capital, initial income levels, and various
of these policy variables significantly increases macroeconomic policy variables. Regression
the explanatory power of the estimated system 1.4 in table 1 presents the results of this es-
(table 1). timation using the SUR method. The results
The results of regression 1.3 reveal that FDI indicate FDI has a positive effect on domes-
and trade contribute positively to economic tic investment, as the estimated coefficient is
growth, but the estimated coefficients are not positive and statistically significant. This pos-
statistically significant (table 1). The stock of itive relationship implies that FDI stimulates
human capital and domestic investment, on or crowds-in domestic investment. This finding
the other hand, have positive and statistically is consistent with Borensztein, Gregorio, and
significant coefficients. The results also indi- Lee. Even though trade, by itself, is not sta-
cate that FDI positively interacts with trade, tistically significant, trade interacts positively
human capital, and domestic investment. But with FDI on domestic investment. The esti-
only FDI–trade interaction is statistically sig- mated coefficient for the FDI–trade interac-
nificant. This implies that FDI and trade com- tion term is positive and significant at the 90%
plement each other in advancing growth rate of confidence level.
income in developing countries. This result is
consistent with the idea that flow of advanced Endogeneity Problems
technology brought along by FDI can increase
the growth rate of the host economy by inter- The correlation between FDI and growth rate
acting with that country’s trade. could arise from an endogenous determination
The diverse experiences from developing of FDI. That is, FDI, itself, may be influenced
countries suggest that FDI and trade, by them- by innovations in the stochastic process gov-
selves, may not guarantee economic growth. erning growth rates (Borensztein, Gregorio,
A country’s economic growth is also affected and Lee). For example, market reforms in host
by its macroeconomic policies and institutional countries could increase both GDP growth
stability. Sound macroeconomic policies and rates and the inflow of FDI simultaneously. In
institutional stability are necessary precondi- this case, the presence of correlation between
tions for FDI-driven growth to materialize. FDI and the country-specific error term would
The estimated coefficients for the three pol- bias the estimated coefficients.
800 August 2004 Amer. J. Agr. Econ.

Table 2. Interrelationship between GDP Growth and FDI and Trade:

Instrumental Variable Estimation
TSLS Estimates
Dependent Variable:
Per-capita GDP Growth Rate
Independent variable 2.1 2.2 2.3
Intercept 0.5490 0.6271 0.9965b
(0.4873) (0.4690) (0.4321)
FDI 0.1806b 0.0427 0.0805

Downloaded from https://academic.oup.com/ajae/article-abstract/86/3/795/63091 by guest on 04 February 2020

(0.0698) (0.0946) (0.0794)
Trade 0.0115 0.0098 0.0066
(0.0116) (0.0112) (0.0096)
Human capital 0.0096 0.0093c 0.0094b
(0.0057) (0.0055) (0.0049)
Domestic investments −0.0361 −0.0912 0.0444b
(0.0942) (0.0980) (0.2025)
Initial GDP per capita 0.0195 0.0008 −0.0188
(0.0215) (0.0197) (0.0169)
FDI ∗ Trade 0.0049a 0.0043a
(0.0018) (0.0013)
FDI ∗ Human capital −0.0036 0.0065
(0.0110) (0.0105)
FDI ∗ Domestic investment 0.0456c 0.0027
(0.0259) (0.0225)
Inflation rate −0.0043a
(0.0009)
Tax rate −0.0042b
(0.0021)
Government consumption −0.0600a
(0.1950)
System R2 0.0704 0.1175 0.2421
a 99% confidence level.
b 95% confidence level.
c 90% confidence level.

The endogeneity problem is addressed Summary and Conclusions

by using the instrumental variables (see
Borensztein, Gregorio, and Lee for more de- This paper analyzes the role of FDI and trade
tails). One of the major problems with the in- in economic growth of developing countries
strumental variable estimation method is the within the endogenous growth-theory frame-
difficulty in identifying instruments that are work. Using cross-section data relating to a
highly correlated with FDI (or trade) but not sample of sixty-six developing counties over
with the error term. We use lagged values of three decades, we show that FDI and trade
FDI, lagged values of trade, and log value of contribute toward advancing economic growth
total GDP as instruments in a TSLS method. in developing countries. There is a strong, pos-
The results of the TSLS model (reported in itive interaction between FDI and trade. FDI
table 2, regressions 2.1–2.3) show that the in- is often the main channel through which ad-
strumental variable estimation yields qualita- vanced technology is transferred to develop-
tively similar results as those obtained by the ing countries. Our results imply that the ben-
SUR method. The estimated coefficients on efits from such investment would be greatly
FDI and trade, by themselves, are positive but enhanced if the host country has a better stock
statistically insignificant. The interactive term of human capital. We also show that FDI stim-
of FDI and trade is positive and statistically sig- ulates domestic investment. Sound macroe-
nificant. This alternative estimation also sug- conomic policies and institutional stability
gests that our results are robust. are necessary preconditions for FDI-driven
Makki and Somwaru Foreign Direct Investment and Trade 801

growth to materialize. Our results imply that tion.” Weltwirtschaftliches Archiv 4(1978):810–
lowering the inflation rate, tax burden, and 33.
government consumption would advance eco- Frankel, J.A., and D. Romer. “Does Trade
nomic growth in developing countries. Cause Growth?” American Economic Review
89(1999):379–99.
Froot, K.A., and J.C. Stein. “Exchange Rates and
References Foreign Direct Investment: An Imperfect Cap-
ital Markets Approach.” Quarterly Journal of
Balasubramanyam, V.N., M.A. Salisu, and D. Economics 106(1991):1191–217.
Sapsford. “Foreign Direct Investment and Greene, W. Econometric Analysis. New York:

Downloaded from https://academic.oup.com/ajae/article-abstract/86/3/795/63091 by guest on 04 February 2020

Growth in EP and IS Countries.” Economic Macmillan Publishing Company, 1990.
Journal 106(1996):92–105. Grossman, G.M., and E. Helpman. Innovation and
Barrell, R., and N. Pain. “An Econometric Anal- Growth in the Global Economy. Cambridge:
ysis of U.S. Foreign Direct Investment.” Re- The MIT Press, 1991.
view of Economics and Statistics 78(1996):200– Lipsey, R. “Home and Host Country Effects of
7. FDI.” Working Paper, National Bureau of Eco-
Borensztein, E., J. De Gregorio, and J. Lee. nomic Research, Cambridge, MA, 2002.
“How Does Foreign Direct Investment Affect Ram, R. “Exports and Economic Growth: Sine Ad-
Economic Growth.” Journal of International ditional Evidence.” Economic Development
Economics 45(1998):115–35. and Cultural Change 33(1985):415–52.
De Mello, L.R. “Foreign Direct Investments in Rodriguez, F., and D. Rodrik. “Trade Policy and
Developing Countries and Growth: A Selec- Economic Growth: A Skeptic’s Guide to
tive Survey.” Journal of Development Studies the Cross-Country Evidence.” Working Pa-
34(1997):115–35. per, National Bureau of Economic Research,
Dollar, D. “Outward Oriented Developing Cambridge, MA, 1999.
Economies Really Do Grow More Rapidly: Sachs, J., and A. Werner. “Economic Reform and
Evidence from 95 LDCs, 1976–85.” Economic the Process of Global Integration.” Brookings
Development and Cultural Change 40(1992): Papers on Economic Activity 1(1995):1–118.
523–44. Salvatore, D., and T. Hatcher. “Inward Oriented
Feder, G. “On Exports and Economic Growth.” and Outward Oriented Trade Strategies.” Jour-
Journal of Development Economics 12(1983): nal of Developmental Studies 27(1991):7–25.
59–73. Solow, R.M. “Technical Change and the Aggregate
Fisher, S., and F. Modigliani. “Towards an Under- Production.” Review of Economics and Statis-
standing of the Real Effects and Costs of Infla- tics 39(1957):312–20.