In spite of a large increase in FDI inflows to developing countries as reported by UNCTAD (UNCTAD, 2009), the effect of FDI flows on economic growth remains ambiguous. However, whether foreign direct investment (FDI) helps to improve economic growth has been one of the fundamental debates in development and international economics. Recently, this question has received a lot of consideration in the economic literature. So far, it seems that this debate has not been conclusive. The recent contribution of modern economic growth  theories in general predicts that FDI can have a positive impact on economic growth and sustainable economy in the receiving countries (Barro and Sala-I-Martin, 1995; De Jager, 2004; Romer, 1990). Empirical studies, how-ever, produce ambiguous results, and suggest that the growth effects of FDI are conditional on the host country characteristics (Alfaro et al., 2004; Balasubramanyam et al., 1996; Bernstein, 2000; Blomstrom et al., 1992; Borensztein et al., 1998; Kinishita and Lu, 2006; Kokko, 1994; Li and Liu, 2005; Sadik and Bolbol, 2001). Besides, De Mello (1999) finds that the growth effects of FDI depend on the degree of complementary with DI in the receiving countries. In contrast, Carkovic and Levine  (2002) investigate whether the growth effect of FDI depends on the host country’s absorptive capacity for a panel of 72 developed and developing countries from 1960 to 1995. They find that FDI does not exert a positive impact on economic growth in the host country and that it is not conditional on its absorptive capacity.

Considering these matters, it is natural to find such interest in investigating the growth effects of FDI in developing countries. The main purpose of this paper is to examine the growth effect of FDI in a selected sample from Asian, African and Latin American countries. The sample is selected form the top ten recipients of FDI inflows in each region from 1970 to 2005. This paper focuses mainly on the role played by the host country’s absorptive capacity in the growth effect of FDI. The paper examines the following specific research question: Does FDI contribute to economic growth in developing countries alone or does it depend on its initial conditions?

Recent empirical studies suggest that the ability of host countries to exploit FDI efficiently depends on a set of absorptive capacities within these countries, which may help in explaining the ambiguity in the previous empirical studies. This paper contributes to this debate by present-ing a deeper insight into the host country conditions that might affect the FDI-growth nexus. This deeper insight is needed, because the majority of previous empirical studies focus on the interaction between FDI and one of the host country characters (e.g. human capital develop-ment, financial market development, technology gap, infrastructure development, trade openness, etc). This paper investigates the impact of a set of these factors simultaneously on the FDI-growth relationship. This paper also contributes to the existing literature by deter-mining the threshold value of absorptive capacity in the host country that positively correlates FDI with growth.

The rest of the paper is structured as follows: Section two presents an overview of existing empirical studies. Section three is the empirical  specification.  Section  four describes the data, variables set, and estimation  method used for empirical tests. Section five is the empirical results. Section six is the sensitivity analysis and section seven is the summary of this paper.

 

An overview of existing empirical studies

The majority of empirical studies on the impact of FDI on economic growth present controversial evidence. The impact of FDI on host country economic growth comes from the fact that FDI inflow is the most important channel for technology diffusion. The diffusion of technology considered as the main source of conditional convergence between countries and achieving sustain-able development (Elmawazini et al., 2008). The literature appears to offer a thoughtful assessment of the impact of the host country’s absorptive capacity on the dynamic relationship between FDI inflows and economic growth. Many of these studies argue that the degree of technology transfer or externality generating from FDI inflows to the host economy depends on the host country’s absorptive capacity. The term “absorptive capacity” takes account of factors such as the level of human capital development, the level of technology gap, the level of financial development, the degree of trade openness, the level of institution quality, etc. The majority of empirical studies show that host countries do indeed need to pass a certain level of absorptive capacity, known as a development threshold, to be able efficiently exploit FDI.

Recent growth theories argue that the availability of human capital quality plays an essential role in economic growth. The quality of human capital is also crucial for a host country in absorbing the FDI externalities. These externalities are the transfer of skills from MNCs to domestic firms through labour mobility or learning-by-doing. Borensztein et al. (1998) investigate the effect of FDI inflows on economic growth in 69 developing countries using cross-country and cross-section regres-sions. They apply panel data for two decades (1970-79 and 1980-89). Both regressions show that host countries must pass a threshold value of human capital development to benefit from FDI inflows. Xu  (2000) also obtains similar results for 40 countries (20 DCs and 20 LDCs) from 1966 to 1994. By applying the panel data, two stages least square (2SLS) method, he finds that developing countries (DCs) benefit positively from technology transfer provided by US MNCs but not in less developing countries (LDCs). He concludes that LDCs do not reach the minimum human capital threshold required. In contrast, Blomstrom et al. (1992) investigate the impact of FDI on economic growth for 101 countries over the period from 1960 to 1985. They find that education level is  not  essential  to  achieve  an  FDI  growth  effect (Carkovic and Levine, 2002). In addition, Blomstrom et al. (1992) find that the host country must pass a certain threshold of economic development to benefit from FDI.

In turn, Colen et al. (2008) argue that the impact of FDI on economic growth expected to depend on the technology gap between the home and host countries. A large technology gap might slow down the knowledge and technological spillovers. If the technology gap is too wide to bridge, the spillovers may not easily spread to the domestic economy. Castellani and Zanfei (2005) also argue that a higher technology gap may in principle increase the possibility that MNCs tend to crowd out domestic suppliers and competitors.

Absorptive capacity of the recipient economy measured by the technology gap used in many empirical studies. Kokko  (1994) uses the technology gap between foreign and domestic firms, as a proxy for absorptive capacity in 216 Mexican manufacturing industries. He finds that domestic firms can benefit from the technology diffusion from foreign firms if the technology gap between them is small. Li and Liu  (2005) reach a more specific conclusion on the role played by the technology gap in the host economy to obtain the FDI growth effect. They find that for the host country to benefit from attracting FDI, it must have a certain level of technological development. They argue that for a country above a certain level of technology gap, FDI inflows will no longer benefit the host economy.

Despite the numerous empirical studies on the growth effect of FDI, the literature on the FDI-growth nexus seems to have ignored the importance of the role not only of the financial development but also of other factors, such as infrastructure development, trade openness and institutional development. The level of financial development is crucial because a lack of financial market development might be preventing the foreign and domes-tic investors from accessing the financial resources required (Massoud, 2008). Alfaro et al. (2004) and Hermes and Lensink (2003) argue that countries with a better financial system can exploit FDI more efficiently. Hermes and Lensink (2003) provide some explanations on the role of financial system development in exploiting FDI inflows efficiently to promote economic growth in the host country. They argue that financial institutions can help to reduce the risks of investment related to upgrading or adopting new technologies, which affect the speed of technological innovation. Financial systems also determine partly the ability of domestic firms to finance their investment plans in the case of external finance needed. Therefore, the quality of financial system may influence the impact of FDI on the diffusion of technology in the host country. Using cross-country data for two samples (49 and 71 countries) from 1975 to 1995,  Alfaro et al. (2004)  find  that  FDI  played  an  important  part  in contributing to economic growth, and those countries with well-developed financial markets gained significantly from FDI. Using panel data for Arab countries from 1975-2000, Sadik and Bolbol (2003) also find that a certain threshold of financial market development must be reached to benefit from FDI inflows.

Many studies of economic growth define infrastructure as an essential factor behind economic growth (Barro and Sala-I-Martin, 1995; Munnell, 1992; Sanchez-Robles, 1998). Munnell  (1992) points out that good infrastructure can increase the productive capacity of the economy, by increasing resources and encouraging the productivity of existing resources. Therefore, the idea is that host economy may benefit from FDI only if it has appropriate infrastructure development. Kinishita and Lu  (2006) and Yamin and Sinkovics (2009) argue that a good infrastructure is not the only FDI inflows driver but also a pre-requisite for positive spillovers from FDI to the host economy. Kinishita and Lu (2006) investigate the effects of FDI on economic growth when a host country has a sufficient level of infrastructure development for 42 non-OECD countries. They find that technology spillovers via FDI take place only when the host country has a certain level of infrastructure development.

Economic literature also recognises the importance of trade openness as one factor in host country’s absorptive capacity. Frankel and Romer  (1999) argue that trade openness can help to facilitate more efficient production of goods and services through shifting production to eco-nomies that have comparative advantages. Grossman and Helpman (1990) also argue that an open trade regime significantly related with good investment climates, technology externalities and learning effects. Therefore, FDI and trade motivate advancing economies to be more innovative and allow developing ones to draw upon the stock of knowledge of more advanced coun-tries. Adhikary  (2011) also cites that FDI can increase the technological spillover benefits to the host country through widening the scope of international competition and strengthening the supply side capabilities for producing and selling goods and services. These effects lead to a fostering of economic growth as pointed out by Pugel (2007). Edwards (1998) also argues that a country with a greater degree of openness can absorb the new technology brought by FDI at a faster rate than a country with a lower degree of openness. Empirically, Balasubramanyam et al. (1996) and Makki and Somwaru (2004) find that the effect of FDI inflows on economic growth is dependent on the degree of openness.

Although a number of studies investigate the impact of FDI on economic growth, they do not consider the role played by  institution  quality  in  determining  investment efficiency and economic growth, including, for example, that of   Alfaro  et  al.  (2004),   Balasubramanyam  et  al. (1996), Borensztein et al. (1998), Carkovic and Levine (2002) and  Li and Liu (2005).

Olofsdotter (1998) argues that the ability to absorb the new technology provided by FDI inflows can be emphasised in countries with better institution quality. Empirically, Olofsdotter (1998) finds that the strong positive impact of FDI on economic growth reached in countries that have high institution quality. Similarly, Edwards (1998) examines the role played by institution quality in determining the effects of FDI on economic growth for 80 countries from 1979 to 1998. He finds that FDI inflows are more beneficial in countries with higher levels of institutional (as measured by business regulation index and property rights index). Edwards (1998) also finds that the host country that passes a minimum threshold of institution quality enjoys a positive impact of FDI on economic growth. In line with the same argument, the authors in (5) examine the effect of institution quality measured by economic freedom index components, on economic growth in 58 countries from 1975 to 1990. Their findings indicate that economic freedom index has a positive impact on economic growth. They point out that reports on economic freedom suggest that economic growth increased with reduced direct involvement of government in economic activities.

The above review suggests that the growth effect of FDI remains extremely controversial. This may be due to the use of different samples and data by different authors, and partly because of various methodological problems. Moreover, a number of studies do not take into account the role of different factors of host country absorptive capacity on the growth effect of FDI, and the certain level of absorptive capacity required to benefit from FDI. Overall, the above discussion shows that previous empirical studies are sensitive to the measure of absorptive capacity used. To overcome these limitations, this paper investigates a set of factors, as measures of host country absorptive capacity in selected sample from developing countries. This may help to explain the ambiguities in the literature of the contribution of FDI or in exploiting FDI more efficiently to promote economic growth.

 

Empirical specification

To investigate the hypothesis of this paper empirically, the growth rate of real GDP per capita of the host econo-my will be used as a dependent variable. Furthermore, since the data available in DI already included the flows of FDI, so DI will not be controlled in the growth  equation (Bengoa and Sanchez-Robles, 2003; Carkovic and Levine, 2002; Kinishita and Lu, 2006; Li and Liu, 2005 Li and Liu,  2005).  Alfaro  et  al.  (2004)  and   Bengoa  and Sanchez-Robles (2003) do not control DI in their growth equation to avoid the collinearity of DI with FDI. Conversely, one could argue that FDI can have a positive impact on growth, because DI is not controlled in the growth equation. Therefore, for further robustness, DI will be added to the list of independent variables in the growth equation in the sensitivity analysis section.

For enlarging the sample size, the choice of countries and the time-period determined by the availability of the data on the top ten recipients of FDI inflows in Asian, African and Latin American countries. All data were sampled at five-year intervals for 36 years from 1971 to 2005, that is, 1971-1975, 1976-1980, 1981-1985, 1986-1990, 1991-1995, 1996-2000, and 2001-2005. Thus, data permitting, there are seven observations per country. Transforming data from annual observations to five-year averages has several advantages. For example, it may assist in limiting the influence of business cycles on the estimated coefficient such as FDI. Net FDI inflows vary widely from year to year, resulting in large fluctuations that may make the effect of persistent factors ambiguous (Bengoa and Sanchez-Robles, 2003).

This paper follows the contributions of Romer  (1990), and extends the hypothesis of Borensztein et al. (1998) who are the first authors to examine the absorptive capacity of the host country. The paper includes in the Growth equation not only human capital as a proxy of host country’s absorptive capacity but also the techno-logy gap, financial market development, infrastructure development, institution quality and trade openness. This paper also considers most of the explanatory variables in the Growth equation that have been used in previous studies, such as FDI inflows, human capital development (HC), the technology gap between host and home country (TG), the financial market development (MS), infrastructure development (IFR), institution quality (EFW) and trade openness (DOP). The theory predicts that these variables positively related to Growth, except TG that is negative. In addition to these explanatory variables, the empirical model includes a set of control variables that are likely to affect economic growth in developing countries. These variables are also included for testing the hypothesis of this paper and for the robustness of the results.

Among this set of variables, the empirical model includes macroeconomic stability (IFL), government size (GS), black market premium (BMP) and two dummy variables, one for African countries (Africa) and another one for Latin American countries (Latin). These variables also include the interaction term of FDI inflows  with  both of these variables, the human capital, the technology gap,   the   financial   market  development,  infrastructure development, and trade openness and institution quality. The theory predicts that  inflation  rate,  government  size,black market premium variables negatively related to economic growth.

By considering all of these explanatory variables in the Growth equation, the model used in this paper has the following formula:

LGrowthi,t  = α0 + α1 Lagged Growthi,t + α2 LFDIi,t + α3 LHCi,t + α4 LTGi,t + α5 LIFRi,t + α6 LMSi,t+ α7 LDOPi,t+ α8 L(1+IFL)i,t + α9 LGSi,t + α10 L(1+BMP)i,t + α11 LEFWi,t + α12 Africai,t + α13 Latini,t + α14 (LFDI*ABS)i,t+ ηi + εi,t                                                  (1)

Since, it is not simple to measure the technology gap between leading country and following one, a measure of the productivity gap can be used, as in (33, 35, 36). The technology gap measured as the ratio of the gap between US GDP per capita as the world’s technological leader country and host country GDP per capita, relative to host country GDP per capita at constant US dollars. Therefore,

TGi,t= (Ymaxt- Yi,t)/ Yi,t                                  (2)  

                                      

Where, Ymaxt is the GDP per capita of United States, and Yi,t is the GDP per capita of the host economy.

Note that all the variables are in the natural logarithm.

η: unobserved country-specific effect; ε: The disturbance term; i and t: Country and time period, respectively.

(LFDI*ABS): The multiplication of FDI by the host country’ absorptive capacity variables, which capture the interaction terms of FDI with host country’s absorptive capacity factors. This variable allows for testing the hypothesis that the impact of FDI on economic growth determined by the host country’s absorptive capacity. The term “ABS” includes LHC, LTG, LIFR, LMS, LDOP and LEFW variables.

From the model specification, three possible results can assess the role played by the host country’s absorptive capacity factors in determining the contribution of FDI in economic growth.

1.  If α2 and α14 both have a positive (negative) sign in the growth equation, then FDI inflows have an unambi-guously positive (negative) effect on economic growth.

2.  If α2 is positive, but α14 is negative, then FDI inflows have  a positive  effect  on  growth,  and  this effect diminishes with the improvements in the host country’s absorptive factors.

3.  If α2 is negative and α14 is positive, then this means that the host country has to achieve a certain threshold level (in terms of absorptive  capacity  developments)  for FDI inflows to have a positive impact on economic growth.

The threshold of the host country’s absorptive capacity calculated by finding the partial impact of FDI on Growth as follows:

(∂Lgrowth/∂LFDI)= α2 + α14 ABS=0, then the threshold of host country’s absorptive capacity (ABS) = - α2/ α14 (3)

The sensitivity of the growth model specified is tested by controlling for other determinants of economic growth, by including a set of host country’ absorptive capacity variables and by applying panel of GMM estimations. To gain some robustness, the list of countries is expanded, changing the time-period and removing the observations outlier also carried out in the next section.

 

Data, variables and estimation method

The empirical test is based on 24 developing country recipients of FDI inflows selected from three regions; Asia, Africa and Latin America over the period from 1971 to 2005. The choice of countries and the time-period determined by the availability of data. This paper identifies countries with high-FDI flows over the entire thirty-six year sample period. The motivation for employing the size of FDI flows is to examine the hypothesis of this paper within successful developing countries. A list of the economies integrated in the sample, the variables used in the empirical test and the data sources themselves are presented in Appendix.

To gain robustness results, the paper uses the method of GMM-in-System estimator of Blundell and Bond (1998). This technique can help to overcome the endogeneity problem of some regressors, especially FDI, which leads to inconsistent estimations. So far, endogeneity is dealt with by using lagged period of endogenous variables as effective instruments in panel dynamic techniques (Arellano and Bond, 1991). The Hansen and Sargan tests were also used to approve the validity of the overall appropriateness of the instruments used. The Arellano-Bond test also was used for testing second-order serial correlation in residuals¹

Column 1 of Table 1 reports the results of the Growth equation. As expected all the explanatory variables have a right sign and are statistically significant. This column shows that countries with low level of initial GDP per capita grow faster as shown by the negative sign of the lagged  GDP  per  capita  growth  rate  (Lagged Growth)².

 

 

Column 1 also shows that FDI inflows significantly and positively related to economic growth, which is consistent with the empirical literature and economic growth theory, stating that FDI inflows in general have a positive impact on economic growth. The coefficient on LHC, the measure of human  capital  development,  also  positively and significantly related to growth as reported in column 1. This result highlights the importance of education in the growth process of these economies³. The govern-ment size proxy has a negative and significant impact on economic growth, suggesting that a higher government spending to GDP ratio leads to lower economic growth. The black market premium is also negatively and significantly related to economic growth, where higher international price distortions lead to lower economic growth. The two dummy variables also significantly and negatively relate to economic growth. These results suggest that African and Latin American countries tend, ceteris paribus, to grow more slowly than Asian countries. This finding is not surprising given the fact that Africa and Latin America countries suffer the most from slower economic growth, compared to Asia economies. Column 1 also shows that the inflation rate has a right sign, but statistically significant at lower (10%) confidence level, confirming the findings of Borensztein et al. (1998)⁴.

Column 2 presents the estimated results for testing the growth effect of FDI through a well-educated workforce by including the interaction term of FDI with the human capital development proxy (LFDI*LHC)⁵ in the growth equation. Column 2 shows that FDI has a negative impact on economic growth, while the interaction term of FDI with human capital significantly and positively relate to economic growth⁶. These facts suggest that a minimum level of human capital is required for FDI to contribute positively to growth, confirming the results of Borensztein et al. (1998). This suggests that all econo-mies with gross ratio of secondary school enrolment⁷ above 24.77 will benefit positively from FDI inflows. In this case, by taking the average value of gross ratio of secondary school enrolment in each country for the period from 1971 to 2005, 20 out of 24 countries satisfy this threshold. Note that there are four countries below the minimum estimated threshold including Pakistan, Angola, Congo and Madagascar.

Column 3 presents the estimated results for testing the growth effect of FDI through the effect of the techno-logical gap between developing countries and developed ones by including the technology gap variable along with the interaction term of FDI with the technology gap proxy (LFDI*LTG)⁸ in the growth equation. This column shows that the technology gap (LTG) variable appears to have a significant negative impact on economic growth. This implies that a wide technology gap between home and host country tends to slow down economic growth  of  the host  country,  as  suggested  by  a  number  of  empirical studies, such as those by Krogstrup and Matar (2005), Li and Liu (2005) and Lim and McAleer (2002). Column 3 also shows that the coefficient of FDI is positive and the coefficient of the interaction term of FDI with technology gap is  significantly  and  negatively  related  to  economic growth. This suggests that a certain level of technological development is required for FDI to contribute positively to growth⁹, confirming Li and Liu (2005)’s findings. Column 3 shows that not all economies will benefit positively from attracting FDI when the technology gap level is above 5014.05¹⁰. The sample suggests that 5 out of 24 countries can no longer exploit the positive impact of FDI on growth¹¹.

Column 4 tests the hypothesis that the contribution of FDI to economic growth is conditional on the levels of infrastructure development. Column 4 confirms the hypothesis that the relation between FDI and growth is contingent on the level of infrastructure development. This suggests that host country must reach a certain level of infrastructure development to benefit positively from FDI. This confirms previous findings of empirical studies, such as that of Kinishita and Lu (2006), Bernstein (2000), Lumbila (2005) and Munnell (1992). From column 4, the certain level of pre-infrastructure required equals 25.27. In this case, 17 out of 24 countries can satisfy a requested pre-telephone network requirement to exploit the positive impact of FDI on growth over the average of the period¹².

Column 5 shows that the financial market development has a significant positive impact on economic growth in line with Alfaro et al. (2004), Barro (1991), King and Levine (1993), Mankiw et al. (1992) and Romer (1993). Column 5 also shows that the certain level of financial development is required to benefit positively from FDI equals 22.19, confirming the findings of Alfaro et al. (2004) and  Durham (2004).  Generally, 8 out of 24 countries cannot satisfy a requested M2 as a share of GDP requirement to exploit the positive impact of FDI on growth are the average of the period under consideration¹³.

Column 6 also shows that trade openness is significantly and positively related to economic growth, confirming empirical studies on the impact of trade openness on economic growth (Balasubramanyam et al., 1996; Makki and Somwaru, 2004; Yanikkaya, 2003). Column 6 also shows that a threshold of degree of openness equals to 46.99. Thus, 12 out of a selected sample that can satisfy a requested degree of trade openness requirement to reap the positive impact of FDI on growth over the average of the period.

Column 7 examines whether economies with better institutional quality can exploit FDI more efficiently. In line with the literature, as can be seen in Column 7, the  resultconfirms that a higher quality of institution positively  affects economic growth in these economies. The calculated threshold for the economic freedom index is 5.25, thus practically any improvement in the EFW index above this threshold would yield a positive growth effect of FDI. The estimated threshold shows that 11  out  of  23 economies 14 do not pass this threshold.

 

Sensitivity analysis

The empirical results presented above based on a small sample of 24 top developing countries that are successful in attracting FDI inflows in three regions; Asian, African and Latin American regions. The reason for using that sample is to test the hypothesis of this paper within successful countries. As a result, the findings might be sensitive to the sample choice. Thus, the robustness of the results tested by using a larger country sample. To enlarge the sample size, the choice of countries and the time-period determined by the availability of the data on most developing countries. Since the majority of developing countries have started attracting FDI inflows from the early 1980s, the time-period of this section covers 1981 to 2005. All data were sampled at five-year intervals for 25 years from 1981 to 2005, that is, 1981-1985, 1986-1990, 1991-1995, 1996-2000, and 2001-2005, thus data permitting there are five observations per country. These changes increase the sample size from 24 to 76 countries and the number of observations from 168 to 380. A list of the economies integrated in the sample and used in the empirical investigation presented in Appendix.

Economic growth literature shows that the rate of physical capital formation positively affects economic growth, as concluded, for example, by Barro, (1991), Kormendi and Meguire (1985) and Li and Liu (2005). Thus, the robustness of the results is also tested by including domestic investment (DI)¹⁵ in the growth equation and by reducing omitted variables biases. This section also examines the outliers observed to gain some robustness. A common statistical test is Cook’s distance measure, which provides an overall measure of the influence of an observation on the estimated regression coefficient. The higher the value of the Cook’s D the more frequent outliers are the observations, and lowest value of the Cook’s D, zero or near-to-zero is the assumed. The potential critical value is 4/number of observations. Appendix includes a table that shows the outliers result of Cook’s D test, which obtained from regression all explanatory variables in the growth equation by applying OLS estimation. The multicollinearity check among explanatory variables also reported in Appendix. The  test shows   that  the  problem  of  multicollinearity  does   not exist and estimated coefficients are stable.

Table 2 presents the results of the growth equation obtained by applying the GMM estimator. As can be seen from column 1 of Table 2, FDI still has a positive and significant impact on growth, confirming previous findings of this paper. Column 1 also  shows  that  the  impacts  of HC, IFL, GS and BMP on economic growth confirmed. Column 1 also shows that two dummy variables have a right sign and are statistically significant. Columns (2, 3, 4, 5, 6 and 7) show that the hypothesis that the relation between FDI inflows and economic growth is contingent on the host country’s absorptive capacity confirmed. The results indicate that FDI inflows contribute positively to economic growth, only if the host countries have reached a certain level of human capital development, technolo-gical gap, infrastructure development, financial system development, degree of trade openness and institutional development.

 

 

These results suggest that changing the sample size and omitted variables do not affect the main findings of this paper. Namely, FDI contributes positively to econo-mic growth of the host countries, but the magnitude of this effect depends on the host country absorptive capacity.

To gain more robustness results, we re-estimated the growth equation after excluding outliers in observations. The results of GMM estimators presented in Appendix. The results indicate that there is a threshold level of host country’s absorptive capacity development, and the countries gain the most from FDI spillovers, if they reach this threshold.

A large number of empirical studies examine the growth effects of FDI in developing countries. However, the results of these studies fail to confirm whether FDI helps to improve economic growth in the host countries. Thus, the main purpose of this paper is to examine the growth effect of FDI on the host economies in selected samples, from Asian, African and Latin American countries, for the data from 1971 to 2005. The paper investigates firstly this hypothesis among the most successful countries, and then in most of Asian, African and Latin American countries 1981 to 2005. Particularly, the paper examines the following specific research question: Does FDI contribute to economic growth in developing countries alone or does it depend on its initial conditions?

The results of this paper confirm the numerous empirical studies and economic growth theories studying the growth effect of FDI, stating that FDI has in general  a positive  impact  on  economic  growth. The results of this paper clearly show that domestic investment, human capital, infrastructure development, financial market development, trade openness, and institution quality positively related to economic growth. In contrast, the technology gap, government size, black market premium and the inflation rate negatively related to economic growth. The result of this paper  also  shows  that  African and Latin American countries are, assuming other factors remaining fixed, more likely to grow less than Asian countries.

The main finding of this paper is that FDI can have a positive impact on economic growth, but its magnitude depends on the host country conditions, as suggested by the significant impact of the interaction terms of FDI with a set of host county characteristics. These findings suggest that a certain level of absorptive capacity is required for FDI to be beneficial to the host economy. These findings are in line with many empirical studies on this topic, although it is contrary to the findings of Carkovic and Levine (2002) for panel data, Blomstrom et al. (1992) for cross-section data, and Herzer et al. (2008) for time series data. Furthermore, change in applied techniques, omitted variable, sample countries used or observations outlier influences the results of this paper.

Overall, the findings of this paper support the fact that policies considered to attract more FDI are not satisfactory in generating spillovers for economic growth. Improving the investment environment through developing the host country’s absorptive capacity factors should be a priority for policymakers in these countries to exploit FDI efficiently.

This investigation suggests that further empirical studies and researches are required to re-examine which type of foreign capital inflows fosters economic growth in the host country. However, this claim requires further analysis to empirically test whether such a specific capital inflow forms exist, and if so, how significant it is.

The author has not declared any conflict of interest.