Evaluation of FDI flows into the MENA Region Dr. Ahmed Kamaly*

*

Assistant Professor, Economics Department, The American University in Cairo. Address: The American University in Cairo, 113 Kasr El-Aini, P.O Box 2511, 11511 Cairo, Egypt. Email: [email protected]. Tel. +(02) 797-6786. Fax: +(02) 795-7565. 1

I. Introduction One of the main features characterizing the recent surge in capital flows to developing countries is its composition. Equity flows in the form of equity portfolio flows and foreign direct investment (FDI) constituted more than 80 percent of total flows. Taking a closer look at to the composition of capital flows, one can easily discern that the most dynamic and resilient component of capital flows, during the 1990s, was FDI. This settled behavior of FDI during good as well as bad times, as opposed to the observed procyclical pattern of other types of inflows, made FDI the most favorite “diet” among developing countries’ capital accounts, especially after the Asian crisis. However this exuberant movement in FDI did not proliferate into all countries or regions. The Middle East and North Africa (MENA) region did not secure for itself a significant share of FDI proportional to its size. The recent surge of capital flows in the 1990s witnessed the dominance of equity flows, mainly portfolio flows and FDI. Emerging markets in Latin America, East Asia and in transition economies absorbed large flows of capital in the form of equity flows. During the 1990s, capital flows were not restricted to a handful of countries. A wide range of developing countries has managed to attract a reasonable amount of flows relative to the size of their economies. However, the portion of FDI directed to the MENA region was small both in absolute and relative terms. The existing empirical literature on capital flows and their determinants is biased toward countries which have secured for themselves high level of capital flows. This could explain partially the paucity of studies dealing with capital flows directed to the MENA region. As for studies dealing with the determinants of FDI flows to the region, they are almost nonexistent due first to a general limitation within the literature, that is the tendency of the literature to treat capital flows as one coherent entity without breaking down its components; and second to the relative insignificance of the region when it comes to capital flows including FDI. The paper is organized as follows. Section two starts by a quick overview of the current literature on the determinants of FDI with an emphasis on the empirical studies and how they relate to the MENA region. This part discusses briefly the recent debate regarding the relative importance of global vis-à-vis domestic factors in driving capital flows. Section three of the paper examines the distribution and trend of FDI flows to countries in the MENA region during the 1990s. This is the analytical part of the paper where the level of FDI secured by the MENA region is compared and contrasted with other regions in world, such as East Asia and Latin America. Issues like the general trend of FDI, its variability, and distributional characteristics across time and countries are analyzed. At the end of this section, a reducedform model which makes use of the recent progress in panel model analysis, is estimated using data from the MENA region to gauge the chief macroeconomic factors affecting the flow of FDI to the region. Finally, section four concludes.

2

II. Brief Exposition of Relevant Literature This section attempts to shed some light on the empirical literature dealing with FDI and its determinants, with an emphasis on the literature dealing with FDI directed to the MENA region. This is a brief exposition of the literature and it should be regarded as a complete survey of the literature. Generally, one can divide the literature into two branches. The first considers the decision of the firm to engage in FDI. The second, which is more recent, discusses the forces that make countries attract capital flows, part of which is FDI. In this paper, we focus on the latter branch. This branch of the literature starts to develop with the resurgence of capital flows to developing countries in the beginning of the 1990s where a number of studies attempted to analyze the forces behind this phenomenon. These are, by and large, macro-based studies aiming at identifying the principal variables shaping the movements of capital flows to developing countries. In the beginning of the 1990s, this line of research seemed to be the natural result of the sudden unexpected1 surge in capital flows after the drought following the debt crisis in the 1980s. Most of the studies were aimed at explaining the forces behind this notable change in capital flows directed to developing countries. At that point, most of these studies did not find it worthwhile to differentiate between different types of capital flows, perhaps since initially the main types of capital flows, portfolio investment and FDI, were on the rise and followed more or less the same path. This is evident from the correlation coefficients between portfolio investment and FDI during the first half of the 1990s, as depicted in Table 1. However, this strong positive comovement between portfolio investment and FDI disappeared, and was even reversed, during the second half of 1990s (Table 1). Hence, aggregating the different types of capital flows, or using current account deficit, hides the variability of the various components of capital flows, making empirical results unreliable due to aggregation bias. Insert Table 1 Soon enough two distinct stories were proffered to explain this surge in capital flows: the push and the pull stories. According to the push story, external factors like international interest rates are the main driving force behind the recent surge in capital flows. On the other hand, according to the pull story, domestic factors are perceived to be more important than external ones in attracting capital flows. This debate over the relative importance of external versus domestic factors started with the seminal paper of Calvo, Leiderman and Reinhart (1993), which aimed at explaining the forces behind the surge in capital flows in Latin America in the beginning of the 1990s. The study pointed out the importance of external factors (US macro variables) in driving capital flows to the region. This result has a critical implication: sustainability of flows should not be taken for granted, as changes in external 1

After the end of the debt crisis in the second half of the 1980s, economists were very skeptical about the chances for developing countries in regaining access to the international capital market due to the deterioration in countries’ creditworthiness and poor economic performance. For example, in their study of private capital flows, Cardoso and Dornbusch (1989) wrote “The foreseeable future will be overshadowed by the continuing debt crisis”. 3

factors over which a host country has no control could bring about sudden swings in flows, endangering the stability of the host country financial market. Insert Figure 1 Despite the plethora of FDI flows during the 1990s, and especially after the Tequila crisis (see Figure 1), there have been a relatively limited number of studies that look closely into the forces directing FDI and shaping its distribution among developing countries. Some of these studies examine the effect of some particular policies on FDI. Singh and Jun (1995) gauge the effect of socioeconomic instability on FDI for the 1970-93 period. Gastanaga, Nugent and Pashamova (1998) assess the relative importance of host country reforms in driving FDI flows. The majority of the studies though looked at FDI as part of capital flows, like for example Hernandez and Rudolph (1995), Calvo and Reinhart (1998), and Lensink and White (1998)2. Other problems in the existing literature on FDI include the use of inadequate estimation techniques. A number of these studies did not take into account both time series and cross section dimensions of the data and ignored the dynamics that characterize FDI behavior. Edwards (1992) and Fernández-Arias and Hausmann (2000b) take the data as averages of the period covered. Claessens, Oks, and Polastri (1998) on the other hand, use panel analysis, but formulate their model as a static panel model with no dynamics. Nonetheless, FDI, as most macroeconomic variables, is a dynamic process and failing to take this into account could bias the obtained results. Other recent studies such as Singh and Jun (1995), Bathattachaary, Montiel, and Sharma (1997) and Gastanaga, Nugent and Pashamova (1998)3 use a dynamic model with a lagged endogenous variable, but apply fixed or random effects specifications leading to biased and inconsistent results, as detailed in the next section. Recently, Kamaly (2002) address some of these problems such as aggregation bias, endogeneity, and static representation of dynamic relations. Main results of this study indicate that first, FDI is highly persistent, which implies that the long-term effects of various factors on FDI are much greater than their short-term effects. Second, high international interest rates deter on FDI flows directed to developing countries. Third, non-traditional variables, such as democracy level and exchange rate variability do influence FDI in the expected directions. Finally, large devaluations have adverse effects on FDI flows. As for the literature on FDI to the MENA region, the size and the coverage of this literature is quite comparable to the scanty share of the MENA region in FDI relative to other regions. Very few studies have studied FDI or even capital flows directed to the MENA region. One of these studies is the recent study by Onyeiwu (2000), which examines the macroeconomic determinants of FDI outflows from the Arab world. This study uses data from ten Arab countries covering the period from 1987-1997. The paper applies Seemingly Unrelated Regression (SUR) method on a reduced-form model which has FDI outflows as its 2

For a recent survey of this literature, see Kamaly (2002). The dynamic panel model in Gastanaga, Nugent and Pashamova’s (1998) study suffers from yet another problem, since future real growth rate of GDP which clearly is neither a predetermined nor an exogenous variable is used in the regression without correcting for the possible endogeneity problem. 3

4

dependent variable. However, the model used is a static one as it does not include lagged dependent variable. Results show that only the interest rate, the exchange rate, and the lagged inflation rate carry the expected signs, but interestingly, the rate of growth of real GDP carries the opposite sign indicating that high real GDP growth is associated with more outflows! Other studies like El-Naggar (1990) and Bisat (1996) dealt with FDI as yet another type of investment where they stress the importance of maintaining solid fundamentals to elevate the investment level in the MENA region. El-Erian and El-Gamal (1997), on the other hand, distinguish between “good” and “bad” types of incentives to attract FDI that could be adopted in Arab countries.

III. Data, Statistical Analysis, and Results A. Analysis of the Data The data for this study is obtained from the World Economic Outlook (WEO). Data is available from 1990 till 1999 for the following eleven countries: Algeria, Egypt, Kuwait, Lebanon, Libya, Morocco, Oman, Syria, Tunisia, Turkey, and Yemen4. Table 2 shows some basic statistics including the mean and the standard deviation of FDI and the ratio of FDI to GDP5. (Insert Table 2) From this table, we notice the following. First, over the decade of the 1990s and on average, all countries except Kuwait show positive FDI flows. Second, all countries except Lebanon have managed to attract very little FDI in absolute term and also relative to their economies as indicated by FDI figures and the ratios of FDI to GDP respectively. Third, Lebanon is the country in the sample with the highest average FDI flows as well as the highest ratio of FDI to GDP. In fact, the average FDI directed to Lebanon is more than double the amount of FDI directed to the second highest FDI country. In addition, the ratio of FDI to GDP in Lebanon is more than 5 fold the same ratio in any other country in the region. Other countries like Turkey and Egypt have attracted modest level of FDI over the 1990s period; however, it is Tunisia that has the second largest ratio of FDI to GDP indicating that Tunisia attracted a significant amount of FDI relative to its economic size. Fourth, again Lebanon comes in first in terms of the highest standard deviation in both FDI flows and the ratio of FDI to GDP over the 1990s period. However correcting for the means, the countries that exhibit the highest level of variability in FDI and the ratio of FDI to GDP as measured by the 4

Originally, Saudi Arabia, Sudan, and United Arab Emirates were part of the sample but the FDI figures for these countries according to WEO database are all zeros, however, it is hard to imagine that there is no FDI flows coming from or to these countries, hence a decision was made to drop these countries from our sample. 5 It is quite misleading to compare countries in terms of FDI flows without referring to their respective economic sizes. It should not come as a surprise that China attracts much more FDI than Botswana. However if one is indeed interested in meaningful comparison between these two countries, or any other country for that matter, then one should control for the size of the economy by comparing the amount of FDI per a dollar of GDP 5

coefficient of variation are Yemen followed by Kuwait as appearing in Table 2. On the other hand, the country with the smallest variability in terms of FDI flows is Turkey. Insert Figure 2 If we examine the trend of the total FDI flows for those eleven MENA countries as shown in Figure 2, we notice that first, the sum of FDI flows in the second half of the 1990s is higher the one in the first half; however, this does not imply that there is a clear upward trend in FDI flows directed to the region similar to the one observed in Figure 1 associated with the sum of FDI directed to developing countries. Second, FDI flows directed to those MENA countries around the middle of the 1990s, where they reach a sum of $ 6.11 billion and $6.07 billion in 1994 and 1995 respectively. Third, there is an evidence for a slight upward trend in FDI flows starting from 1997 continuing till the end of our sample period. It would be interesting to see whether this trend has continued through the new millennium or not especially that FDI flows to developing countries have experience a regression in the last few years. Insert Table 3 Comparing the MENA region with other regions in the world in terms of the amount of FDI per a dollar of GDP, it becomes clear how trivial is the share of FDI in the economies of the MENA region. As appearing from Table 3, the average ratio of FDI to GDP for the eleven MENA countries over the 1990-decade is 1.76 percent. One should note that when excluding Lebanon from the sample, the ratio drops to a mere 0.67 percent. It is quite obvious that Lebanon is an outlier even relative to all developing countries. Even with this outlier included in the sample, the average FDI flows per dollar of GDP for the MENA region is less than the one of Latin America (2.0) and way less the one of East Asia (2.28).

B. Model and Estimation Procedure In this part, we analyze the factors influencing FDI flows to the MENA region using data from the previously mentioned eleven MENA countries. I adopt a reduced form approach with the ratio of FDI to GDP as the dependent variable. One should note that almost all6 the empirical literature on the determinants of FDI and private capital flows employs reduced-form equations that are not derived from a micro-founded theoretical model (Edwards 1992; Bathattachaary, Montiel, and Sharma 1997; Calvo and Reinhart 1998; Claessens, Oks and Polastri 1998 are examples of such studies). Singh and Jun (1995), and Hernandez and Rudolph (1995) assume a partial adjustment model for FDI flows without providing a 6

A notable exception is Fernández-Arias (1996). In his study, Fernández-Arias develops a theoretical model of private capital flows to developing countries based on a non-arbitrage condition between the investment return in a given developing country and the opportunity cost of making such an investment, i.e. the expected return in developed countries. Despite the fact that this model does not deal with a specific type of capital flow, but rather considers private flows in general, in reality its behavioral relations, describing returns and creditworthiness, imply that the model mainly applies to portfolio investment. In fact, the empirical part of this study focuses solely on portfolio investment, taking it as its dependent variable. 6

theoretical background or justification for such a specification. Furthermore, as argued by Edwards (1992), since there exists no unified accepted theory of FDI, any empirical study on FDI should adopt a pragmatic approach in selecting the explanatory variables to be included in the regressions. I assume that the FDI to GDP ratio follows the following data generating process: yit = α + µi + δyit −1 + xit' β + uit u i ~ iid (0, σ )

(1) i = 1,2,...N , t = 1990,....T

2 u

Where y it : FDI to GDP ratio

xit'

: The matrix of explanatory variables besides the lag dependent variable

N: Total number of countries T: End of the period (1999) µ i : Country individual effect. α , δ and β are unknown parameters u it : Error term

Here I assume that the error term uit follows a one-way error component model with constant variance σ u . This is a fixed-effects formulation with a lagged dependent variable. The reason for including the lagged dependent variable in the model is to take into account the dynamics associated with FDI flows as mentioned in the previous section. The fixed effects representation captures the fact that countries have individual specific effects such as institutional settings, geographical characteristics, and cultural norms, which do influence FDI but are fixed in the short and medium terms. 2

The dynamic panel model used in this study calls for a special type of estimation to deal with the problems generated by including the lagged dependent variable in the regression. In these models, if the usual panel regression techniques, such as the Within estimator, are used, the results could suffer from inconsistency and bias, especially if the time dimension is small and the cross section dimension is large as in our case (see Nickell, 1981; and Anderson and Hsiao 1981 for more details). The recent development of dynamic panel models has provided a number of solutions to deal effectively with this problem. Among the first suggested solutions is the one introduced by Arellano and Bond (1991) which is often referred to in the literature as the GMM-IV or the standard GMM estimator. This estimator makes use of the orthogonality conditions that exist between the instruments and the model after some type of transformation (e.g. first differences or orthogonal deviations). Performing GLS on this transformed model 7

produces the one-step GMM-IV estimator. If the error terms are heteroskedastic, the one-step residuals are used to obtain a more reliable estimate, the two-step GMM-IV estimator. This study, however, adopts another GMM estimation technique based on Arellano and Bover (1995), and Blundell and Bond (1998), which builds and improves upon the standard GMM estimator. In addition to the moment conditions associated with the standard GMM estimator, this estimator employs additional restrictions based on the

fact that any available instrument that is not correlated with the group effects µ i can serve as an instrument for (1) in levels without any transformation. Combining all these restrictions in one system and using GMM produces a more efficient estimator, the GMM-SYS estimator.

The transformation used in this study is orthogonal deviations7 as opposed to firstdifferencing. The advantage of this transformation is that autocorrelation between transformed errors will be absent if it is absent among the original errors. In fact, the orthogonal deviations procedure is equivalent to applying first, first-differencing transformation to get rid of fixed effects, and then using GLS to eliminate first degree autocorrelation resulting from first-differencing (Arellano and Honoré, 2000).

C. Empirical Results The initial list of the explanatory variables follows Kamaly (2002) and includes the lagged dependent variable, the lagged real GDP growth, a proxy for openness (the ratio of exports plus imports to GDP), and the weighted average bond yield in the Group of Seven (G7) countries8. Using data from 114 developing countries over the 1990-1999 period, Kamaly (2002) confirms that these above variables affect FDI flows in the anticipated directions. Also, this study points to the importance of inertia in driving FDI flows to developing countries. Starting with the same specification, I estimate (1) using the data of the eleven MENA countries over the 1990s period. The initial estimation of the model points to the significance of only the lagged dependent variable and the lagged real GDP growth. Only these two variables are retained together with the second lag of the real GDP growth which was found to have a significant effect on FDI flows. Insert Table 4

7

This is a transformation introduced by Arellano (1988), and Arellano and Bover (1995) where each observation is transformed into a weighted deviation from the average of future observations for the same group:

z i , t +1 + ... + z i ,T  z it* =  z it − T −t 

 T − t    T − t + 1  

12

t = 1,2,3..., T − 1 . Note that this transformation also standardizes the

variance. 8 See Kamaly (2002) for a justification for the use of this list of explanatory variables 8

Table 4 shows the results of estimating (1) using three different estimators: OLS, fixed-effects and finally, GMM-SYS. Statistically, all coefficients turn out to be at least significant at the 5 percent level, and carry the expected signs, except for the coefficient of the first lag of the real GDP growth in the case of the Within estimator. The most reliable estimators as mentioned before is the GMM-SYS which deals with the problems associated with the inclusion of the lagged dependent variable in the panel model as well as accounting for the existence of country-specific effects. In addition, the Wald test which assesses the joint significance of variables points to the significance of the explanatory variables. The GMM-SYS results show a trace of negative first order serial correlation and the absence of second order serial correlation among residuals, indicating that the disturbances are not serially correlated, and hence, establishing the consistency of those estimators. In addition, the Sargan test indicates the validity of the used instruments9. The results from the three estimators are quite comparable in terms of the magnitudes of the coefficients. However, focusing on the more reliable GMM-SYS estimates, one can notice first, the large coefficient on the lagged dependent variable (0.72) indicating that FDI exhibits a high degree of inertia. This high level of inertia also implies that the long-term effects10 of other explanatory variables are almost a quadruple of their point estimates (short-term effects). Second, the real growth of GDP has a significant effect on FDI flows to the MENA countries. According to the point estimates, a one percent increase in the real growth rate of GDP raises the ratio of FDI to GDP by approximately 0.13 percent in the next period and by 0.14 percent in the following one. These are the short-term effects; the long-term effect, however, is much larger amounting to an equivalent one percent increase in the ratio of FDI to GDP. This result points to the importance of economic growth in the MENA countries in attracting FDI to the region both in the short-term and the long-term. Next, we test the effect of a number of additional regressors on FDI flows to the MENA region. These additional regressors appear in Table 5. These regressors vary from the variability of the nominal exchange rate as a measure of the macroeconomic risk11 in the economy to democracy index, which assesses how democratic a country is. Insert Table 5 Insert Table 6 Table 6 reports the results obtained from tossing these potential explanatory variables into the base regression. In each of these different regressions, the three 9

See Kamaly (2002) for a detailed discussion of these tests.

The long-term impact of β is given by β (1 − δ ) −1 . 11 Although not stressed by the author, Wei’s (2001) results indicate that exchange rate volatility has a significant negative effect on FDI as a ratio of total flows (see last column of Table 10b in Wei, 2001). Also, studies by Frankel and Wei (1995), Rose (2000), and Frankel and Rose (2001) find evidence that exchange rate variability discourages trade in developing countries. My conjecture is that high exchange rate variability depresses FDI, as in the case of trade. 9 10

variables appearing in the base regression are retained and only one additional regressor is added, with its name appearing on the top of the column. Only the GMM-SYS estimator is used to estimate the different equations. As shown in Table 6, Wald test points to the joint significance of explanatory variables in all the estimated regressions. In fact, in all of these regressions the Wald test shows an improvement over the base regression. In general, the explanatory variables included in the base regression retained their significance and magnitudes in face of the inclusion of additional explanatory variables. The variable with the largest coefficient was, again, the lagged endogenous variable, which hovers around 0.7. This variable is always significant at the 1 percent level. This high statistical and quantitative significance, combined with the stability of point estimates, suggest the importance of inertia in driving FDI flows to the MENA region as well as the larger effect of the explanatory variables in the long-term relative to the short-term. As shown in Table 6, all the additional regressors took the correct signs; however none of them were statistically significant. These findings are somewhat different than the findings of Kamaly (2002) which found a systematic evidence indicating the significant effects of some of these variables on FDI flows directed to developing countries. One possible explanation for the insignificant of these additional regressors as well as openness and the international interest rate on FDI flows to the MENA region could be due to the fact that the MENA region is a relatively newcomer in the international capital market. Since the MENA region is a newcomer with a relatively low stock of FDI, FDI flows could be only sensitive to economic fundamentals surrogated by the real GDP growth. This view is analogous to the conventional wisdom related to the determinants of FDI which argues that FDI is more sensitive to host countries’ fundamentals- since it is a function of long-term profitability expectationsthan cyclical variables, such as the interest rate (World Bank 1997). To reconcile this finding with the one of Kamaly (2002), one could argue that when a typical developing country enters into the international capital market especially the FDI market, the one thing that matters the most is the country’s fundamentals. However, as this developing country attracts more and more FDI and the stock of FDI increases, the international capital market starts to look at other policy and non-policy variables which are bound to affect the profitability of investment when FDI flows (or stock) reach a certain level or threshold. Another possibility explanation is related to the sample itself. Given the limited size of the sample and the fact that most of the MENA countries share a number of characteristics and features, it is possible that the sample has limited variability which is reflected negatively on the significance of the explanatory variables.

10

IV. Conclusion During the last surge of capital flows in the 1990s, FDI was the main conduit through which capital was channeled to developing countries. As opposed to other types of capital flows, FDI has shown a remarkable stability even in tremors when other types of flows hasten away from capital importing countries. The MENA region, on the other hand, did not participate in this jamboree. The share of FDI collected by the MENA countries is very small in absolute terms and relative to the size of their economies. This paper attempts to shed some light on FDI flows directed to the MENA region as well as identifying the factors that influence their flow. The paper is an empirical in nature. It uses data from eleven different MENA countries over the 1990 period to assess the general trend of FDI, its variability, and distributional characteristics across time and countries. The paper confirms the scanty share of FDI directed to the MENA countries compared to other regions like for example, Latin America and East Asia. The trend of FDI to the MENA countries indicates that it does not follow the general trend of FDI directed to developing countries. The paper analyzes the flow of FDI to these eleven countries identifying their trend and variability. In addition, the paper looks at the ratio of FDI to GDP to compare between the MENA countries after correcting for the size of their economies. Basic statistics reveals the poor track of FDI and the ratio of FDI to GDP for almost all these countries with the notable exception of Lebanon which is regarded as a perfect outlier even among all developing countries. The study makes use of the recent development in dynamic panel model to estimate a reduced-form model aiming at identifying the critical factors affecting the flow of FDI to our sample of MENA countries. Results conform the traditional views on FDI that it follows more countries’ fundamentals than cyclical variables like the international interest rate. In addition, results point to the importance of inertia in driving FDI flows to these countries implying that the long-term effect of various factors on FDI is much greater than their short-term one. Other additional explanatory variables are found to carry the right signs but without being statistically significant. Some of these results, mainly the insignificance of openness, the international interest rate, the variability of the nominal exchange rate, and democracy measure, do not conform the results of Kamaly (2002) which looks at FDI flows to all developing countries. Two possible explanation where given to reconcile these differences. One is related to the nature of the MENA region as being a newcomer to the international market of capital flows; the other is related to the nature of the sample as having a low degree of variability. One obvious policy recommendation from this paper is the importance of having a solid set of fundamentals. It appears that for a newcomer to the international capital market, especially FDI market, maintaining a strong macroeconomy is key in attracting FDI flows.

11

REFERENCES

Agarwal, J (1980) “Determinants of Foreign Direct Investment: A Survey” Weltwirtschaftliches Archiv, 116, pp. 739-73. Ahn, S. and P. Schmidt (1995) “Efficient Estimation of Models for Dynamic Panel Data” Journal of Econometrics, 68, pp. 5-27. Anderson, T.W., and C. Hsiao (1981) “Estimation of Dynamic Models with Error Components” Journal of the American Statistical Association, 76, pp. 598-606. Antzoulatos, A. (2000) “On the Determinants and Resilience of Bond Flows to LDCs, 19901995” Journal of International Money and Finance, 19, pp. 399-418. Arellano, M. (1993) “On the Testing of Correlated Effects with Panel Data” Journal of Econometrics, 59, pp. 87-97. Arellano, M. and S. Bond (1991) “Some Tests of Specification for Panel Data: Monte Carlo Evidence and an Application to Employment Equations” Review of Economics Studies, 58(2) pp. 277-97. Arellano, M. and O. Bover (1995) “Another Look at the Instrumental Variables Estimation of Error-Components Models” Journal of Econometrics, 68, pp. 29-51. Arellano, M. and B. Honoré (2000) “Panel Data Models: Some Recent Developments” Working Paper 0016, Spain, www.cemfi.es Baltagi, B. (1995) Econometric Analysis of Panel Data, New York: Wiley. Bathattachaary, A. Montiel, P and S. Sharma (1997) “Private Capital Flows to Sub-Saharan Africa: An Overview of Trends and Determinants” in Z. Iqbal and R. Kanbur (eds.), External Finance for Low-Income Countries, IMF. Binder, M., Hsiao, C and M. Pesaran (2002) “Estimation and Inference in Short Panel Vector Autoregressions with Unit Roots and Cointegration”, www.inform.umd.edu/econ/mbinder/ Bekaert, G., Campbell, H., R. Lumsdaire (1999) “Structural Breaks in Emerging Capital Flows” NBER Working Paper 7219. Blonigen, B. (1997) “Firm-Specific Assets and the Link Between Exchange Rates and Foreign Direct Investment” American Economic Review, 87(3), pp. 447-65. Blöndal, S. and H. Christiansen (1999) “The Recent Experience with Capital Flows to Emerging Markets Economies” Economics Department Working Paper 211, OECD, Paris. 12

Blundell, R and S. Bond (1998) “Initial Conditions and Moment Restrictions in Dynamic Panel Data Models” Journal of Econometrics, 87, pp. 115-43. Bohn, H and L. Tesar (1996) “U.S Equity Investment in Foreign Markets: Portfolio Rebalancing or Return Chasing?” The American Economic Review, 86(2), pp. 77-81. Brock, W. and S. Durlauf (2001) “Growth Empirics and Reality” World Bank Economic Review 15(2), pp. 229-72. Cailloux, J and S. Griffith-Jones (1998) “International Capital Flows, What Do We Know About Their Volatility? Evidence From Latin American Countries” CESSEFI Working Paper 9711, Sorbonne, Paris. Calvo, G., R Sahary, and C.A Vegh (1995) “Capital Flows in Central and Eastern Europe: Evidence and Policy Options” IMF Working Paper 95/57. Calvo, G., L. Leiderman, and C. Reinhart. (1993) “Capital Inflows and Real Exchange Rate Appreciation in Latin America: The Role of External Factors”, IMF Staff Papers 40(1), pp. 10851. _________ (1996) “Inflows of Capital to Developing Countries in the 1990s” Journal of Economic Perspectives, 10(2), pp. 123-39. Calvo, G., and C. Reinhart. (1998) “The Consequences and Management of Capital Inflows” Lessons from Sub-Saharan Africa” Expert Group on Development Issues, 2. _________ (1999) “When Capital Inflows Come to a Sudden Stop: Consequences and Policy Options” www.bsos.umd.edu/econ/ciecalvo.htm. _________ (2000) “Fixing for Your Life” Paper presented for the Brookings Trade Forum 2000, Policy Challenges in the Next Millennium, April 27-28, 2000 Washington DC. Cardoso, E. and R. Dornbusch (1989) “Foreign Private Capital Flows” in H. Chenery and T.N. Srinivasan Handbook of Development Economics Vol. II, Elsevier Science Publishers. Chuhan, P. S. Claessens and N. Mamingi (1993) “Equity and Bond Flows to Asia and Latin America: The Role of Global and Country Factors” World Bank Policy Research Working Paper, 1160, July. Chuhan, P et al. (1996) “International Capital Flows: Do Shot-Term Investment and Direct Investment Differ?” Policy Research Working Paper 1669, World Bank, Washington DC. Claessens, S, M. Dooley and A. Warner (1995) “Portfolio Capital Flows: Hot or Cold” World Bank Economic Review, 9(1), pp. 153-74.

13

Claessens, S., Oks, D. and R. Polastri (1998) “Capital Flows to Central and Eastern Europe and Former Soviet Union” World Bank Working Paper 11/4/98. DPD Package Ver. 1.2 by Doornik, J. M. Arellano and S. Bond (2001), www.nuff.ox.ac.uk/users/doornik/ Dadush, U. Dhareshwar, A and R. Johannes (1994) “Are Private Capital Flows to Developing Countries Sustainable?” “Policy Research Working Paper 1397, World Bank. Diamonte, R., Liew, J. and R. Stevens (1998) “Political Risk in Emerging and Developed Markets ” in Richard Levich (ed.) Emerging Market Capital Flows. Kluwer Academic Publishers. Dooley, M., E Fernández-Arias, and K. Kletzer. (1996) “Is the Debt Crisis History? Recent Private Capital Inflows to Developing Countries” The World Bank Economic Review 10, pp. 2950. Dunning, J (1981a) “Explaining the International Direct Investment Positions of Countries: Towards a Dynamic or Development Approach” Weltwirtschaftliches Archiv, 117, pp. 30-64. Dunning, J (1981b) “A Note on Intra-industry Foreign Direct Investment” Banca Nazionale del Lavoro-Quarterly Review, 139, pp. 427-37. Edwards, S. (1992) “Capital Flows, Foreign Direct Investment, and Debt-Equity Swaps in Developing Countries” In H. Seibert (ed.) Capital Flows in the World Economy. Institut Für Weltwirtschaft an Der Univerität Kiel. Edwards, S. (1999) “How Effective are Capital Controls?” NBER Working Paper 7413. Eichengreen, B. (1998) “International Lending in the Long Run: Motives and Management” in Richard Levich (ed.) Emerging Market Capital Flows. Kluwer Academic Publishers. Feenstra, R (1999) “Facts and Fallacies About Foreign Direct Investment” in M. Feldstein (ed.), International Capital Flows, The University of Chicago Press. Fernández-Arias, E. (1996) “The New Wave of Private Capital Inflows: Push or Pull?” Journal of Development Economics 48, pp. 389-418. Fernández-Arias, E. (2000) “The New Wave of Capital Inflows: Sea Change or Tide?” Paper Prepared for the Seminar: The New Wave of Capital Inflows: Sea Change or Just Another Tide? Annual Meetings of the Board of Governors, Inter-American Development Bank and InterAmerican Investment Corporation, Working Paper 415. Fernández-Arias, E. and R. Hausmann (2000a) “Is FDI a Safer Form of Financing?” Paper Prepared for the Seminar: The New Wave of Capital Inflows: Sea Change or Just Another Tide? 14

Annual Meetings of the Board of Governors, Inter-American Development Bank and InterAmerican Investment Corporation, Working Paper 416. Fernández-Arias, E. and R. Hausmann (2000b) “Foreign Direct Investment: Good Cholesterol” Paper Prepared for the Seminar: The New Wave of Capital Inflows: Sea Change or Just Another Tide? Annual Meetings of the Board of Governors, Inter-American Development Bank and Inter-American Investment Corporation, Working Paper 417. Fernández-Arias, E., and P. Montiel (1996) “The Surge in Capital Inflow to Developing Countries: An Analytical Overview” The World Bank Economic Review 10, pp. 51-77. Frankel, J and A. Rose (2001) “An Estimate of the Effect of Common Currencies on Trade and Income” www.ksg.harvard.edu/fs/jfrankel Frankel, J and S. Wei (1995) “Emerging Currency Blocs” in H. Genberg (ed.) The International Monetary System: Its Institutions and Its Future, Berlin. Gastanaga, VC., J. Nugent and B. Pashamova (1998) “Host Country Reforms and FDI Inflows: How Much Difference Do They Make?” World Development, 26(7), pp. 1299-314. Griffith-Jones, Stephany (1998) Global Capital Flows: Should They Be Regulated?, MacMillan Press, London, Great Britain. Hansen, L.P. (1982) “Large Sample Properties of Generalized Method of Moments Estimators” Econometrica, 49, pp. 1377-98. Haque, N., Pesaran, M. and S. Sharma (1999) “Neglected Heterogeneity and Dynamics in CrossCountry Savings Regressions” mimeo. Hernandez, L., and H. Rudolph. (1995) “Sustainability of Private Capital Flows to Developing Countries” Policy Research Working Paper, 1518, October. Kamaly, A. (2002) “Essays on FDI Flows to Developing Countries” PhD Dissertation, Department of Economics, University of Maryland at College Park. Khan, M. and C. Reinhart, (eds.) (1995) Capital Flows in the APEC Region, IMF Occasional Paper 122. Kim, Y (2000) “Causes of Capital Flows in Developing Countries” Journal of International Money and Finance”, 19, pp. 235-53. Knight, M. (1998) “Developing Countries and the Globalization of Financial Markets” World Development, 26(7), pp. 1185-1200. Lehmann, A (2001) “Foreign Direct Investment in Emerging Markets: Income, Repatriations and Financial Vulnerabilities” mimeo. 15

Leiderman, L. and A. Thorne (1996) “The 1994 Mexican Crisis and Its Aftermath” in G. Calvo, M. Goldstein and E. Hochreiter (eds.), Private Capital Flows to Emeging Markets After the Mexican Crisis Institute for International Economics, Washington DC. Lensink, R. and H. White (1998) “Does the Revival of International Private Capital Flows Mean the End of Aid? An Analysis of Developing Countries’ Access to Private Capital” World Development, 26(7), pp. 1221-34. Lucas, Robert, B. (1993) “On the Determinants of Direct Foreign Investment: Evidence from East and Southeast Asia” World Development, 21(3), pp. 391-406. Marshall, M.G. and K. Jaggers (2000) Polity IV Project: Political Regime Characteristics and Transitions, 1800-1999, Dataset Users Manual, www.bsos.umd/cidcm/polity/showfiles.asp Manzocchi, Stefano (1999) Foreign Capital in Developing Countries, MacMillan Press, London, Great Britain. Montiel, Peter and Carmen Reinhart (1999) “Do Capital and Macroeconomic Policies Influence the Volume and Composition of Capital Flows? Evidence from the 1990s”, mimeo. Nickell, S. J. (1981) “Biases in Dynamic Models with Fixed Effects” Econometrica, 49(6), pp. 1417-26. Ridder, G and T. Wansbeek (1990) “Dynamic Models for Panel Data” Advanced Lectures in Quantitative Economics, Harcourt Brace Jovanovich, Academic Press. Rose, A (2000) “One Money, One Market? The Effect of Common Currencies on International Trade” Economic Policy, 0(33), pp. 7-33. Pesaran, H and R. Smith (1995) “Estimating Long-run Relationships From Dynamic Heterogeneous Panels” Journal of Econometrics, 68, pp. 79-113. Sargan, J. (1958) “The Estimation of Economic Relationships Using Instrumental Variables” Econometrica, 26, pp. 393-415. Schadler, S et al. (1993) “Recent Experience with Surges in Capital Inflows” IMF Occasional Paper 108, IMF. Schneider, F. and B. Frey (1985) “Economic and Political Determinants of Foreign Direct Investment” World Development, 13(2) pp. 161-75. Singh, H. and K. Jun (1995) “Some New Evidence on Determinants of Foreign Direct Investment in Developing Countries”, Policy Research Working Paper 1531, World Bank, Washington DC.

16

Taylor, M., and L. Sarno (1997) “Capital Flows to Developing Countries; Long and Short-term Determinants” The World Bank Economic Review 11, pp. 451-70. Temple, J. (1998) “Robustness Tests of the Augmented Solow Model” Journal of Applied Econometrics, 13, pp. 361-75. Torrisi, R. (1985) “The Determinants of Direct Foreign Investment in a Small LDC” Journal of Economic Development, 10(1) pp. 29-45. Turner, P (1991) “Capital Flows in the 1980s: A Survey of Major Trends” Bank of International Settlement Economic Papers 30. Vernon, R. (1966) “International Investment and International Trade in the Product Cycle” Quarterly Journal of Economics, 80(2) pp. 190-207. Wei, S. (2001) “Negative Alchemy? Corruption, Composition of Capital Flows, and Currency Crises” NBER Working Paper 8187. Williamson, J. (2001) “Issues Regarding the Composition of Capital Flows” Development Policy Review, 19(1), pp. 11-29. Windmeijer, F. (2000) “A Finite Sample Correction for the Variance of Linear Two-Step GMM Estimators” Institute for Fiscal Studies Working Paper W00/19. World Bank (1997). Private Capital Flows to Developing Countries: The Road to Financial Integration. Oxford: Oxford University Press. World Bank (2000) Global Development Finance. World Bank (2001) Global Development Finance.

17

Table 1: Correlation Coefficient between Different Types of Flows

Correlation between different types of flows FDI and Portfolio flows FDI and other investment Portfolio Flows and other investment

1990-99 0.044 -0.662 -0.250

1990-94 0.995 -0.777 -0.782

1995-99 -0.174 -0.933 0.181

Table 2: Basic statistics on FDI in the MENA countries in the 1990s

Algeria Egypt Kuwait Lebanon Libya Morocco Oman Syria Tunisia Turkey Yemen

Average FDI Average (billion $) FDI/GDP (%) 0.14 0.30 0.64 1.18 -0.27 -1.11 1.44 12.50 0.11 0.33 0.54 1.67 0.10 0.76 0.09 0.61 0.37 2.22 0.65 0.39 0.10 0.24

SD FDI 0.22 0.32 0.95 1.08 0.03 0.40 0.05 0.07 0.17 0.09 0.40

Coefficient of variation of FDI SD FDI/GDP 0.46 1.52 0.85 0.49 3.80 -3.45 10.42 0.75 0.10 0.28 1.10 0.75 0.38 0.52 0.51 0.80 1.00 0.44 0.09 0.15 3.63 4.13

Table 3: The Average FDI to GDP in Different Regions over the 1990s Region MENA MENA without Lebanon East Asia Latin America

Average FDI to GDP ratio (%) 1.76 0.676 2.28 2.0

Coefficient o variation of FDI/GDP 1.53 0.72 -3.43 0.83 0.32 0.66 0.50 0.83 0.45 0.22 15.38

Table 4: Base Regression (Alternative Estimators) Explanatory variables

OLS

Constant

-0.053 (-0.167) 0.782 (29.7)*** 0.075 (2.98)*** 0.08 (3.03)*** 97 11 19670***

Lagged FDI/GDP Lagged real GDP growth Lagged twice GDP growth Number of observations Number of countries Wald χ 2 R2 Sargan’s overidentifying test (P-Value) A & B test of 1st order autocorrelation (P-Value) A & B test of 2nd order autocorrelation (P-Value)

0.67

GMM-SYS (Robust)

FixedEffects (Robust) .

.

0.414 (10.0)*** 0.079 (1.59) 0.13 (3.08)*** 97 11 1896***

0.716 (16.5)*** 0.126 (4.06)*** 0.140 (2.07)** 97 11 543.3***

0.5

. 5.532 (1.0) -1.145*** (0.25) 0.912 (0.326)

Table 5: List of Potential Explanatory Variables Variable Nominal exchange rate variability Democracy1 Rate of change of the nominal exchange rate Financial deepening (M2/GDP) Creditworthiness variables2

1

Expected Sign Negative Positive Ambiguous Positive Negative

This index varies from –10 (strongly autocratic political system) to 10 (strongly democratic political system). This index is computed by subtracting an indicator of autocracy from another indicator of democracy. The details of the construction of these indicators are found in the Polity IV Project document by Marshall and Jaggers (2000). 2 These variables are the ratios of long-term debt to GDP, total debt to GDP, and the difference between total debt and international reserves to GDP

Table 6: FDI and Additional Explanatory Variables Dependent variable: FDI/GDP Explanatory variables

Lagged Exchange Rate Variability

Lagged Democracy Index

Lagged FDI/GDP

0.693 (12.5)*** 0.132 (4.47)*** 0.14 (2.12)** -11.758 (-1.35) 95

Lagged once real GDP growth Lagged twice real GDP growth Additional regressor Number of observations Number of countries

Wald χ

2

Sargan’s overidentifying test (P-Value) A & B test of 1st order autocorrelation (P-Value) A & B test of 2nd order autocorrelation (P-Value)

Lagged M2 to GDP)

Lagged Long-term Debt to GDP

07.08 (15.2)*** 0.117 (2.61)** 0.134 (2.70)*** 0.002 (0.416) 97

Lagged Change in Nominal Exchange Rate 0.725 (12.5)*** 0.110 (2.42)** 0.132 (2.20)** 0.141 (0.449) 96

0.646 (8.17)*** 0.106 (1.91)* 0.134 (2.16)** 0.0194 (0.736) 87

0.7289 (11.9)*** 0.125 (3.94)*** 0.135 (2.12)** -0.023 (-0.601) 97

11 244.9***

11 529.1***

11 500***

10 307.1

11 449.1

5.40 (1.00) -1.173 (0.241)

5.093 (1.00) -1.131 (0.258)

4.808 -1.127 (0.26)

5.115 (1.00 -1.137 (0.255)

4.60 (1.00) -1.127 (0.260)

0.9059 (0.365)

0.9033 (0.366)

1.013 (0.311)

0.9208 (0.357)

0.9158 (0.360)

Notes: t-statistics are in brackets. *** Significant at 1% level or more, ** significant at 5% level or more, * significant at the 10% level or more. Robust specifies that the White estimator of variance is to be used in place of the traditional calculation of variance. This variance estimator produces consistent standard errors even if the data are weighted or the residuals are not identically distributed

Figure 1: Trends and Rate of Change of Capital Flows Components 1 6 0 .0 0

6 0 .0 %

1 4 0 .0 0

5 0 .0 %

1 2 0 .0 0

4 0 .0 %

$billion

1 0 0 .0 0 3 0 .0 %

FDI

2 0 .0 %

G r o w th R a te o f FDI

8 0 .0 0 6 0 .0 0 1 0 .0 %

4 0 .0 0

0 .0 %

2 0 .0 0

1999

1998

1997

1996

1995

1994

1993

1992

1991

1990

1989

-1 0 .0 % 1988

0 .0 0

Year

120

200.0%

100

150.0% 100.0%

$ billion

80

Portf. Inv. 50.0%

60 0.0%

Growth Rate of Portf. Inv.

40 -50.0%

1999

1998

1997

1996

1995

1994

1993

1992

1991

-150.0% 1990

0 1989

-100.0%

1988

20

Ye a r

300.0%

100 80

200.0% 60 40 100.0%

0.0% 1999

1998

1997

1996

1995

1994

1993

1992

1991

1990

-20

1989

0 1988

$ billion

20

Growth rate of other Inv. -100.0%

-40 -60 -200.0% -80 -100

Other Inv.

-300.0%

Figure 2: Sum of FDI Flows Directed to the Eleven MENA Countries

7 6

4

Total FDI 3 2 1

Years

1999

1998

1997

1996

1995

1994

1993

1992

1991

0 1990

$ billion

5