“Impact of foreign direct investment on economic growth in Africa”

Several studies have been conducted to examine the influence of foreign direct investment (FDI) inflow on economic growth. Indeed, the overall evidence is best characterized as mixed. This paper investigates the effect of FDI on economic growth in some randomly selected African economies from 1980 to 2013, using a modified growth model by Agrawal and Khan (2011). This model consists of Gross Domestic Product, Human Capital, International Technology Transfer, Labor Force, FDI and Gross Capital Formation (GCF). Ordinary least squares and generalized method of moments were used as the estimation techniques. Of all the results, only Gross Capital Formation, Human Capital, and International Technology Transfer in the Central African Republic were found not to have any statistically significant influence on economic growth. In general, the impact of FDI on economic growth in African countries is limited or negligible. Consequently, this study observes that a 1% increase in FDI would result in a 0.12% increase in GDP for South Africa, a 0.05% increase in Egypt, a 0.03% increase in Nigeria, a 0.02% increase in Kenya, and a 1% increase in GDP in the Central African Republic. The findings also reveal that South Africa’s growth is more affected by FDI than the other four countries. The study also provides possible reasons behind South Africa’s great show of FDI and the lessons other African countries could learn from South Africa better utilization of FDI. This study integrates the related drivers of the effectiveness and success of FDI.


Introduction 
Today, the impact of globalization can never be overemphasized.The tremendous growth in foreign direct investment (FDI) is one of the important influences of globalization (Whalley & Xin, 2009;Mohamed & Sidiropoulos, 2010).Although, African countries are not major beneficiaries of FDI flows in the world; that notwithstanding, FDI inflows have experienced a dramatic upsurge in many countries in Africa over the past two decades (Agrawal and Khan, 2011; Ozturk, 2007).In addition to inward FDI from developed and 'emerging economies', the last two decades have also seen a marked increase in the FDI activities of African firms -especially within the region (UNCTAD, 2015).Indigenous companies in South Africa, Nigeria, Togo, and Kenya have operations in more than two countries in the region (Anyanwu and Yameogo, 2015;UNCTAD, 2015).This unique upsurge in the crossborder activities of African firms presented a unique opportunity to investigate the patterns and outcomes of FDI in these countries (Agrawal and Khan, 2011; Aregbesola, 2014; UNCTAD, 2013).In addition, these unprecedented levels of competition for inward FDI in the region are on the premise that inward FDI, whether in the form of portfolio or direct investments, also presents a unique opportunity for African economies to achieve improved economic growth (Asiedu, 2002;Johnson, 2006).As a result, most African countries are now competing more vigorously for FDI inflows in their quest for rapid economic growth (Anyanwu, 2012;Lee, 2005;Li & Liu, 2004).Furthermore, the increasing competition for FDI in the past two decades in most developing and developed countries is also stimulating the extended and controversial debate about the cost and/or benefits of FDI (Agrawal and Khan, 2011; Wijeweera, Villano & Dollery, 2010).While many scholars concur about the positive influence of FDI on economic growth, given various incentives and appropriate policies, others highlight the potential drawbacks to include the negative impact on both balance of payments and competition in the host country (Johnson, 2006;Ozturk, 2007).Durham (2004) and Agrawal and Khan (2011) also observed the efficacy of exploiting FDI by nations with superior financial market regulations (Ozturk, 2007;Lee, 2005;Li & Liu, 2004).Although various studies have examined the influence of FDI inflow on economic growth, with most reporting mixed results, there seem to be few empirical analyzes on this phenomenon (Durham, 2004).Moreover, many of the previous studies often utilized GDP per capita as a substitute for economic growth, despite the fact that FDI can only moderate the reward for labor (Agrawal and Khan, 2011).In addition, Saggi (2002) also suggests the need for a survey of relationships between inward FDI and economic growth, given its influence on economic development.
The above considered, the main objective of this study was to investigate the effect of FDI inflows on economic growth in some selected African economies from 1980 to 2013.The analytical focus was premised on developing countries, due to the submission of Wyk and Lal (2008).This paper, therefore, argues that the practice of pooling developed and developing economies together in analyzing the impact of FDI on growth is inappropriate.This is on the premise that while developed economies have substantial amounts of two-way FDI flows, developing economies, on the other hand, are almost exclusively recipients of FDI flows (Wyk & Lal, 2008;El-Wassal, 2012).Section two of this paper is the literature review.Section three details the adopted methodology.Section four includes the analysis of the various data collected, results and discussion of findings.Section five presents the conclusion, and the recommendations and implications of the study.

Review of related literature
FDI can be in the form of joint venture, Greenfield investment, mergers and acquisitions, management knowhow, portfolio investment and transfer of technology from one country to another (Johnson, 2006;Madsen, 2007).The neoclassical models of growth and endogenous growth models support most of the empirical work on the FDI-growth relationship (Ozturk, 2007;Solow, 1956).According to the neoclassical growth theory, economic growth generally comes from two sources: factor accumulation and total factor productivity (TFP) growth (Felipe, 1997).These are often dependent on the host country's bargaining power (assets that it is able to provide to investors) in attracting FDI, such as its market size, human capital, geographical location, and infrastructure (Fedderke and Romm, 2005).In other words, Fedderke and Romm (2005, p. 758) summarized that FDI inflow is strongly influenced by the risk profile and net rate of return of FDI liabilities.
Generally, most empirical literature focuses more on the growth of factor inputs.This may not be unconnected with the fact that factor growth is easier to analyze and quantify, while difficulties abound in the measurement of TFP growth due to the unavailability of appropriate data and lack of appropriate econometric modeling techniques (Johnson, 2006;Madsen, 2007).Given the limited contributions of the neoclassical growth theory, the endogenous growth literature posits that FDI not only contributes to economic growth through technology transfers and capital formation (Lucas, 1988;Merican, 2009;Blomstrom et al., 1996), but also through the augmentation of the level of knowledge via skill acquisition and labor training (De-Mello, 1999; Solow, 1956).Consequently, the framework of endogenous growth models identified three main channels through which FDI can affect economic growth.First, FDI may increase capital accumulation in the host nation through the introduction of new inputs and technologies (Dunning, 1993).Second, FDI can raise the level of skills and knowledge in the host country through training (De-Mello, 1999).Third, FDI can increase the level of competition in the host country through reduction in entry barriers and the market power of existing firms (Johnson, 2006).Consequently, the findings of David Ricardo (comparative advantage), Heckscher-Ohlin's (factor proportions), and Porter's competitive advantage clearly establish the practical inevitability of foreign direct investment, as the foundation upon which a nation's economic growth and prosperity can be built (Aregbesola, 2014).

FDI and economic growth.
A number of interesting studies on the role of FDI in stimulating economic growth have appeared in the past three decades.Bhagwati (1978) was the first to present a theory on the dynamic effect of trade policy regime on FDI, trade and growth in a given host country.This theory was presented as an extension to his theory of immiserizing growth, which was further developed by Bhagwati (1994) and Brecher and Findlay (1983).Also, referred to as the "Bhagwati hypothesis", the theory postulates that FDI inflows in a restrictive, import-substitution (IS) regime often retard, rather than promote economic growth.This is based on the fact that with the low comparative advantage of the host developing country, FDI often becomes an avenue for multinational companies to maintain and even increase their profitability, market share and the economic rent created by the highly protected domestic market in an IS regime (Aitken, Hanson & Harrison, 1997;Brecher and Findlay, 1983).
In addition to the above two postulations, many theoretical and empirical studies also acknowledge several channels by which FDI may negatively or positively influence economic growth.Apart from improved capital accumulation and efficiency via contract and demonstration effects in the host economy, technological change, competition, improved exports and human capital augmentation are other channels (Mohamed & Sidiropoulos, 2010;Blomstrom and Kokko, 1998).FDI also accelerates growth by generating employment in the host country, and through sharing of knowledge and management skills through forward and backward integration in host countries (Brecher and Findlay, 1983).Investment in learning and innovations by local firms may also generate productivity spillovers for the host economy (Whalley & Xin, 2009).
As a byproduct of technology transfer, technology diffusion is also a channel for improving economic growth (Grossman and Helpman, 1990).This was corroborated by Romer (1990), who highlighted the importance of international trade in technology diffusion, as an impetus to economic growth.Apart from Romer (1990) and Grossman and Helpman (1990), the importance of technological diffusion in enhancing economic growth in developing economies was posited by other endogenous growth theorists (Eaton and Kortum, 1999;Young, 1991).Other pro-FDI liberalization studies like Love and Chandra (2004) and Chakraborty and Basu (2002) also supported the importance of FDI in enhancing economic growth.Consequently, Love and Chandra (2004) concluded that FDI should be encouraged if the objective of accelerating economic growth in developed economies is to be realized.That notwithstanding, Durham (2004) warned that although FDI may promote economic growth in most developing countries, the extent of the benefits depends on labor-force skills, trade policies and the absorptive capabilities of local firms (Felipe, 1997;Durham, 2004).
However, in a deviation from the above positive sentiments, many past studies on the impact of FDI on economic growth have produced mixed results.Singh (2003), Chakraborty and Basu (2002) and Young and Lan (1996) are not so positive about the influence of FDI on economic growth.Specifically, Singh (2003) and Young and Lan (1996) observed the influence of FDI on economic growth to be industry biased, and the level of development in the home nation also distorts the gains from FDI.It may even depend on the social and economic climate in the host nation (Young and Lan, 1996).This negative sentiment was also echoed by Chakraborty and Basu (2002) about the influence of FDI and trade in the diffusion of technology on economic growth.In addition, FDI may negatively influence the economic growth of the host nation due to significant reverse flows by the multinational companies (Akinlo, 2004).According to Akinlo (2004), these reverse flows could be in the form of dividends, large concessions and remittances of profits.Many empirical studies also provided evidence of significant "crowd in" on other investments at the macro level (Mohamed & Sidiropoulos, 2010;Durham, 2004).Some empirical studies also observed insignificant influence of FDI on economic growth and that FDI is no more productive than domestic investments (Love and Chandra, 2004).

Methodology
This research employed time series data of the selected African countries, from 1980 to 2013.For representativeness and ease of generalization (Singh, 2003), the following countries were randomly selected from each of the regions in Africa: Nigeria (West Africa), Egypt (North Africa), Kenya (East Africa), South Africa (Southern Africa), and Central African Republic (Central Africa).The ordinary least squares regression (OLS) and the generalized method of moments (GMM) were the two estimation techniques used.Firstly, we performed an OLS with time fixed effects.However, due to a probable endogeneity bias and serial correlation of the error term, one serious defect of OLS might be unstable and inconsistent parameter estimates (Ozturk, 2007).To deal with this probable problem, the GMM technique was used as compliment to provide consistent estimates (Singh, 2003).

Multiple regression analysis: OLS and GMM.
Based on a modified growth model by Agrawal and Khan (2011), the proposed model was constructed from the basic production function below: where Y = Output level (i.e., GDP), K = Capital (gross capital formation (GCF) as percentage of GDP) and L = Labor (country's labor force).This is based on the assumption that both factors of production and production technology determine the level of output in an economy (Ogutcu, 2002).Consequently, given constant technology, any increase in labor and/or capital will increase output level in the economy.
According to the new growth theory (Ogutcu, 2002), to analyze the effect of FDI on economic growth, the Coub-Douglas Production Function was expanded with the addition of FDI and human capital (H) (El-Wassal, 2012).
Consequently, the augmented production function is stated as follows: Alternatively, equation 2 can be re-written as follows: ,0 , 1 , .
In the abridged equation 3, Y i,t is the logarithm of real per capita GDP, FDI i,t is FDI inflows as a percentage of GDP, while X i,t represents the set of relevant explanatory variables, μ i is the time-invariant country specific effects, and ε i,t is the error term.Based on the assumptions of the GMM dynamic estimator in equations 4 and 5, and given the small sample (randomly selected countries) size, s = 2 and t = 3,...,T, equation 3 can be translated and expanded to our multiple regression equation 6: ,0 , , where Y = Gross Domestic Product, K = Gross Capital Formation, L= Labor Force, FDI = Foreign Direct Investment, H = Human Capital, ITT = International Technology Transfer, β 0 = Total Factor Productivity, μ i = Country Specific Effects and ε i,t = The Error Term.
GDP (PPP) was measured in current international US Dollars in millions, and FDI inflow was measured as a percentage of GDP.Gross Capital Formation was also measured as percentage of GDP, as a proxy for capital; while Labor Force was proxied by the Human Development Index (HDI).
The importance of education to economic growth was reflected via human capital; accordingly, human capital was proxied by the ratio of secondary and tertiary institution enrolment in the population (El-Wassal, 2012).In addition, the import of machinery was used as a proxy for International Technology Transfers.Total Factor Productivity (TFP) was factored in to explain output growth, and was not accounted for by the other explanatory variables.However, log values of the variables were used to facilitate the use of the ordinary least square method (Agrawal and Khan, 2011).The result of the unit root test assumed stationarity of the series for all the variables by the rejection of the null hypothesis for second difference at all the critical values (maximum lag of one).Therefore, the models follow an integrating order of 1(2) process and are a stationary process (Ozturk, 2007).

Estimated regression results.
The estimated regression results for all countries are shown in Table 2 (below).In the model of South Africa, the entire variables were positive and significant for both OLS and GMM.The correlation coefficient (r) was 0.67 for South Africa.This implies a positive relationship between economic growth and all the explanatory variables in the model.The adjusted R 2 was 0.59, which implies that about 59% variations in economic growth could only be explained by FDI and other explanatory variables, while the remaining 41% were due to other variables outside the regression model and which also affect growth.The assumption of independent errors was tested with the Durbin-Watson statistics, which monitors for serial correlations between errors (Akinlo, 2004).A value of 1.99 complies with the assumption of no independent errors, since a value less than 1 or greater than 3 are definitely cause for concern (Singh, 2003).It also shows that there is no problem of autocorrelation (Keller, 2002).The estimate of FDI was positive and significant, although limited, for both OLS and GMM, which indicates that FDI inflows contributed minimally in explaining the level of economic growth in South Africa during the study period.Specifically, a coefficient of 0.12 implies that a 1% increase in FDI would result in a 0.12% increase in GDP for South Africa.
The results from the model of Egypt also showed that all variables were positive and significant for both OLS and GMM results.The coefficient of FDI was 0.05, which implies that a 1% increase in FDI would result in 0.05% increase in GDP for Egypt.However, the results from the model of Nigeria were not as impressive as those of South Africa and Egypt.For instance, the estimate for Human Capital was positive and significant for OLS.This indicates that a previous level of Human Capital contributed positively in explaining the level of economic growth in Nigeria.However, the second lag was not significant at the GMM and had a negative sign.This shows that Nigeria's human capital deteriorated with time (Usman & Ibrahim, 2012).That notwithstanding, other explanatory variables were significant at 10% level of significance, both at OLS and GMM (Ozturk, 2007).The coefficient of FDI was 0.03 -which implies that a 1% increase in FDI would result in a 0.03% increase in GDP for Nigeria.
The results from the model of Kenya were also not as impressive as those of South Africa and Egypt.Although all the variables were positive and significant for both OLS and GMM, only Human Capital (H) was significant at 5%, while FDI and the other explanatory variables were only significant at 10% level of significance (Usman & Ibrahim, 2012; Ogutcu, 2002).The coefficient of FDI was 0.02, which implies that a 1% increase in FDI would result in a 0.02% increase in GDP for Kenya.Lastly, the Central African Republic recorded negative indicators based on the results from the model.Only Labor Force (L) was significant at 10%.FDI and the other explanatory variables were not significant.The correlation coefficient (r) was also very low at 0.09.Although the estimates of FDI were positive, they were not significant for both OLS and GMM.This indicates that FDI inflows failed to explain the level of economic growth in Central African Republic during the study period.That notwithstanding, a coefficient of 0.01 implied that a 1% increase in FDI would result in a 0.01% increase in GDP for the Central African Republic during the study period.

Discussion of findings.
The OLS and GMM results suggest that FDI inflows to the selected African countries over the past three decades have had a limited impact on economic growth.This is not surprising, given similar results from previous studies (Fedderke & Romm, 2005

Conclusions
This paper investigates the effect of FDI on economic growth in some randomly selected African economies from 1980 to 2013, using a modified growth model.Specifically, except for Central African Republic, the estimate of FDI was positive and significant for both OLS and GMM in all the selected countries.However, despite the significant and positive coefficients of FDI, yet the most important feature of the coefficients is the extremely small magnitude.In econometrics, these results imply a minimal or negligible impact of FDI on economic growth.For instance, the study found that a 1% increase in FDI would result in a 0.12% increase in GDP for South Africa, a 0.05% increase in GDP for Egypt, a 0.03% increase in GDP for Nigeria, a 0.02% increase in GDP for Kenya, and a 1% increase in GDP for the Central African Republic.That notwithstanding, this study provides possible reasons for South Africa's better use of FDI and the lessons other African countries could learn from South Africa and Egypt so that they can make better utilization of FDI.In general, this study revealed that South Africa's growth is more affected by FDI than the other four countries.Therefore, this paper posits that South Africa, and, indeed, other African economies, has great potential to improve its inflow of FDI and the growth benefits accruing from FDI (Anyanwu & Yameogo, 2015; UNCTAD, 2015).

Theoretical and managerial implications of the study
This paper contributes to the existing literature in four ways.First, unlike previous studies which largely consider either developed and developing economies, or a group of both developed and developing economies, our study focuses solely on African countries.Second, the analytical focus was premised on developing countries, due to Wyk and Lal (2008) submission.This paper, therefore, argues that the practice of pooling developed and developing economies together in analyzing the impact of FDI on growth is inappropriate (Wyk & Lal, 2008;El-Wassal, 2012).Third, the study adds to the growing literature by examining a range of variables that seem to play an important role in shaping the relationship between FDI and economic growth.Lastly, the study also integrates the related drivers of the effectiveness and success of FDI, by confirming the significant relationship between FDI and economic growth in the selected countries (Anyanwu, 2012;Aregbesola, 2014).In summary, this study posits that FDI tends to have a significant effect on economic growth through multiple channels like gross capital formation, human capital enhancement, labor force, and technology transfer and spillover (Fedderke and Romm, 2005).It is also imperative for policy makers in Africa to understand the role of incentives.To achieve economic and social objectives, incentives can compensate for market failures (OECD, 2007, p. 3).It is also important for developing countries to know that, contrary to expectations, FDI may not automatically lead to economic growth, as is insinuated by many policy makers in the region (Akinlo, 2004).Rather, African countries should focus on the impact of policies on technological change, as well as the diffusion of knowledge through FDI from developed countries (UNCTAD, 2013; Anyanwu, 2012; Aregbesola, 2014).In addition, policy makers in both Nigeria and the Central African Republic should prioritize the expansion of existing human capital and improve the educational system to raise the stock of Human Capital.
However, care must be taken in using the output of this study, due to some inherent limitations.Similar to most empirical literature on the FDI-growth relationships using cross-country evidences, the study suffers from two major econometric weaknesses.First is the problem of endogenity, since most explanatory variables are likely to be jointly endogenous with economic growth (Anyanwu, 2012;Hailu, 2010).The second weakness relates with the presence of periods and country-specific omitted characteristics or variables affecting both inflows of FDI and economic growth (El-Wassal, 2012).That notwithstanding, generalized method of moments (GMM) approach was adopted in this paper to address the potential endogeneity of the regressors (El-Wassal, 2012).That notwithstanding, future studies might consider the inclusion of omitted variables, due to data limitation (Anyanwu & Yameogo, 2015).
The data set was collected from the following sources: UNCTAD (United Nations Conference on Trade and Development), World Bank Databank, United Nations Commodity Trade Statistics (UNCTS) Database, Matrade Database and World Trade Organization (WTO) Statistics Database, the International Monetary Fund (IMF), the United Nations Statistics Database (UNdata), and publications of national central banks and other agencies of the governments of the selected countries.

Table 1 .
Summary of unit root test results

Table 2 .
OLS and GMM estimation regression results for FDI inflow