Download A Solution to Africa's Foreign Direct Investment Woes - Lecture Notes | 031 043 and more Papers Psychology in PDF only on Docsity!
Aggressive Trade Reform and Infrastructure Development: A Solution to Africa’s Foreign Direct Investment Woes Elizabeth Asiedu Department of Economics, University of Kansas August 2002 Abstract This paper examines why countries in Sub-Saharan Africa (SSA) did not partake in the foreign direct investment (FDI) boom to developing countries in the 1990s, despite improvements in infrastructure and trade openness. The subdued response of FDI to policy changes can be partly explained by the fact that policy reform in SSA was marginal compared to other regions. As a consequence, SSA’s global relative standing declined in the 1990s. Thus, even though SSA implemented reform, SSA was less open and had weaker infrastructure (in relative terms) in the 1990s than in the 1980s, suggesting that SSA was more attractive for FDI in the 1980s than in the 1990s. These results imply that in order to attract FDI in this era, SSA needs to improve its global relative standing—this will require aggressive trade reform and drastic improvements in infrastructure.
Key words : Africa, foreign direct investment, infrastructure, openness. Please address all correspondence to: Elizabeth Asiedu Department of Economics, University of Kansas 213 Summerfield Hall, Lawrence, KS 66045Phone: (785) 864-2843; E- mail: asiedu@ku.edu
I am grateful to Kwabena Gyimah-Brempong, Nicholas Charalambides, Ted Juhl, Mehrene Larudee, Donald Lien and Francis Owusu for helpful comments.
“Despite the efforts of African governments to attract foreign direct investment by improving their policy frameworks, and despite signs of renewed economic activity in Africa, Africa has been largely bypassed by the recent foreign direct investment boom” (UNCTD 1999, p.1)
1. INTRODUCTION When it comes to foreign direct investment (FDI) to Africa, what comes to mind is investment in natural resources, mainly investments in minerals and oil. 1 This is not surprising since the region has large endowments of natural resources. For example, more than half of the world’s cobalt and manganese and over a third of bauxite are located in Africa. Indeed, the common perception is that FDI in Africa is solely driven by natural resource availability. This perception, if true, is problematic because it suggests that FDI in the region is largely determined by an exogenous and uncontrollable factor, and that natural resource-poor countries will attract very little or no FDI, irrespective of the policies the country pursues. This sentiment resonates in the above quote which suggests that FDI has not responded to policy changes in SSA. Specifically, a number of African countries have improved their regulatory framework for FDI, making them more open, permitting profit repatriation, and providing other incentives to attract investment. Yet, Africa’s share of FDI to developing countries continues to decline from about 25 percent in the 1970s, to 10 percent in the 1980s and to about 3 percent in the 1990s. One objective of this paper is to examine why FDI to Africa has been unresponsive to policy changes. A broader goal is to: (i) find the determinants of FDI to Africa; (ii) identify the policy-related (and endogenously determined) variables that affect FDI flows to the region; and (iii) determine whether the impact of the policy-related variables on FDI has changed over time.
(^1) All references to Africa refer to Sub-Saharan Africa. I employ a narrow definition of natural resources, to include only oil and mineral deposits.
The “uniqueness” of Africa is confirmed by the empirical results in Asiedu (2002, p.14), who notes that “Countries in SSA have on the average received less FDI than countries in other regions by virtue of their geographical location—there is a negative effect on FDI for being an African country.” But African leaders can learn from each other. Hence, an empirical analysis that focuses on performance within the continent will have greater credibility among African policymakers. The remainder of the paper is organized as follows: Section 2 provides some background information on the trends in FDI flows, infrastructure development, and openness to trade to SSA and other regions. Section 3 briefly reviews the empirical literature on the determinants of FDI to developing countries and Section 4 describes the data and the explanatory variables. Section 5 presents the empirical results, Section 6 discusses the results and Section 7 concludes.
2. FDI, Openness and Infrastructure in SSA: Some Background Information This section provides a brief summary of the trends in FDI/GDP, infrastructure availability and trade openness for SSA and other regions.^3 I use trade/GDP and tariff rates as measures of openness. To measure infrastructure availability, I use the number of telephones per 1000 population. The measures of infrastructure and openness have been used in several studies (e.g., Wheeler and Mody (1992), Gastanaga el al (1998), Noorbakhsh and Paloni (2000)). The supporting data is provided in Tables 1 and 2.
Insert Tables 1. Insert Table 2.
(^3) See Rodrik (1998) and Collier and Gunning (1998) for a detailed discussion of trade and trade policy for Africa and other regions.
2.1. Trends in FDI flows (i) The importance of natural resources as a determinant of FDI to SSA has decreased. (ii) Between 1980-89 and 1990-99, FDI to SSA increased, however the growth rate was lower than the rate for developing countries as a whole.
In the 1980s, the annual average of FDI/GDP for the top eight natural resource-rich countries in SSA exceeded the average for the region (Table 1). In contrast, in the 1990s, only 3 out of the 8 countries (Nigeria, Zambia and Angola) had FDI ratios higher than the region’s average. This suggests that natural resource availability has become less important as a determinant of FDI. It also indicates that the sectoral composition of FDI has changed and that FDI is now less concentrated in extractive industries. This conjecture is supported by a United Nations report which notes that “For all three major host countries (Germany, United Kingdom and the United States) FDI in mining or oil played a relatively less significant role during the 1990s than during the 1980s,” [cf., UNCTD (1998) p. 14]. The report also documents that the share of primary sector in total U.S. FDI to SSA dropped from 79 percent in 1986 to 53 percent in 1996, from 25 percent in 1990 to 16 percent in 1996 for Germany, and from 37 percent in 1980 to 27 percent in 1995 for the United Kingdom. Table 2 shows that although FDI to SSA increased in the 1990s, the growth rate was much less than the rate for other regions. Specifically, over the period 1980-89 and 1990-99, FDI/GDP for SSA increased by about 183 percent. This compares with an increase of 332 percent for East Asia, 1,690 percent for Europe and Central Asia and 266 percent for all
the average tariff rate in the 1990s was 21.8 percent for SSA and 17.6 percent for all developing countries. Furthermore the gap between the rates for SSA and developing countries widened in the 1990s: rates were 6 percent higher in SSA in the 1980s and 23 percent higher in 1990s. This suggests that SSA’s global relative standing declined in the 1990s. An important implication of this result is that with regards to openness, SSA seemed more attractive for FDI in the 1980s than in the 1990s.
2.3. Infrastructure Development (i) SSA’s physical infrastructure has improved over time. However, the growth rate of infrastructure was considerably less than the rate for other developing countries. In addition, SSA’s infrastructure remains inadequate compared to other regions, suggesting that improvements in infrastructure did not go far enough.
(ii) Widespread improvements in infrastructure in other developing countries has led to a decline in the global relative standing of SSA: with regards to infrastructure availability, SSA looked more attractive for FDI in the 1980s than in the 1990s.
Between the 1980-89 and 1990-99, SSA’s infrastructure increased by about 43 percent (Table 2). This number although large, pales in comparison to the growth rates in other regions. For example, over the same period, the number of telephones per capita increased by 440 percent for East Asia, 286 percent for South Asia, 100 percent for Latin America and 153 percent for developing countries as a whole. Table 2 also shows that despite improvements in infrastructure,
SSA’s infrastructure remains inadequate. The number of telephones per capita for SSA in the 1990s is about 50 percent less than the average for developing countries in the 1980s. In addition the gap between SSA’s physical infrastructure and that of other regions is huge, and the gap has widened over time. The average number of telephones in SSA was about one-third of the average for developing countries in the 1980s and one- fifth of the average in the 1990s. This suggests that SSA’s global relative standing declined in the 1990s. Thus, with regards to infrastructure availability, SSA looked more attractive for FDI in the 1980s than in the 1990s.
In summary, this section has provided evidence that: (i) the composition of FDI to SSA has changed in favor of investments in non-extractive industries and (ii) the outcome of policy reform in the 1990s for SSA may be characterized as absolute progress but relative decline. In section 5, I argue that these two factors partia lly explain the subdued response of FDI to policy changes in SSA.
3. Brief Literature Review According to the “eclectic” theory of FDI, countries that have a “locational advantage” will attract more FDI (Dunning 1988). Location-specific advantage embodies any characteristic (economic, institutional and political) that makes a country attractive for FDI. This includes large domestic markets, availability of natural resources, an educated labor force, good infrastructure, low labor cost, good institutions, political stability, to mention a few. The empirical literature on the determinants of FDI to developing countries has generally focused on identifying the
trade and infrastructure development, in directing FDI flows to SSA. Clearly, this type of analysis is of paramount importance to policymakers in the region. Below, I describe the data and the variables included in the empirical analysis.
4. Description of the Data and the Variables. The analysis covers 34 countries in SSA over the period 1980-2000. As is standard in the literature, the dependent variable is the ratio of net FDI flows to GDP. My choice of independent variables was constrained by data availability. For example, data on important factors such as real wages, tax legislation and trade policies are not readily available for countries in SSA, and therefore I’m unable to test the impact of these important variables on FDI. Below, I describe the independent variables that were significant for my sample. It is worth pointing out that several of the variables employed in previous studies were not significant for my sample.^8 This may be partly explained by the fact that there is little variation in these variables, since my analysis covers only countries in SSA.
4.1. Description of Explanatory Variables (i) Natural Resource Intensity : As posited by the eclectic theory, I hypothesize a positive relationship between natural resource intensity and FDI. I use the share of fuel and minerals in total exports to measure natural resource availability. Very few studies on the determinants of FDI control for natural resource availability. Here, the relationship between natural resource intensity and FDI is far from unanimous. For example, Morisset (2000) finds a positive relationship between resource intensity and FDI. In contrast, Gastanaga et al (1998) find an inverse relationship while Noorbakhsh and Paloni (2002) find no significant relationship (^8) Variables considered include measures corruption, political instability and the inflation rate.
between the two variables. There are two plausible explanations for the conflicted results. First, the differences in results may be partly attributed to sample selection. The analysis of Gastanaga et al (1998) and Noorbakhsh and Paloni (2002) employ data for several developing countries while Morisset (2000) focuses exclusively on African countries. Second, the authors employ different measures of resource intensity. Gastanaga et al (1998) include a dummy variable for oil exporters, Noorbakhsh and Paloni (2002) use the share of minerals in total exports, and Morisset (2000) measures resource intensity by the sum of primary and secondary sectors, minus manufacturing. 9 Note that the measures of resource intensity employed in Gastanaga et al (1998) and Noorbakhsh and Paloni (2002) are not appropriate for analyzing FDI to Africa. The reason is that these measures are not comprehensive they capture resource richness in either oil or minerals. However, as pointed out earlier, both oil and minerals are relevant for FDI to SSA. As a consequence, I use the share of minerals and oil in total exports, a measure that encompasses both types of resources. This measure of natural resources has been used in several studies, including Asiedu and Esfahani (2001) and Warner and Sachs (1995).
(ii) Openness of the Host Country : It is standard hypothesis that openness promotes FDI. In the literature, the ratio of trade to GDP is often used as a measure of openness of a country. This ratio is also often interpreted as a measure of trade restrictions.^10 As pointed out earlier, data on trade policy (e.g., tariff rates and trade taxes) is not readily available for most of the countries in
(^9) The dummy variable equals an index of real oil prices for oil exporters and it takes on value zero for non-oil exporters. (^10) Using trade/GDP as a measure of trade policy assumes that there is an inverse relationship between trade ratios and trade restrictions. This assumption is innocuous. Based on data for countries in Sub-Saharan Africa, Rodrik(1998) finds an inverse and robust relationship between several measures of trade policy and trade/GDP.
produced and sold in the local market.^13 Following the literature I include GDP per capita and GDP growth rates as measures of market attractiveness in my regressions.
4.2. Description of the Data The data was obtained from the World Bank’s World Development Indictors. Summary statistics of the variables are reported in Table 3. Table 4 compares the averages of FDI/GDP, openness and infrastructure development for the 1980s and 1990s.
Insert Table 3
Insert Table 4
Table 4 shows that the countries included in my sample improved their infrastructure and became more open in the 1990s. In addition, FDI to these countries increased in the 1990s. The trends in FDI, infrastructure and openness is similar to that observed for the whole region (Table 2), suggesting that my sample is a good representation of countries in the region.
5. Empirical Analysis The vast majority of the studies on the determinants of FDI to developing countries use pure cross-sectional data to determine the extent to which differences in host country characteristics explain the variation in FDI across country. These types of analysis answer the question: suppose country A and country B differ in variable X by one unit. What is the expected
(^13) The importance of the host country’s market depends on the type of FDI—whether market seeking or non-market seeking.seeking FDI serves domestic markets. As a consequence domestic demand is relatively more important for market Unlike non-market seeking FDI, where goods are produced in the local markets but sold abroad, market seeking FDI.
difference in FDI/GDP between the two countries? The focus of this paper is different. Specifically, the paper analyzes the impact of policy changes within country. Here, the relevant question is this: suppose country A increases variable X by one unit. What is the expected change in Country A’s FDI/GDP? An appropriate framework for such an analysis is the fixed- effects panel estimation. 14 This approach has two advantages. First, it allows the analyst to focus on changes within different units over time. Second, the estimates remain unbiased even when data is missing for some time periods for some cross-sectional units.^15 This advantage of a fixed- estimates estimation is particularly important for an analysis for SSA since data is not available for some years for several countries in the region. I start my analysis by determining the non- natural resource based variables that have a significant impact on FDI. The results reported in column (1) of Table 5 indicates that openness to trade, better infrastructure, higher income and better growth prospects promote FDI to Africa. These variables have been identified in several studies as important determinants of FDI to developing countries. In column (2), I include the share of fuel and minerals in total exports, NAT RESOURCE, to examine the effect of natural resource availability on FDI flows. As expected, the estimated coefficient of NAT RESOURCE is positive and significant. As indicated in Table 2, SSA has made some improvements (albeit marginal compared to other regions) in improving its physical infrastructure and trade openness. A natural question that arises is this: why has FDI not responded to these policy changes? To address this seemingly paradoxical issue I examine the impact of the policy-related variables, INFRAC and OPEN, over two time periods: 1980-89 and 1990-2000. Specifically, I
(^14) To the best of my knowledge only two papers on the determinants of FDI to developing countries employ a fixed- effects panel estimation (cf., Gastanaga et al (1998) andthe impact of human capital on FDI flows and report that their estimates for the fixed effect model “were poor”. Morisset (2000)). Noorbakhsh and Paloni (2002) examine They obtain better estimates using pooled-OLS. (^15) The unbalanced panel causes no problem if the missing data is not correlated with the idiosyncratic errors.
In this section I summarize the results, provide plausible explanations for the results and also discuss policy implications.
Result 1. Natural resource intensity, physical infrastructure and openness to trade have a positive impact on FDI to Africa. This suggests that (i) FDI to SSA is not solely determined by natural resource availability policy is important. (ii) In order to realize two of the most important benefits of FDI i.e., employment creation and technological spillovers countries in SSA need to improve their infrastructure and open their markets.
The significance of OPEN and INFRAC after controlling for natural resource availability suggests that FDI to SSA is not determined by an exogenous and uncontrollable factor and that policymakers can play an important role in directing FDI. The result also indicates that countries that open up their markets and improve their infrastructure will attract significant amounts of investments in non-extractive industries. Non-natural resource based FDI (e.g., investments in manufacturing and technologically intensive industries) is important to SSA because such investments foster technological spillovers (the process where domestic firms acquire the “superior” technology employed by foreign firms). In contrast, very little technology transfer occurs in extractive industries. The reason is that one of the most common and least expensive ways via which foreign technology gets diffused in host countries’ is through labor turnover, as domestic employees (especially employees in higher level positions) move from foreign firms to
domestic firms.^18 There are only a few (if any) domestically owned firms operating in extractive industries in SSA. As a result, technological spillovers are less likely to occur in such industries. Another advantage of non- natural resource based FDI over investments in extractive industries is that the latter requires very little labor hence it does not foster employment creation. This result is important because it suggests that in order to realize the full benefits of FDI, countries in SSA need to improve their infrastructure and open up their markets.
Result 2 : The increased investments observed in the 1990s were not driven entirely by better country conditions.
As pointed out in section 2, FDI to SSA increased in the 1990s (Table 2). Thus the fact that DUM90 remains significant after controlling for relevant country characteristics (specifically openness, infrastructure, natural resources and attractiveness of the host market) suggests that the increased FDI flows observed in the 1990s cannot be fully explained by changes in country conditions, in particular, better domestic policies. Indeed, the observed phenomena may be part and parcel of the FDI boom to developing countries that occurred in the 1990s (Table 2).
Result 3 : The estimated coefficients of the interaction terms, INFRACDUM90 and OPENDUM90 are negative and significant, suggesting that openness and infrastructure have become less significant in promoting FDI.
(^18) Bloom (1992) finds substantial technological transfer in South Korea when production managers left multinationals to join domestic firms. Indeed, foreign firms sometimes pay higher wages in order to retain theirworkers, and thereby prevent domestic firms from appropriating their superior technology [cf., Glass and Saggi (2001)].
very restrictive regimes prior to implementing reform. In the 1980s, most countries in SSA had very restrictive trade regimes, much more restrictive than countries in other regions. For example, over the period 1980-89, import duties (% total imports) for SSA averaged 26. percent. This compares with an average of 12.3 percent for East Asia and 11.2 percent for Latin America (World Bank, 2001).^20 Thus, countries in SSA may require drastic trade reform in order to attract FDI. However, as pointed out in section 2, liberalization did not go far enough. As a consequence, increased openness did not generate more FDI. An alternative explanation for the insignificance of openness in the 1990s is that widespread trade liberalization has caused foreign investors to completely discount the importance of openness. In this case, increased openness will have no impact on FDI.
Result 4 : INFRAC had a positive impact on FDI/GDP both in the 1980s and 1990s, however, the marginal impact of INFRAC was greater in the 1980s than in the 1990s. This suggests that in the 1990s, countries need to provide better infrastructure in order to receive investments at levels comparable to the 1980s.
The results follow from the fact that a one unit increase in infrastructure lead to a 1. percent increase in FDI/GDP in the 1980s. In the 1990s, however, more than a one unit increase in infrastructure is required in order to obtain a 1.12 percent increase in FDI/GDP. There are at least four plausible explanations. First, similar to openness, widespread improvement in the quality of infrastructure has made infrastructure availability a norm rather than an exception in the 1990s (Table 2). As a consequence, foreign investors have discounted the importance of (^20) Indeed, SSA seems even more restrictive when one compares data on export duties. The average export duties (% total exports) for SSA in the 1980s was 8.5 percent. This compares with an average of 1.7 percent for East Asia and2.1 percent for Latin America.
infrastructure. Second, the world has become more integrated and more competitive. This has increased the options available to foreign investors and made FDI more footloose. As a result, host countries need to provide better amenities, including infrastructure, in order to attract FDI in the 1990s. Third, as pointed out in section 2, the sectoral composition of FDI to SSA has changed in favor of investments in non-extractive industries. Note that physical infrastructure is more relevant for non-natural resource based investments.^21 As a consequence, host countries need to provide infrastructure of much better quality than the infrastructure available in previous years, in order to attract the “new” types of FDI. Fourth, as pointed out in Section 2, SSA’s global relative standing with respect to infrastructure declined in the 1990s. As a consequence, countries in the region had to provide better infrastructure in order to attract FDI at levels comparable to the 1980s.
7. Conclusion This paper has analyzed the impact of natural resources, infrastructure and openness to trade on FDI flows to Africa and examined why SSA has not attracted more FDI despite improvements in infrastructure and trade openness. The results indicate that the effect of infrastructure and openness on FDI has changed over time and the requisites for attracting FDI are higher than in previous years. Specifically, countries that liberalized trade and opened up their markets in the 1980s were rewarded with more FDI. In contrast trade reform did not generate a significant amount of FDI in the 1990s. Infrastructure development generated more FDI both in the 1980s and 1990s, however the marginal benefit from increased infrastructure was less in the 1990s.
(^21) For example, firms in extractive industries in SSA often locate in remote areas, which typically lack access to basic amenities such as electricity and water.
