Form for thesis proposal
Your name: Menghan Wang
Your student number: 10180109
Specialization: Economics and Finance
Field: Macroeconomics
Number of credits thesis: 12
Title of your research proposal: the effect of Foreign Direct Investment on the economic growth
Abstract
This paper examines whether the foreign direct investment has a significant positive effect on the economic growth by using the cross- countries data for 78 countries over the period of 1984-2004. The empirical results imply that FDI inflows can promote the economic growth based on the total sample and the subsample of the developed countries. However, these results do not suggest a significant positive correlation between FDI inflows and the economic growth for the developing countries. Furthermore, these results stress that the absorptive capacity is significant important for the developing countries to obtain the positive effects of the FDI inflows.
I. Introduction
According to UNCTAD (2001, p. xiii), foreign direct investment (FDI) inflows grew by 18 per cent in 2000, which increased faster than other economic aggregates like world production, capital formation and trade. The FDI inflows, enlarging the role of international production in the world economy, reached a record of $1.3 trillion in 2000 and continued to expand rapidly. This increasing role of FDI has been synchronous with a shift in emphasis among policymakers in developing countries that they put more efforts to attract more FDI inflows, especially after the 1980s debt crisis and the recent turmoil in emerging economies (Alfaro et al., 2004, p. 90). The essential reason for the policymakers to do so is they have beliefs that FDI inflows have several positive effects on the local economic growth, including productivity gains, technology spillovers, the introduction of new processes, managerial skills and know-how in the domestic market, employee training, and international production networks and access to markets (Alfaro et al., 2004, p. 90).
Although these beliefs provide the basis for most of the empirical works on the effect of FDI inflows, the empirical evidences for whether FDI inflows can generate positive effects on the economic growth is ambiguous at both micro and macro levels (Alfaro et al., 2009, p. 111). At the macro level, Borensztein, De Gregorio and Lee (1998, p. 115) provide little evidence that FDI inflows have an exogenous positive effect on economic growth and they further suggest that FDI inflows may contribute to the economic growth only when a sufficient
absorptive capability of the advanced technologies is available in the host economy. While, a recent study which includes both developed and developing countries finds no significant effect of FDI inflows on the economic growth and the empirical results show that the FDI inflows only have irregularly s ignificant effects on the economic growth in a five year periods (Carkovic and Levine, 2002, p. 13).
Empirical evidence at the micro level is not straightforward as well. Javorcik (2004, p. 605) cites case study evidence of Lithuania in his paper, and the evidences provided by him are consistent with the assumption that positive productivity spillovers from FDI inflows take place through the contacts between foreign affiliates and their local suppliers in upstream sectors. On the other hand, Görg and Greenaway (2004, p. 186), reviewing the micro evidence on externalities from foreign-owned to domestically-owned firms, conclude that much of the work fails to find positive horizontal spillovers and some studies even report negative effects of multinational presence on domestic productivity.
This paper examines whether FDI inflows affect the economic growth for the host country at the macro level. The test is taken from following aspects. Firstly, this paper use a more recent time period (1984-2004) for a large cross-country (78 countries). The primary reason to choose this time period is that majority countries had a boost of FDI inflows after the 1980s debt crisis (Alfaro et al., 2004, p. 90). Moreover, data for this period shows that there exists a more synchronized trend of growth rate between FDI inflows and the economy (UNCTAD, 2002, p. 4). Therefore, as most literatures focus on the correlation between them over the period of 1970-1990, the recent data will gives a more clear and significant relationship between them. Additionally, due to the economic globalization and integration, the recent data may eliminate some unfavorable factors for the test, such as the missing data for data set, the changes of national trade policies and the development of local productivities. As a result, it will get a more unambiguous and practical results. Secondly, it accepts the assumption that there exists an endogenous relationship between FDI inflows and the economic growth which is actively tested by a large of previous papers (Li & Liu, 2004, p. 394), so that it will use the instrumental variables regression (IV estimation) instead of the ordinary least square regressions (OLS). Thirdly, it will examine the effects separately for the developed and developing countries as the effect will be influenced by the different endowments of these
countries. Finally, the interaction terms of FDI inflows with human capital, infrastructure, and the technology gap are introduced to examine whether FDI inflows affect the economic growth by itself directly or through the interaction terms indirectly (Li & Liu, 2004, p. 394). The Section II offers a literature review on the previous findings of the relationship between the FDI inflows and the economic growth. The Section III provides an introduction to the methodology and data used in the empirical study. The Section IV presents the regression results and interprets the meaning of them. The Section V draws the conclusion and lists the limitation of this paper.
II. Literature Review
FDI inflows have grown dramatically in the past 20 years, exceeding the growth rate of world production and the growth rate of international trade. Although most FDI flows are within the developed countries, it has become increasingly one of the significantly important economic sources for many developing countries (Herzer, Klasen & Nowak-Lehmann, 2008, p. 793). The main reason for this situation is that many policy makers believe that FDI inflows could bring capital to the host country and help create jobs. Also, they view the FDI inflows as an important source of advanced technologies and management transfer to the host economy (Gao, 2005, p. 158).
The theoretical foundation for empirical studies on FDI inflows and the economic growth derives from either neo-classical models of growth or endogenous growth models. As Cipollina et al. (2012, p. 1601) cited, the standard neoclassical Solow growth model suggests that FDI inflows increase the capital stock, and thus stimulate the economic growth for the host economy simply by the capital accumulation. In the neoclassical growth model, however, because of the diminishing returns to capital, FDI inflows have only a short-run effect on the economic growth, and the effect may decrease after each country reaches its new steady state. In the endogenous growth model, FDI inflows can have a short-effect on growth but, more interestingly, they can also have a permanent effect through the technology transfer, diffusion, and spillover effects (Cipollina et al., 2012, p. 1604).
knowledge created in the developed countries with their relatively high endowments of human capital can be transferred to the developing countries through FDI inflows to them. Also, using the cross-country data for a sample of 46 developing countries, they provide evidence that the trade openness is a crucial factor for the host countries to acquire the potential positive impact of FDI inflows. Also, the authors interpret the positive correlation between them as the more open the economies, the higher volume of FDI inflows the countries can attract, and therefore, the more efficient utilization the countries can promote (Balasubramanyam, Salisu & Sapsford (1996, p. 97). Moreover, the empirical results indicate that FDI inflows have a stronger effect on the economic growth than the domestic investment has, and the FDI inflows can be recognized and act as an important vehicle for international technology transformation.
On the other hand, Borensztein, Gregorio and Lee (1998, p. 117) examine the effect of FDI inflows on the economic growth by using a framework of cross-country regressions and a data set on FDI inflows from 69 developing countries over the period of 1970-1990. Their results confirm the findings of Balasubramanyam, Salisu & Sapsford and also suggest that FDI inflow is an important vehicle for the transfer of technology spillovers and it can contributes to the economic growth in a larger measure than domestic investment. Moreover, their empirical results imply that the effect of FDI inflows on the growth depends on the level of human capital in the host country, in other words, the FDI inflows have a positive growth effect on the economy only if the level of education is higher than a given threshold (Borensztein, Gregorio & Lee 1998, p. 119).
Similarly, Alfaro et al. (2004, p. 107) examine the links between FDI inflows, financial markets and economic growth by using cross-country data from 71 developing and developed countries. Their empirical evidences suggest that FDI inflows play an important role in contributing to the economic growth, however, the empirical results also address that the level of development of the local financial markets is crucial for these positive effects to be realized, which has not been shown in the previous papers. Additionally, Alfaro et al. (2004, p. 107) provide evidence that the link between FDI inflows and growth is causal, where FDI inflows promote the economic growth indirectly through the financial markets.
identify a direct, unmitigated positive relationship between FDI inflows and the economic growth, but instead, he suggests that the effects of FDI inflows are contingent on the ‘absorptive capability’ of host countries, with particular respect to financial or institutional development.
Zhang (2001, p. 175), using data from 11 countries in Asia and Latin America, presents that FDI inflows are expected to stimulate the economic growth for the host countries. However, the empirical results also address that whether the FDI inflows can stimulate the economic growth mainly depends on the specific characteristics of each country. Further, he demonstrates that FDI inflows tend to be more likely to promote the economic growth when host countries adopt a liberalize trade regime, improve the education and thereby human capital conditions, encourage export-oriented FDI, and maintain macroeconomic stability The above discussion shows that the impact of FDI on economic growth is far from conclusive. In accordance with Li and Liu (2004, p. 395), the role of FDI inflows can be positive, negative, or insignificant, depending on the economic, institutional, and technological conditions in the recipient economy.
Because the important role of financial market development in the transmission of positive externalities of FDI inflows on the host country, Azman-Saini, Law and Ahmada (2010, pp. 211-212) present new evidence and re-examine the impact of FDI on growth through a different approach, using data from 91 countries over the period 1975–2005. Their empirical results confirm the findings of Alfaro et al that the positive effect of FDI inflows on the economic growth ‘kick in’ only after the financial markets development exceeds a threshold level. This finding stress the important role of the local government to emphasize on diffusion aspect in formulating policies to attract FDI inflows, as the knowledge diffusion is not sustained on welfare ground. Therefore, policies which tend to directly attract the FDI inflows should go hand in hand with, not precede, policies that aims at promoting the local financial market developments (Azman-Saini, Law & Ahmada, 2010, p. 213).
Meanwhile, an earlier survey of 11 studies by de Mello (1997, p. 8) find that majority of the literatures present a positive effect of FDI inflows on the economic growth, and they also show that the stronger effects of the FDI inflows are always associated with a greater openness or export-promotion policies or with a higher level of development in the local
economy.
On the contrary, Choe (2003, p. 44) finds that the strong positive association between the economic growth and FDI inflows do not necessarily mean that a high FDI inflow can lead to a rapid growth of the economy. He tests the causal relationships between the economic growth and FDI inflows by using a data set which consists of 80 countries from 1970 to 1995. By using a panel VAR model, he gets the empirical results that the FDI inflows can cause the economic growth, and vice versa; however, the effects from the economic growth on FDI inflows are rather more significant than from FDI inflows to the economic growth.
In addition, Li and Liu (2004, p. 393) use the data from 84 countries over the period of 1970-1999 to investigate whether the FDI inflows affect the economic growth. Not just simply assumed, they test the endogeneity between FDI inflows and the economic growth and then find a significant endogenous relationship between them in the period after the mid-1980s. Furthermore, they imply that FDI not only directly promotes the economic growth by itself but also indirectly does so through its interaction terms. They stress that the interaction term of FDI inflows with human capital exerts a strong positive effect on the economic growth in developing countries, while the interaction term of FDI inflows with the technology gap has a significant negative impact on the economy.
The above literature review suggests that the impact of FDI inflows on the economic growth remains extremely controversial. This is partly due to the use of different samples by different authors and partly due to various methodological problems (Li & Liu, 2004, p. 396).
III. Methodology and Data
The data set covers 78 countries over the period of 1984-2004, including 22 developed and 56 developing countries. The reason for choosing this specific time period is that most of the FDI inflows increased dramatically after the debt crisis in 1980s and the analysis wants to eliminate the negative effect of GDP growth from the newly global financial crisis.
Following the contributions of Li and Liu (2005, p. 396) and others to the development of the new growth theory, the core explanatory variables for the economic growth include initial per capital GDP, population growth, domestic investment, foreign direct investment (FDI) and
human capital. Also, the equation includes a group of other variables, X, which may also affect the economic growth (Li & Liu, 2005, p. 396).
The basic specification for the model is therefore
gi,t = ß0+ß1*In yi,84 + ß2*POPi,t + ß3*SCHi,84 +ß4*INVi,t + ß5*FDli,t + ß6*Xi,t +Ɛi.t (1) where gi,t is the real GDP per capita growth of country I which represents the economic growth, yi,84 is the real GDP per capita in 1984 which represents the initial economic level of country I, POPi,t is the population growth of the country I, SCHi,t is the level of secondary school attainment in 1984 which is used to be the proxy for the initial level of human capital in the host economy, INVi, t represents the domestic investment level and is measured as a ratio of gross domestic investment to GDP, and FDIi,t represents the foreign direct investment level and is measured as a ratio of FDI inflows to GDP.
The group of other variables X consists of the country-group dummies and other variables that are frequently used as the determinants of the economic growth in cross-country analysis. To construct the country-group dummies, the analysis uses the developed countries as the base group, and then creates other four dummy groups for the developing countries, including Latin American, African, fast growing and other developing countries. In other words, if a country is from the African continent, then the African dummy variable equals to 1 for this country, and the other dummy variables equal to 0.
Moreover, the group of X also includes the infrastructure variables and technology gap variables. The infrastructure variable is measured as the telephone lines per capita for the test, and the technology gap is measured by the following equation as:
GAPi,t = (Ymax-Yi,t)/Yi,t, (2) where Ymax is the GDP per capita of the United State which represents the highest level of productivity in the world.
Additionally, in order to examine the absorptive capacity of foreign technology through the FDI inflows, the analysis uses the interaction term of FDI with human capital, the technology gap and infrastructures. All these interaction terms are included in the group of variables, X, therefore the analysis can find out whether the FDI inflows can affect the economic growth indirectly through these factors. Furthermore, the significance level of these interaction terms
may indicate whether there exists a minimum level of human capital, infrastructure or the technology gap for the host country to be beneficial from the FDI inflows.
Except for the human capital variables, the rest of variables can obtain the data from the World Bank Database. Data on human capital is taken from Barro-Lee 2000 and it is the percentage of secondary school attainment of the total population aged 15 and over in the year 1984. It should be noted that the data on schooling is only available on average of every 5 years. However, because the data evolves very slowly, it is fairly reasonable to assume that the secondary school attainment in 1984 could be replaced with data of the attainment level in 1985. Also, the list for the countrygroup dummy variables is provided in the Appendix A. In addition, it should be noticed that the model presented above may be subject to
endogeneity, which has been tested by Li and Liu (2005, p. 397) and means that the FDI inflows can promote the economic growth for the host countries and then it will in turn attract more FDI inflows for the country (Li & Liu, 2004, p. 394).The problem may stem from the simultaneous causality, omitted variables or errors in variables, and therefore, it may results in an inconsistent in the OLS estimators and causes the biased estimated coefficients for the analysis (Stock &Watson, 2011, p. 420). To solve this problem, the analysis will use the instrumental variables estimation (IV estimation) and construct the instruments to isolate the FDI variable from the error term of the model. However, one of the major problems with the IV estimation method is the difficulty in identifying the appropriate instruments which are correlated with FDI variable but are uncorrelated with the error term.
The analysis tries to control for the endogeneity problem by using the one-period lagged value of FDI as the instruments. The one-period lagged value of FDI at time t is the value of FDI at time t-1. As the value of the FDI is always predicted and affected by the past, it is reasonable to assume that the one-period lagged FDI is correlated with the FDI variables. Moreover, because of the endogeneity problem, the FDI variables is correlated with the GDP growth at time t and the error tem ut , and also the one-period lagged FDI variables is
correlated with the GDP growth at time t-1and the error term ut-1. As a result, it can assume that one-period lagged FDI variables can be uncorrelated with error term ut as it can affect the GDP growth at time t only indirectly through the FDI variables. These two assumptions are confirmed by the correlation matrix below
FDI Lagged FDI Residual FDI 1.0000 Lagged FDI 0.6467* 1.0000 0.0000 Residual 0.0000 0.0156 1.0000 1.0000 0.5279
From the matrix above, it shows a significant positive correlation between FDI variables and one-period lagged FDI variables and an insignificant correlation between one-period lagged FDI variables and the residuals which, therefore, cannot reject the hypothesis that the one-period lagged FDI is uncorrelated with the residual.
Also, according to the results of the relevance test shown below, the F-value of this test is larger than 10, as a result, it can eliminate the weak instruments problems for the analysis.
Number of obs 1638
Source SS df MS F( 1, 1636) 1176.07
Model 9343.96924 1 9343.96924 Prob > F 0.0000
Residual 12998.1199 1636 7.94506104 R-squared 0.4182
Total 22342.0891 1637 13.6481913 Adj R-squared 0.4179
Root MSE 2.8187
FDI Coef. Std. Err. t P>t [95% Conf. Interval] Lagged
FDI
0.6173674 0.0180022 34.29 0 0.5820575 0.6526773
_cons 0.783697 0.0829767 9.44 0 0.6209453 0.9464486
In conclusion, the analysis may use the one-period lagged FDI as the instrument for the IV estimation and get the reduced form for FDI as
FDIi,t=a0+a1*FDIi,t-1+ut (3) where FDIi,t-1 represents the one-period lagged value of FDI.
The analysis may firstly estimate the coefficients from the reduced form (3) by OLS to get the predict value of FDIi,t , and then regress gi,t on the predict value of FDIi,t which is calculated from the first stage, and the rest regressors in formula (1) by OLS to estimate the coefficients (Stock & Watson, 2011, p. 432). The analysis may avoid the endogeneity problem as it drops the problematic component of the FDI variables in the first stage and then estimates the coefficient for FDI variables by only using the problem-free part which is uncorrelated with the error termbut correlated with the dependent variables in equation (1).
IV. Empirical Results
Section IV presents the estimation results for this analysis. This section firstly demonstrates and discusses the results of single equations for all countries and then compares the results for the subsamples of developed and developing countries.
(a) The results of single equations for all countries in the sample (Table 1)
Column 1 shows the results from the basic model with the main core variables and all of them are statistically significant at 5% level. This column shows that the estimation coefficient on Log (initial GDP) is statistically negative, while the coefficients on the domestic investment and secondary school attainment are significant positive. In other words, the empirical results suggest that the growth in the GDP per capita is positively correlated with a low initial level of GDP, a high initial level of secondary school attainment as well as a high domestic investment ratio. According to Wang (2009, p. 996), the significant negative effect of initial level of GDP on the economic growth supports the assumption that the economic growth is conditional convergent. In addition, as the coefficient on population growth variable is significant negative, it indicates that the growth of population will have a negative effect on the economic growth.
From the Column 2, the equation includes the FDI inflows variables and then shows that the coefficient on FDI inflows is statistically significant positive, indicating that the FDI inflows can significantly contribute to the economic growth. Also, the coefficient implies that a 1 percent increase in the FDI inflows will enhance the economy by around a 0.106 percent among all the countries. However, compared to the coefficient on the domestic investment,
Table 1
which implies that a 1 percent increases in the domestic investment inflows will enhance the economy by a 0.132 percent, the FDI inflows may not stimulate the economy as much as predicted by the previous literatures. In contrast to what Borensztein, Gregorio & Lee (1998, p. 115) have found in their earlier papers, these empirical results imply that the domestic investment contributes more to the economic growth than the FDI inflow does.
After adding the country- group dummy variables, the column shows that all coefficients on the dummy variables are at 5% significance level. The empirical results present that the coefficients on the Africa dummy, Latin American dummy and other developing dummy are negative while that on the fast developing dummy is positive. The significance of all these dummy variables indicates that the location and the local economic financial market may have a significant effect on the GDP growth for the host countries.
(1) (2) (3) (4) (5) (6) (7) (8)
GDP Growth GDP GrowthGDP GrowthGDP GrowthGDP GrowthGDP GrowthGDP GrowthGDP Growth
Log(initial -0.971*** -0.985*** -1.232*** -1.062** -1.615*** -1.615*** -1.587** -1.549** GDP) (-3.80) (-3.87) (-3.79) (-2.81) (-3.34) (-3.35) (-3.28) (-3.20) Population -0.310** -0.303** -0.230* -0.242* -0.213* -0.213 -0.218* -0.210 Growth (-3.13) (-3.07) (-2.19) (-2.29) (-1.99) (-1.96) (-2.02) (-1.95) Domestic 0.148*** 0.132*** 0.108*** 0.106*** 0.103*** 0.103*** 0.108*** 0.107*** Investment (12.15) (9.91) (7.63) (7.21) (6.99) (4.96) (6.33) (6.35) Human 0.0312*** 0.0307*** 0.0152 0.0168 0.0166 0.0168 0.0218 0.0154 Capital (3.43) (3.39) (1.61) (1.75) (1.73) (1.39) (1.39) (0.85) FDI 0.106** 0.103* 0.113* 0.114* 0.117 0.250 0.232 (2.68) (2.52) (2.54) (2.58) (0.77) (0.86) (0.79) Africa -1.193* -1.394* -1.482** -1.481** -1.466** -1.434** (-2.45) (-2.54) (-2.69) (-2.73) (-2.72) (-2.65) Latin -1.086** -1.325** -1.555** -1.553*** -1.587*** -1.581*** America (-3.05) (-2.87) (-3.25) (-3.37) (-3.39) (-3.37) Fast Growing 1.357** 1.268** 1.081* 1.087* 0.949* 1.022* (2.95) (2.67) (2.23) (2.16) (1.96) (2.09) Other -0.888* -1.084* -1.255** -1.251** -1.343** -1.304** developing (-2.25) (-2.36) (-2.68) (-2.71) (-2.88) (-2.78) Infrastructure 0.0105 0.00732 0.00722 0.00920 0.00291 (0.85) (0.59) (0.61) (0.76) (0.21) Technology -0.00799 -0.00799 -0.00399 -0.00350 Gap (-1.83) (-1.83) (-0.65) (-0.58) FDI*Human -0.0000977 -0.00263 0.000458 Capital (-0.03) (-0.41) (0.06) FDI*Technology -0.00225 -0.00245 Gap (-0.92) (-1.04) FDI* -0.00281 Infrastructure (-1.24) _cons 1.502 1.641* 3.917** 3.728** 5.807*** 5.806*** 5.411** 5.314** (1.86) (2.03) (3.30) (3.10) (3.52) (3.52) (3.20) (3.16) N 1638 1638 1638 1638 1638 1638 1638 1638 t statistics in parentheses * p<0.05 ** p<0.01 *** p<0.001"
According to the Column 4, the analysis includes the infrastructure variable, which is defined as the telephone lines per capita for every country. Although the sign of this coefficient indicates a positive influence for the economic growth, it is statistically insignificant in this test. The primary reason for the insignificance is that many countries, especially the
developing countries, do not have shortages in the number of infrastructures, but they need to improve the efficiency of the usage of them (Li & Liu, 2005, p. 400). Therefore, even though some previous theories predict that the infrastructure should have a significant positive contribution to the economic growth; it is still reasonable to believe that this result is valid and it suggests that the infrastructure itself may have no effect on the economic growth. In Column 5, the technology gap enters the analysis and has a negative effect on the economic growth. The negative sign of the coefficient implies that the technology gap has a negative effect on the economic growth and can slow the local economy. However, rather than 5%, the result is significant at 10% level, indicating that the negative effect from the technology gap may not significant enough to be considered by the policymakers as the main reason for the economic differences.
To be more specific to investigate the effects of FDI inflows on the economic growth for the host countries, this analysis may examine the indirect effects of FDI inflows by adding the interaction terms of FDI inflows in the basic equation. According to Durham (2004, p. 285), the interaction term could be considered as the proxy for the absorptive capacity of the country, and it includes the interaction of FDI inflows with secondary school attainment, technology gap and infrastructure. Therefore, the analysis may test whether the minimum level of human capital, infrastructure and technology gap is significantly conditional for the host countries to gain the positive externalities from the FDI inflows.
After the involvement of the interaction terms with FDI inflows and human capital, infrastructure and technology gap, the coefficients on the FDI inflows become statistically insignificant. This insignificance can be explained by that the positive effect of FDI inflows on the economic growth may exist if the host countries meet some requirements.
The coefficient on the interaction term of FDI with secondary school attainment is
respectively negative and positive in the Column 6, Column7 and Column 8. However, all the coefficients are not significant at 5% level. This situation suggests that the FDI inflows may
increase the economy in the host country even though the initial level of human capital is low. Furthermore, it indicates that there is no significant correlation between the positive effect of FDI inflows on the economic growth and the initial human capital level. In other words, the empirical results deny the assumption that the high level of human capital could enhance the ability of the host countries to absorb technological spillovers through the FDI inflows and the finding that the FDI inflows may have a positive effect on the economic growth if there is a minimum level of human capital, which are confirmed by Borensztein, Gregorio & Lee (1998, p. 115).
The coefficients on the interaction term of FDI with the technology gap are both statistically insignificant negative in Column 7 and Column 8, suggesting that the economic growth is unrelated with the transmission of technological spillovers through the FDI inflows for the host countries. This finding is consistent with some previous papers that the positive effect of FDI inflows does not depend on the local technology gap.
In the last column, the interaction term of FDI inflow with the infrastructure is also included. As the coefficient on this term is insignificant negative, it supports the finding that there is no correlation between the positive externalities of the FDI inflows and the infrastructure. Because the coefficients on the interactions terms are all insignificant, they suggest that the positive effect of FDI inflows on the economic growth may not be significantly correlated by their absorptive capacities for the total sample. More specific, the insignificant correlation may suggest that there already exists a minimum level of human capital, infrastructures and technology gap in the host countries to make they benefit from the FDI inflows (Li & Liu, 2005, p. 402), therefore, the effect of absorptive capacities may not be such significant for the host countries.
(b) The comparison of developed and developing countries (Table 2 and Table 3) From Table 2 to Table 3, the analysis uses the subsamples of developed and developing countries to examine and compare the roles of FDI inflows in these two different groups. Column 1 in both tables presents the results from the basic equation and these estimation results are similar to those are found in part (a). The initial level of GDP and population growth still negatively affect the economic growth, however, the coefficient on the former is statistically insignificant in the developed countries while it is significant in the developing
countries, and the latter variable is the other way around. The insignificant of the estimate coefficient on the population growth for the developing countries indicates that the
developing countries may be benefit from population growth as they may be benefit from the labor-intensive industries. In addition, it is interesting to find that the coefficient on the secondary school attainment is statistically insignificant negative in the developed countries while it is statistically significant positive in the developed countries. The primary reason may be that the percentage of secondary school attainment is already very high in the developed Table 2
countries, and it changes slowly during these years. Therefore, the initial human capital level may not have a significant effect on the economic growth for the developed countries and the correlation between them is not strong in this test (Li & Liu, 2005, p. 402). At the same time,
(1) (2) (3) (4) (5)
GDP Growth GDP Growth GDP Growth GDP Growth GDP Growth
Log(initial -0.811 0.0541 4.059*** 3.459** 5.832 GDP) (-1.25) (0.08) (3.83) (3.03) (1.89) Population -0.365* -0.541** -0.558** -0.537** -1.281* Growth (-2.25) (-3.24) (-3.23) (-3.11) (-2.35) Domestic 0.116*** 0.143*** 0.103** 0.101** 0.164* Investment (3.96) (4.75) (3.27) (3.21) (2.41) Human -0.00389 -0.0132 -0.0265* -0.0286* 0.141 Capital (-0.38) (-1.26) (-2.35) (-2.51) (1.89) FDI 0.369*** 0.522*** 0.515*** 9.931* (5.68) (6.34) (6.31) (2.17) Infrastructure 0.0886*** 0.0922*** 0.112 (5.40) (5.48) (1.32) Technology -0.273 4.010 Gap (-1.19) (1.96) FDI*Human 0.0133 Capital (1.41) FDI*Technology 0.000302 Gap (0.12) FDI* -0.00865** Infrastructure (-2.79) _cons 3.265 -0.967 -11.24** -8.403* -38.58* (1.34) (-0.37) (-3.26) (-2.07) (-1.98) N 462 462 462 462 462 t statistics in parentheses * p<0.05 ** p<0.01 *** p<0.001"
due to the low initial level of human capital in the developing countries, it still has the significant positive impact of the economic growth.
In Table 2 and 3, the coefficients on FDI inflows in both developed and developing countries are positive. The coefficient of FDI inflows is 0.369 for the developed countries, which is larger than what is found in the developing countries. In other words, for the same amount of FDI inflows, the developed countries could enhance more for the local economy than the developing countries. However, it only implies a statistically significant positive effect on the Table 3
economic growth for the developed countries and shows insignificant positive correlation between them for the developing countries. This finding partly confirms what has been found in part (a) and concludes that the FDI would stimulate the economy only for the developed countries. At the same time, these results verify the findings of Alfaro et al. (2010, p. 242)
(1) (2) (3) (5) (6)
GDP Growth GDP Growth GDP Growth GDP Growth GDP Growth Log(initial -1.293*** -1.341*** -1.448** -2.035*** -1.925** GDP) (-3.67) (-3.77) (-3.17) (-3.45) (-3.23) Population -0.183 -0.172 -0.158 -0.133 -0.113 Growth (-1.43) (-1.34) (-1.18) (-0.99) (-0.84) Domestic 0.148*** 0.141*** 0.120*** 0.116*** 0.120*** Investment (10.40) (8.89) (6.85) (6.57) (6.10) Human 0.0527*** 0.0522*** 0.0280* 0.0252 -0.00476 Capital (4.07) (4.04) (1.98) (1.76) (-0.19) FDI 0.0464 0.0305 0.0370 -0.163 (0.94) (0.57) (0.68) (-0.50) Other -2.084*** -2.109*** -2.139*** developing (-3.99) (-4.04) (-3.96) Infrastructure 0.00985 0.0129 0.0344 (0.55) (0.71) (1.63) Technology -0.00818 -0.00853 Gap (-1.57) (-1.21) FDI*Human -0.0889* Capital (-2.21) FDI*Technology -2.114* Gap (-2.14) FDI* -0.0812* Infrastructure (-1.98) _cons 1.733 1.903 5.102*** 7.227*** 7.262*** (1.68) (1.82) (3.83) (3.81) (3.71) N 1176 1176 1176 1176 1176 t statistics in parentheses * p<0.05 ** p<0.01 *** p<0.001"
that the an increase in the FDI inflows leads to a higher additional growth under developed countries which has a well-developed financial market relative to the developing countries which has a under-developed financial market.
From the data shown in the Column 3, the coefficient of infrastructure in the developed countries is statistically significant positive, while it is insignificant positive in the developing countries. Be similar as what has been explained in the part (a), many developing countries may not lack of infrastructures, but lack of efficient usage of it. At the same time, because the developed countries use the infrastructures more efficiently, the increased number of
infrastructures may have a positive effect on the economic growth, and therefore, the infrastructure has a significant positive coefficient with the economic growth.
As also shown in the Column 4, the technology gap has insignificant negative influence on the economic growth in both developing and developed countries. However, even though the coefficient for the developing countries is not at 5% significance level, it still has a higher statistically significance level than the developed countries. Therefore, compared to the developed countries, the technology gap may generate more significant negative effects on the economic growth for the developing countries. As the policy makers believe that they can minimize the technology gap by introducing the FDI inflows, the larger negative effect of the technology gap for the developing countries may explain why the developing countries rather than the developed countries are more interested in attracting the FDI inflows.
Interestingly, in the last column of the table, the interaction terms of FDI with secondary school attainment, technology gap are both statistically insignificant for the developed countries. The insignificance indicates that the human capital and technology gap have no correlation with the positive externalities of the FDI inflows. Moreover, it shows that the developed countries can benefit from the positive effect on the economic growth without any threshold, as the initial level is already high in the developed countries. However, the
interaction terms of FDI inflows with infrastructures are still significant negative, indicating that the infrastructure has a significant role in obtaining the positive externalities of the FDI inflows for the host countries. In addition, as shown in the last column, the coefficient on the FDI inflows is significant positive and it indicates that a 1% increase in the FDI inflows can directly generate a 9.95% increase in the economic growth.
At the same time, as confirmed by Borensztein, Gregorio & Lee (1998, p. 117), the
coefficient on the FDI inflows for the developing countries is insignificant positive, indicating that the FDI inflows have no significantly direct effect on the economic growth. However, the interaction terms of FDI in the developing countries are all statistically significant. The empirical results suggest that as the low level of absorptive capacity, the positive effect of FDI inflows on the economic growth may exist only if the developing countries have a threshold for the human capital, infrastructures and technology gap. Therefore, the empirical results suggest that the FDI inflows may not enhance the economic growth for the developing countries unless they enhance their poor absorptive capacity. These results have the same finding as what has been found by Borensztein, Gregorio & Lee (1998, p. 115) that the developing countries are required to have a minimum threshold stock of human capital to make them benefit from the technology transfer, therefore to increase their economic growth. In addition, the empirical results have the same findings as OECD (2002, p. 10) that the FDI inflows contribute positive effect to the economic growth only if the host countries have attained a certain degree of absorptive capacity.
V. Conclusion
This paper examines the impact of FDI inflows on the economic growth in both developed and developing countries by using a large cross-country sample for the period 1984-2004. These empirical results indicate that there exists a significant positive relationship between the FDI inflows and the economic growth based on the total sample. However, the significant positive relationship between FDI inflows and economic growth exist only in the developed but not in the developing countries. Furthermore, FDI inflows have a stronger positive influence on the economic growth in the developed countries than that in the developing countries.
However, the empirical results partly confirm the findings of Li and Liu (2005, p. 404) that as the absorptive capacity level is high, the FDI inflows can positively affect the economic growth for the developed countries. On the other hand, the test results suggest that there exists a strong positive interaction effect of FDI with the human capital, technology gap and
infrastructure on the economic growth for the developing countries, implying that the FDI inflows may positively affect the economy only if there is a minimum level of human capital, technology and infrastructure.
As the empirical results from this analysis confirm that the FDI inflows tend to stimulate the economic growth based on the total sample, they provide a new and recent evidence to strongly support the endogenous economic growth theories. In addition, the empirical results prove that the absorptive ability is one of the most important factors to decide whether the FDI inflows can have a significant positive effect on the economic growth for the host countries. The insignificant results for the developing countries indicate that there is not such minimum level for them to obtain the positive externalities of the FDI inflows.
One of the limitations of this analysis is the selection of the econometric solution of the endogenous problems. The choice of different instrumental variables may have an effect on the significance level of the analysis results. Furthermore, as suggested by the test, the efficiency usage of the infrastructures is more important than the quantity when people try to examine the effect on the economic growth, therefore, it should choose a more effective proxy for the infrastructure in the test. Finally, as there is an increased number of people prefer to study abroad and there exists more frequency of the international academic exchanges, it is difficult to quantitative the data for the human capital for the host countries, and the test should have a more appropriate variable instead of the secondary level attainment to be the proxy for the human capital.
Bibliography
Alfaro, L., Chanda, A., Kalemli-Ozcan, S., & Sayek, S. (2004). FDI and economic growth: the role of local financial markets. Journal of international economics, 64(1), 89-112. Azman-Saini, W. N. W., Law, S. H., & Ahmad, A. H. (2010). FDI and economic growth: New
evidence on the role of financial markets. Economics letters, 107(2), 211-213.
Balasubramanyam, V. N., Salisu, M., & Sapsford, D. (1996). Foreign direct investment and growth in EP and IS countries. The Economic Journal, 92-105.
affect economic growth?. Journal of international Economics, 45(1), 115-135.
Carkovic, M. V., & Levine, R. (2002). Does foreign direct investment accelerate economic growth?. U of Minnesota Department of Finance Working Paper.
Choe, J. I. (2003). Do foreign direct investment and gross domestic investment promote economic growth?. Review of Development Economics,7(1), 44-57.
Cipollina, M., Giovannetti, G., Pietrovito, F., & Pozzolo, A. F. (2012). FDI and Growth: What Cross‐country Industry Data Say. The World Economy,35(11), 1599-1629.
de Mello Jr, L. R. (1997). Foreign direct investment in developing countries and growth: A selective survey. The Journal of Development Studies, 34(1), 1-34.
Durham, J. B. (2004). Absorptive capacity and the effects of foreign direct investment and equity foreign portfolio investment on economic growth.European economic
review, 48(2), 285-306.
Gao, T. (2005). Foreign direct investment and growth under economic integration. Journal of
international economics, 67(1), 157-174.
Görg, H., & Greenaway, D. (2004). Much ado about nothing? Do domestic firms really benefit from foreign direct investment?. The World Bank Research Observer, 19(2), 171-197.
Herzer, D., Klasen, S., & Nowak-Lehmann D, F. (2008). In search of FDI-led growth in developing countries: The way forward. Economic Modelling, 25(5), 793-810.
Javorcik, B. S. (2004). Does foreign direct investment increase the productivity of domestic firms? In search of spillovers through backward linkages. American economic review, 605-627.
Li, X., & Liu, X. (2005). Foreign direct investment and economic growth: an increasingly endogenous relationship. World development, 33(3), 393-407.
Stock, J. H., & Watson, M. W. (2011). Introduction to econometrics . Boston: Addison Wesley.
UNCTAD, 2001. World Investment Report: Promoting Linkages. United Nations.
Zhang, K. H. (2001). How does foreign direct investment affect economic growth in China?. Economics of Transition, 9(3), 679-693.
Appendix A Classification of country groups
Developed countries
Australia, Austria, Canada, Denmark, Finland, France, Greece, Germany, Iceland, Ireland, Israel, Italy, Japan, Netherlands, New Zealand, Norway, Portugal, Spain, Sweden,
Switzerland, United Kingdom, United States.
Latin-American countries
Argentina, Belize, Bolivia, Brazil, Chile, Colombia, Costa Rica, Dominican Republic,
Ecuador, E1 Salvador, Guatemala, Guyana, Honduras, Jamaica, Mexico, Nicaragua, Panama, Paraguay, Peru, Uruguay, Venezuela.
African countries
Benin, Botswana, Ghana, Kenya, Lesotho, Mauritius, Mozambique, Senegal, Sierra Leone, South Africa, Sudan, Swaziland, Uganda, Zambia.
Fast growing countries
China, Hong Kong, Korea, Singapore, Thailand.
Other developing countries
Algeria, Bangladesh, Cyprus, Fiji, Hungary, India, Indonesia, Iran, Jordan, Malaysia, Nepal, Pakistan, Papua New Guinea, Philippines, Sri Lanka, Tunisia, Turkey.
