Policy Research Working Paper 9103 The External Wealth of Arab Nations Structure, Trends, and Policy Implications Mahmoud Mohieldin Ahmed Rostom Chahir Zaki Office of the Senior Vice President for The 2030 Development Agenda United Nations Relations and Partnerships January 2020 Policy Research Working Paper 9103 Abstract The paper makes two main contributions. First, it analyzes is positively and statistically significantly associated with net foreign assets and liabilities in selected Arab coun- foreign direct investment in Arab countries. Financial tries over the past two decades, emphasizing the relative development (defined as credit to the private sector as a significance of direct versus portfolio investment. It distin- percentage of gross domestic product) is also statistically guishes between foreign direct investment, portfolio equity significant across various regressions. The more financially investment, official reserves, and external debt. Second, the developed a country is, the more it should invest in riskier paper examines the effects of policy variables that affect assets, such as portfolio assets. But Arab investors are more the accumulation of net foreign assets and its components, risk averse than investors elsewhere. Oil-exporting countries analyzing how the existence of a sovereign wealth fund, tend to invest more in debt assets than in portfolio assets. the country’s exchange rate regime, and the development For oil-importing countries, financial development is the of its financial system affect its net foreign assets. The main most important determinant of foreign direct investment. findings show that the presence of a sovereign wealth fund This paper is a product of the Office of the Senior Vice President for The 2030 Development Agenda, United Nations Relations and Partnerships. It is part of a larger effort by the World Bank to provide open access to its research and make a contribution to development policy discussions around the world. Policy Research Working Papers are also posted on the Web at http://www.worldbank.org/prwp. The authors may be contacted at arostom@worldbank.org. The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent. Produced by the Research Support Team The External Wealth of Arab Nations: Structure, Trends, and Policy Implications1 Mahmoud Mohieldin,2 Ahmed Rostom,3 and Chahir Zaki4 Keywords: Net Foreign Assets (NFA), Foreign Direct Investment (FDI), External Debt, Sovereign Wealth Funds (SWFs), Arab economies. JEL classification: F41, O11. 1 The findings, interpretations, and conclusions expressed in this paper are those of the authors in their personal research capacities. They do not represent the views of the World Bank Group, its affiliated organizations, the Executive Directors of the World Bank, or the governments they represent. The authors wish to thank Mr. Rabah Arezki and the participants of the Middle East Economic Association (MEEA) Annual Conference held in Philadelphia; January 2019 for helpful comments and suggestions. They also wish to thank Ms. Mariam Hoda El Maghrabi and Mr. Ramy Zeid for excellent research support. 2  The World Bank Group and Faculty of Economics and Political Science, Cairo University.  3  The World Bank Group and George Washington University.  4  Faculty of Economics and Political Science, Cairo University and the Economic Research Forum.  1. Introduction Over the past two decades, global economic cycles were coupled with remarkable dynamics in cross-border holdings of foreign assets These holdings altered the structure and composition of flows of foreign direct investment (FDI) and portfolio flows to emerging economies. In addition, the recent downturn in oil prices severely affected the realized surpluses of resource-rich countries, particularly the Gulf Cooperation Council (GCC) countries, which suffered significant deterioration in their fiscal balances. Fiscal accounts swung from a surplus equivalent to 3.8 percent of GDP in 2014 to a deficit estimated at $147 billion (10 percent of GDP) in 2015 and $162 billion (7 percent of GDP) in 2016 (IIF 2015). In the absence of flexible exchange rates, oil revenues in local currency terms have fallen sharply. At the same time, political uncertainty in Arab countries that underwent political and social transformation reduced investors’ interest. These factors gave rise to large external imbalances in Arab countries. They revived interest in analyzing the pattern and assessing the adjustment mechanism as well as the dual role played by exchange rate regimes in influencing both net capital flows and net capital gains in external holdings in Arab countries. Net foreign assets (NFA) are the value of overseas assets owned by a country minus the value of its domestic assets owned by foreigners, adjusted for changes in valuation and exchange rates. A country’s NFA position can also be defined as the cumulative change in its current account over time. The NFA position indicates whether a country is a net creditor or debtor to the rest of the world. A positive NFA balance means that it is a net lender; a negative NFA balance indicates that it is a net borrower (Bouchet, Fishkin, and Goguel 2018). The Arab region is of particular interest for four main reasons:  Since the oil boom of the mid-1970s, oil and gas revenues in many Arab countries have been substantially larger than their import needs. Consequently, their foreign asset accumulations have been far greater than those of midsize oil-exporting countries, such as Algeria.  Political and social unrest in many Arab Spring countries resulted in greater fiscal spending and a decline in private investment, leading to internal and external imbalances in many countries.  It is important to understand how Arab countries allocate their assets (mainly oil-exporting countries’ assets from oil revenues) and liabilities (mainly oil-importing ones with oil imports representing the key liability).  This topic is timely for the Arab region, as some countries are creating sovereign wealth funds (SWFs), which may help them increase their NFA. Following the methodology of Lane and Ferretti (2001, 2007), this paper analyzes trends in and the composition of net foreign assets and liabilities in selected Arab countries over the past two decades, with an emphasis on the relative significance of direct versus portfolio investment. 2   It distinguishes between foreign direct investment (FDI), portfolio equity investment, official reserves, and external debt. The paper examines the effect of different policy variables that affect the accumulation of NFA and its components, analyzing how the existence of an SWF, the country’s exchange rate regime, and the development of its financial system affect its NFA. The paper is organized as follows. Section 2 reviews the literature on NFA and their policy determinants. Section 3 provides background information on the countries relevant to this study, identifying attributes unique to each country, and the composition and trends of the NFA. Section 4 presents the methodology and econometric specification. Section 5 analyzes the results. Section 6 will provide policy recommendations and proposals for future research. 2. Literature Review Empirical studies have been conducted on the measurement, acquisition, and management of countries’ NFA. The assessment of such composition is beneficial, as it is indicative to understanding a country’s external assets and liabilities to ultimately estimate its external position. Lane and Milesi-Ferretti ((2001) maintain that substantial changes reported in the composition of capital flows reflect the increasing role of foreign direct investment (FDI) and equity flows. They show that net flows of portfolio equity and FDI, changes in debt associated with assets and liabilities, and net foreign account transfers and exports were key factors in the changing composition of the current account. Vermeulen and de Haan (2014) examine the relationship between the composition of NFA and financial development. They assess the validity of the claim suggested by Mendoza and Smith (2006) that financially developed countries are likely to take more risks than developing countries, because development allows them to insure against risk. Financially developed countries thus tend to have positive net equity positions, whereas developing countries tend to have positive net debt positions. Vermeulen and de Haan show empirically that financial development reduces a country’s long-run net foreign position and that development leads to higher net equity and lower net debt positions. It is therefore advantageous for a developing country to focus on deepening its financial markets in pursuit of higher NFA and consequently development and growth. Henderson and Rogoff (1982) examine the relationship between a country’s NFA position and the stability of its exchange rate regime. A negative NFA position has been associated with unstable regimes, particularly flexible exchange rates. Henderson and Rogoff construct a two- country portfolio model to investigate the stability problem associated with a negative NFA position. They conclude that such a position does not on its own indicate instability. However, in conjunction with static or regressive expectations, instability can arise. Furthering the investigation of stability, Ghironi and Stebunovs (2007) study the domestic and international effects of financial deregulation. Using decreases in monopoly power of financial intermediaries as proxies for deregulation, they reach several conclusions. First, an economy that deregulates experiences a rise in growth, real exchange rate appreciation, and a deficit in the 3   current account. As a result, higher consumption as well as an increase in the number of domestic producers will be observed in economies that lags in deregulation. Second, a decrease in monopoly power leads to less volatile business creation, reduced countercyclicality, and weaker substitutions in labor supply to alleviate the adverse effects to productivity shocks. Third, financial deregulation contributes to a moderation of firm-level and aggregate output volatility; financial ties between two countries therefore reduce volatility in the foreign economy. The study claims that its findings are consistent with findings drawn from data following the deregulation of banks in the United States beginning in 1977. The model of exchange rate determination of Obstfeld and Rogoff (1995) incorporates the role of current account imbalances. However, despite its initial motivation, the paper became fundamental to further research on the relation between the exchange rate and net NFA accumulation, deemphasizing the role of the latter, due to the non-stationarity revealed within the developed model. In the same line, Cavallo and Ghironi (2002) revisited the former with emphasis on the original intent; that is, the authors develop a two-country model of exchange rate determination in which stationarity and foreign asset composition play an explicit role. They explore whether the exchange rate is determined less by the exogenous money supply and more by net foreign assets. They find that the flexible-price model delivers exchange rate appreciation under a relatively favorable productivity shock; the home economy accumulates assets rather than debt. The sticky-price model generates appreciation, net foreign debt, and stock market expansion after a permanent favorable shock to relative productivity. The study proposes further research on the implications of the developed model as well as the consequences of deviations from purchasing power parity. It also suggests exploration of the performance of policy rules. Debelle and Galati (2007) examine the consequences of the current account dynamics for the exchange rate. They show that over the past 30 years, episodes related to current account adjustments in industrial countries were associated with major slowdowns in economic growth and large exchange rate deprecations; the effects on capital flows were indeterminate. The study explains such adjustments as responses to domestic imbalances, particularly within the private sector. At the Middle East and North Africa region level, Lane and Milesi-Ferretti (2001) reach three main findings. First, this region is extremely heterogeneous: Gulf countries hold mainly external assets, and other countries have large external liabilities. Second, in contrast to developing countries, the role of stocks of equity is increasing in Middle Eastern countries as opposed to developing ones. Third, NFA are significant, because they indicate changes in real exchange rates in the long run, as suggested by the degree of depreciation of exchange rates in debtor countries relative to creditor countries as evidence of their theory. As per cross-border investment and foreign assets in the Arab region, Elbadawi and others (2018) study the allocation decisions of SWFs in terms of both the probability of investing abroad (the extensive margin) and the level of investment (the intensive margin). They find that foreign investors have a positive bias for Arab destination countries at the extensive margin level but a negative bias against them at the intensive margin. 4   3. Data and Stylized Facts This paper uses data from the International Monetary Fund–International Financial Statistics (IMF-IFS) database. Lack of data forced the exclusion of Algeria, the Comoros, Lebanon, Libya, Oman, Qatar, Somalia, and the United Arab Emirates. Many data points were also missing for other countries, including Djibouti, Morocco, Saudi Arabia, the Syrian Arab Republic, and Tunisia. The countries’ foreign assets and liabilities grew rapidly over the past four decades, increasing from about 20 percent of GDP in 1970 to 205 percent in 2016, as is shown in figure 1. NFA rose from about 20 percent of GDP in 1970 to about 90 percent in 1994 before falling to 40 percent in 2015 (figure 2). Figure 1 Global ratio of total assets and Figure 2 Global ratio of net foreign assets to liabilities to GDP, 1970–2015 GDP, 1970–2015 2.5 0 Percent of GDP 1970 1973 1976 1979 1982 1985 1988 1991 1994 1997 2000 2003 2006 2009 2012 2015 2.0 Net foreign assets as ‐20 percent of GDP 1.5 ‐40 1.0 ‐60 0.5 0.0 ‐80 1970 1973 1976 1979 1982 1985 1988 1991 1994 1997 2000 2003 2006 2009 2012 2015 ‐100 Total assets/GDP Total liabilities/GDP Source: International Monetary Fund– Source: International Monetary Fund– International Financial Statistics (IMF-IFS) database. International Financial Statistics (IMF-IFS) database. Net foreign assets and/or liabilities are seen as the reflection of a country’s external position and trade and investment with the rest of the world. Central banks traditionally define NFA as the change in net foreign assets as the current account balance. This definition is a bit narrow and more of a short-term nature. A country’s external sector also involves portfolio investments, direct investments as well as debt instruments. This broader context is more reliable. It is captured in the definition proposed by Lane and Ferretti (2001), which relies on two identities. The first identity is NFA = FDIA + EQA + DEBTA + FX – FDIL – EQL – DEBTL (1) where FX = foreign exchange reserves, FDI = the stock of direct investment, EQ = portfolio equity investment, DEBT = debt, A = assets, and L = liabilities. 5   The second identity is CA = (∆FDIA – ∆FDIL) + (∆EQA – ∆EQL) + (∆DEBTA – ∆DEBTL) + ∆FX – ∆KA – EO (2) where CA = the current account, ∆ = flows, ∆KA =capital account transfers, and EO = errors and omissions (assuming this reflects changes in debt assets held by residents abroad). Lane and Ferretti build on the capital flight literature. Disregarding asset valuation changes, they approximate the change in NFA with the current account balance, net of capital account transfers: ∆NFA ≅ CA + ∆KA. (3) According to Lane and Ferretti (2001), given an initial stock of net external assets, the current stock of net external assets can be estimated by cumulating the current account balance, net of capital transfers. The composition of NFA can be obtained by cumulating the relevant flows on the right-hand side of equation 2. Lane and Ferretti claim that their measure represents an improvement over the crude NFA measure along several dimensions. First, they use direct stock measures, rather than cumulative flows, for foreign exchange reserves and gross external debt (with respect to developing countries) and adjust the NFA measure correspondingly, allowing them to take into account the impact of debt forgiveness and reduction agreements, cross-currency fluctuations, misreporting of capital flows, and capital gains and losses on foreign reserves on NFA. Second, they allow, albeit imperfectly, for the impact of changes in relative prices across countries on the value of FDI assets and liabilities, as well as for the impact of variations in stock market prices on portfolio equity stocks. The impact of these adjustments is substantial: For some industrial countries, the correlation between changes in NFA and the current account is zero or negative. As Lane and Ferretti (2007, updated in 2015) note, large discrepancies in the measurement of a country’s foreign asset and liability position are observed. Hence, to account for such a heterogeneity, multiple indicators of the degree of financial globalization are needed to assess the composition of NFA. Lane and Ferretti focus on the global level. This paper focuses on the subset of Arab countries (see figures 3–10). The countries selected reveal wide disparities, as they include both oil exporters and oil importers as well as countries that experienced significant political, economic, and social instability and countries that did not. Adopting a comprehensive data set is critical to understanding the NFA dynamics of these countries and providing accurate comparative analysis that yields relevant policy recommendations. The Arab oil-exporting countries of Saudi Arabia, Kuwait, Bahrain, and Algeria maintained positive net total assets positions over the past two decades (figure 3). Iraq achieved a 6   positive total asset position only after recovering from the Gulf War. The total foreign assets of these countries rose sharply during the late 1990s; that growth slowed, particularly in the past five years, in what seems to be a response to the volatility of and decline in oil prices and the uncertainty caused by the Arab Spring. Non-oil-producing countries (the Arab Republic of Egypt, Jordan, Morocco, and Tunisia) maintained positive net total liabilities positions over the past two decades. Total liabilities in these countries rose sharply over this period, with total assets falling, widening the gap between total liabilities and total assets to 3: 1 over the past 10 years. Figure 3 Net foreign assets as percent of GDP, by country, 2007 and 2015 by country 6 4 2 0 -2 E R M DJI ZA GY RQ OR T T BN LBY AR R MN QAT SAU DN SYR TUN BG EM AR BH CO D E I J KW L M M O S W Y 2007 2015 Source: Authors, based on data from Lane and Feretti 2015. Figure 4 Current account as percent of GDP, by Figure 5 Capital account as percent of GDP, by country, 2007 and 2015 country, 2007 and 2015 by country by country .15 .5 .1 .05 0 0 -.05 -.5 E R M JI ZA Y Q R T N BY R RT N AT U N R N G M E R M JI ZA Y Q R T N BY R RT N AT U N R N G M AR BH CO D D EG IR JO KW LB L MA M OM Q SA SD SY TU WB YE AR BH CO D D EG IR JO KW LB L MA M OM Q SA SD SY TU WB YE 2007 2015 2007 2015 Source: Authors, based on data from Lane and Feretti 2015. Source: Authors, based on data from Lane and Feretti 2015. In order to understand the dynamics of the foreign sector in the oil versus the non-oil sector, we also look at changes in NFA and its key components over the period 2007–15. With the exception of Libya, which has been in severe conflict, oil-exporting countries increased the share of NFA in GDP. The surplus in the current account as a percent of GDP plummeted, except in Iraq. Only Bahrain and Qatar achieved growth in the capital account balance as a percent of GDP; all other oil-exporting countries suffered declines. This simple elevator analysis confirms that the NFA positions of oil-exporting countries are under extraordinary pressure. The change in NFA to GDP was negative in most non-oil- exporting countries, with the exception of Jordan and the Comoros, where it soared, driven mainly 7   by surges in the current account. Egypt, Lebanon, and Djibouti recorded positive changes in their capital accounts as percentage of GDP, but the increases were not sufficient to yield a positive change in the NFA position as a percent of GDP (table 1). Table 1 Change in net foreign assets, current account, and capital account as percent of GDP between 2007 and 2015, by country Country Net foreign assets Capital Capital account transfers account Non-oil-exporting countries Djibouti - - + Mauritania - - - Lebanon - - + Algeria - - Na Jordan + + - Tunisia - - + Sudan + - - Yemen, Rep. - + - Egypt, Arab Rep. - - + Morocco - + - Comoros Islands + + - - - Na Oil-exporting countries Libya - - Na Oman + - - Saudi Arabia + - - Iraq + + - Bahrain + - + Kuwait + - - Qatar + - + Source: Authors, based on Lane and Feretti 2015. Note: ‘Na’ refers to data that not available. 8   Another key component of the external sector position is the FDI position. In all countries studied except Kuwait and Bahrain, FDI assets exceeded liabilities during the past two decades (in Kuwait liabilities exceeded assets; Bahrain maintained a balanced position) (figure 6). Saudi Arabia maintained a large gap between direct investment liabilities and assets. The growth in total direct liabilities slowed in Egypt, Jordan, and Morocco during the early 2000s, which were coupled with the early wave of the global financial crises at the time. Figure 6 Net foreign direct investment as percent of GDP, by country, 2007 and 2015 by country .5 0 -.5 -1 -1.5 E R M JI ZA Y Q R T N BY R RT N AT U N R N G M AR BH CO D D EG IR JO KW LB L MA M OM Q SA SD SY TU WB YE 2007 2015 Source: Authors, based on Lane and Feretti 2015. The net portfolio debt positions indicate that oil exporters are net portfolio investors abroad and net debt asset holders (figure 7). Non-oil-exporting countries, especially Egypt, witnessed volatile net portfolio liabilities positions. Morocco was able to maintain a stable portfolio liabilities position over the past eight years, indicating investor trust in the economy. Figure 7 Portfolio equity and foreign direct investment holdings as percent of GDP, by country, y 2007 and 2015 g by country 3 2 1 0 -1 -2 E R M JI ZA Y Q R T N BY R RT N AT U N R N G M AR BH CO D D EG IR JO KW LB L MA M OM Q SA SD SY TU WB YE 2007 2015 Source: Authors, based on Lane and Feretti 2015. 9   With regard to the debt position, debt liabilities rose steeply in Egypt and Morocco in the past two decades. This uptrend, coupled with a downtrend in the portfolio liabilities position, represents a shift in foreign investors’ preference for holding sovereign debt rather than equities in both countries. Hence, investors are becoming more risk averse evidently during late 2000s. Net debt as a share of GDP is consistent with the earlier results on non-oil-exporting countries’ maintenance of their net borrower position between 2007 and 2015. Egypt, Tunisia, Jordan, Lebanon, Morocco, and Mauritania ended up with larger net debt positions in 2015 than in 2007, indicating a higher propensity to borrow foreign currency. Saudi Arabia, Bahrain, and Oman extended more debt financing to international borrowers in 2015 than in 2007. Libya and Algeria closed 2015 as net lenders. Their lower net position than in 2007 reflects growing domestic financing needs. Almost all oil-exporting countries closed 2015 with a higher ratio of reserves to GDP than in 2007; exceptions were Algeria, Bahrain, and Iraq, which suffered slight declines (figure 8). Non-oil-exporting countries ended up with higher ratios of reserves to GDP than several oil-exporting countries (exceptions are Libya and Saudi Arabia). This pile-up of reserves in non- oil-exporting countries coupled with higher levels of total liabilities, including foreign borrowing, FDI liabilities, and excessive leveraging (figure 9). Reserve cumulation in non-oil-exporting countries is costly, given the sources financing it. As a consequence, the scale of financial integration in oil-exporting countries was greater in 2015 than in 2007 and significantly greater than in non-oil-exporting countries. Figure 8 Net debt as percent of GDP, by country, 2007 and 2015 by country 1 0 -1 -2 -3 -4 E R M JI ZA Y Q R T N BY R RT N AT U N R N G M AR BH CO D D EG IR JO KW LB L MA M OM Q SA SD SY TU WB YE 2007 2015 Source: Authors, based on Lane and Feretti 2015. 10   Figure 9 Reserves as percent of GDP, by country, 2007 and 2015 Reserves to GDP by country 2 1.5 1 .5 0 E R M JI ZA Y Q R T N Y R T N T U N R N AR BH CO D D EG IR JO W LB LB A MR M QA SA SD SY TU BG YEM K M O W 2007 2015 Constructed by the authors using Lane and Feretti (2007) Source: Authors, based on Lane and Feretti 2015. After presenting the structure and trends of NFA in Arab countries, it is important to see how such variables are correlated with macroeconomic determinants. Table 2 shows the interaction between NFA and a series of policy variables, including whether a country has a more developed financial system, an SWF, or an intermediate exchange rate regime (a regime between a fixed and floating one). Countries that are more financially developed and countries with SWFs are more likely to have a better allocation of their assets. Yet, countries with an intermediate exchange rate regime have more uncertainty in their exchange rate policy than countries with fixed peg or flexible regimes, which can adversely affect their external positions. Table 2 highlights these findings. For most of the variables (for both assets or liabilities), countries with an SWF, with a more developed financial system, and without an intermediate exchange rate regime are financially more integrated with the rest of the world. This finding was the key motivation behind including these policy variables in our empirical estimation. 11   Table 2 Financial integration and its policy determinants (millions of US dollars, except where otherwise indicated) Intermediate exchange rate Financial development Sovereign wealth fund Item No Yes Ratio No Yes Ratio No Yes Ratio Assets FDI 4,727.6 504.0 0.11 782.1 4868.0 6.22 422.2 1,1402.8 27.01 Portfolio 30,001.2 4,041.4 0.13 15,605.4 26539.1 1.70 2,470.6 4,4687.5 18.09 Debt 54,809.7 9,340.8 0.17 14,747.5 45724.2 3.10 8,592.9 10,8749.5 12.66 Total 107,720.7 18,590.7 0.17 32,432.6 93895.6 2.90 17,062.7 22,2520.4 13.04 Liabilities FDI 12,368.5 5,684.1 0.46 4,422.0 13510.6 3.06 5,799.2 20,892.7 3.60 Portfolio 11,56.6 211.9 0.18 63.6 1234.5 19.40 318.0 1,919.1 6.03 Debt 23,361.2 13,721.3 0.59 11,213.4 25098.9 2.24 14,103.1 33,645.8 2.39 Total 36,886.3 19,600.9 0.53 15,710.2 40161.4 2.56 2,0326.0 56,457.6 2.78 Aggregates Capital account 173.2 94.4 0.55 212.1 87.8 0.41 163.9 91.2 0.56 NFA 7,0834.5 –1,010.2 –0.01 16,722.4 53734.3 3.21 –3,263.2 166,062.8 –50.89 Reserves 18,887.9 8,268.0 0.44 10,500.4 18516.4 1.76 6,660.5 41,055.4 6.16 Source: Authors, based on Lane and Feretti 2015. Note: Countries whose credit to the private sector is above (below) the median are assumed to have a higher (lower) level of financial development. 4. Econometric Specification To examine the determinants of the accumulation of NFA and its components, we introduce several policy variables. First, we include a variable measuring the existence of an SWF. For long-term investors, SWFs can play a stabilizing role in financial markets by supplying liquidity and reducing market volatility. Second, we include Ln(Credit), a variable measuring financial development (the share of credit to the private sector), as countries with a higher level of financial development are more likely to attract both FDI and portfolio investment. Third, we control for the exchange rate regime (Interm) as well as whether the country is a net oil exporter (Oil). The econometric specification is as follows: Yit = α0 + α1 SWFit + α2Intermit + α3 Ln(Credit)it + α4Oilit + εit Where Y = the share of FDI to GDP (assets and liabilities), debt to GDP (assets and liabilities), portfolio investment to GDP (assets and liabilities), total assets and liabilities, reserves to GDP, and the capital account, and εit is the discrepancy term. Two remarks are worth mentioning. First, we run all regressions using a random effects estimator, as the Hausman test showed that it is more efficient. Second, in some regressions, we distinguish between oil exporters and oil importers, as in general oil exporters have more assets and oil importers more liabilities. 12   The data come from several sources. For financial variables, we use Lane and Ferretti (2007, updated in 2015). For the share of credit to the private sector, we use data from the World Development Indicators. For the exchange rate regime (Eichengreen et al., 2013), we use data from the International Monetary Fund’s Annual Report on Exchange Arrangements and Exchange Restrictions (AREAER), which reports countries’ self-reported exchange rate regimes. The existence of an SWF was constructed using different websites of SWF (see Appendix A). 5. Empirical Findings We used the Lane and Ferretti (2007) data set, which was updated to include data covering the period 1970–2015. We ran a compendium of regressions, including various components of the NFA—including the capital account, FDI, portfolio assets and liabilities, external debt, international reserves, and total foreign assets—as dependent variables. We first present the results on assets (with a focus on oil-exporting countries). We then report the results on liabilities (with a focus on oil-importing countries). We then report aggregates of the balance of payments, especially the capital account and reserves, for both oil-exporting and oil-importing countries. Table 3 shows the basic specification for assets. The SWF variable is statistically significant and positively associated with FDI (Reddy, 2019); it does not seem to matter for other assets (portfolio and debt). This finding shows the extent to which an SWF can help improve the allocation of investment assets toward FDI rather than portfolio investments. Financial development (defined as credit to the private sector as a percent of GDP) is also statistically significant for all asset components, with the highest elasticity for portfolio investments. This result is highlighted by Vermeulen and de Haan (2014), who show that financially developed countries are likely to take more risks than developing countries, because development allows them to insure against risk. For this reason, financially developed countries tend to invest more in portfolio investments than in FDI or debt. 13   Table 3 Determinants of assets – all countries FDI Portfolio Debt Total Sovereign wealth fund 1.105*** -0.129 0.0435 –0.149 (0.395) (0.440) (0.132) (0.120) Ln(Credit) 0.705*** 1.424*** 0.109*** 0.108*** (0.124) (0.140) (0.0411) (0.0374) Intermediate exchange rate –0.156 –1.794*** –0.0166 –0.143** (0.225) (0.259) (0.0737) (0.0669) Oil 1.029 4.343*** 2.296*** 2.376*** (0.903) (0.560) (0.843) (0.661) Constant –2.716*** –4.962*** 3.373*** 3.952*** (0.883) (0.813) (0.640) (0.509) Year dummies Yes Yes Yes Yes Observations 626 634 626 626 Number of countries 21 21 21 21 Note: Standard errors in parentheses. *** p < 0.01, ** p < 0.05, * p < 0.1. An intermediate exchange rate does not have a significant association with FDI or debt; it has a detrimental effect on total assets and portfolio investment, because it increases uncertainty. This is interpreted by the fact that oil-exporting countries are transparently adopting fixed exchange rate regimes supported by the comfortable current account surpluses. The oil dummy is positive and significant for portfolio and debt assets; it is not statistically significant for FDI. Thus, oil exporters in the region tend to invest more in debt assets and less in FDI (Elbadawi, Soto, and Zaki 2018). Table 4 shows the results on the determinants of assets for oil-exporting countries. Three findings are worth noting. First, the coefficient for having an SWF is largest for FDI, followed by the coefficient for debt. This finding suggests that GCC countries are more risk averse and tend to invest in less risky assets, namely debt. Second, financial development matters most for FDI, followed by portfolio and debt assets. This result confirms the finding of Vermeulen and de Haan (2014) that the more financially developed a country is, the more likely it will be to invest in risky portfolio assets rather than debt. 14   Table 4 Determinants of assets of oil-exporting countries FDI Portfolio Debt Total Sovereign wealth fund 1.487*** 0.198 1.227*** 1.081*** (0.432) (0.532) (0.195) (0.201) Ln(Credit) 1.872*** 0.677*** 0.610*** 0.582*** (0.191) (0.236) (0.0862) (0.0891) Intermediate exchange rate 0.752* –5.742*** –1.080*** –0.918*** (0.421) (0.518) (0.190) (0.196) Constant –5.013*** 2.092 4.617*** 5.208*** (1.694) (2.087) (0.763) (0.789) Year dummies Yes Yes Yes Yes Observations 334 334 334 334 Number of countries 11 11 11 11 Note: Standard errors in parentheses. *** p < 0.01, ** p < 0.05, * p < 0.1. As per liabilities, the existence of an SWF is associated with less debt and more portfolio liabilities (table 5). Financial development is also associated with greater portfolio liabilities and less debt. The elasticity of SWF with respect to debt liabilities is higher than the elasticity of portfolio liabilities, suggesting that it can lead to a larger decrease in debt, thanks to a better allocation of resources. As oil-exporting countries are mainly creditors not debtors, the oil dummy is insignificant. Exchange rate policy also turns to be insignificant. Table 5 Determinants of liabilities – all countries FDI Portfolio Debt Total Sovereign wealth fund –0.0443 0.0622*** –0.933*** –0.911*** (0.0351) (0.0109) (0.158) (0.165) Ln(Credit) 0.00554 0.0103*** –0.0824* –0.0754 (0.0110) (0.00348) (0.0491) (0.0515) Intermediate exchange rate 0.00322 –0.00355 –0.00984 –0.00775 (0.0200) (0.00635) (0.0879) (0.0922) Oil –0.0577 –0.0296* 0.866 0.778 (0.0759) (0.0171) (1.158) (1.233) Constant 0.194** –0.0308 0.527 0.708 (0.0763) (0.0212) (0.869) (0.924) Year dummies Yes Yes Yes Yes Observations 626 626 626 626 Number of countries 21 21 21 21 Note: Standard errors in parentheses. *** p < 0.01, ** p < 0.05, * p < 0.1. In oil-importing countries, the SWF dummy drops out from the liabilities’ regression, as only exporting countries have SWFs (table 6). For FDI-receiving countries, financial development matters more for FDI attraction than portfolio or debt liabilities. 15   Table 6 Determinants of liabilities of oil-importing countries FDI Portfolio Debt Total Ln(Credit) 0.189*** 0.0279*** 0.108*** 0.325*** (0.0157) (0.00504) (0.0401) (0.0467) Intermediate exchange rate 0.114*** –0.00821 –0.0200 0.0855 (0.0233) (0.00750) (0.0596) (0.0694) Constant –0.346*** –0.0777*** –0.0930 –0.517* (0.0931) (0.0300) (0.238) (0.277) Year dummies Yes Yes Yes Yes Observations 292 292 292 292 Number of countries 10 10 10 10 Note: Standard errors in parentheses. *** p < 0.01, ** p < 0.05, * p < 0.1. Regarding the balance of payments components, the existence of an SWF positively affects the accumulation of foreign reserves. From a policy perspective, this result is crucial to oil- importing countries that are currently accumulating reserves but will have to create SWFs in order to better allocate their assets. As expected, financial development and the oil exporter dummy have a positive and negative effect on reserves respectively. As in the case of assets, an intermediate exchange rate reduces assets and hence reserves, because it is less transparent than a fixed or flexible exchange rate regime (table 7). The results for the capital account are also in line with theory: The higher the level of financial development, the more the country invests abroad and hence the larger the deficit in its capital account. The same result applies for the oil dummy. Table 7 Determinants of reserves and the capital account – All countries Reserves Capital account Sovereign wealth fund 0.373** 0.000185 (0.189) (0.000172) Ln(Credit) 0.160*** –0.000164** (0.0567) (6.84e–05) Intermediate exchange rate –0.260** –5.75e–05 (0.104) (0.000124) Oil 1.203*** –0.000310** (0.419) (0.000145) Constant 2.748*** 0.000408 (0.400) (0.00118) Year dummies Yes Yes Observations 638 498 Number of countries 21 20 Note: Standard errors in parentheses. *** p < 0.01, ** p < 0.05, * p < 0.1. When we distinguish between oil exporters and oil importers, table 8 shows that financial development matters more for both reserves and the capital account of non-oil-exporting countries, 16   which rely more on FDI than do oil-exporters (table 8) since FDI is sensitive to the level of financial development in receiving countries. Table 8 Determinants of reserves and the capital account in oil-exporting and non-exporting countries Oil-exporting countries Non-oil-exporting countries Reserves Capital account Reserves Capital account Sovereign wealth fund 1.299*** 6.35e–05 (0.229) (0.000123) Ln(Credit) 0.252** 0.000129** 0.704*** –0.000465*** (0.102) (5.82e–05) (0.141) (0.000134) Intermediate exchange rate 0.0754 0.000169 0.658*** –0.000207 (0.224) (0.000120) (0.216) (0.000217) Constant 3.707*** –0.000287 2.210*** 0.00132 (0.795) (0.000716) (0.808) (0.00156) Year dummies Yes Yes Yes Yes Observations 338 258 300 240 Number of countries 11 10 10 10 Note: Standard errors in parentheses. *** p < 0.01, ** p < 0.05, * p < 0.1. 6. Concluding Remarks This paper makes two main contributions. First, using the methodology of Lane and Ferretti (2004, 2007), it analyzes trends in and the composition of NFA in selected Arab countries over the past two decades, with an emphasis on the relative significance of direct versus portfolio investment. It also distinguishes between FDI, portfolio equity investment, official reserves, and external debt. Second, the paper examines the effect of various policy variables that affect the accumulation of NFA and its components. It examines the effects of the presence of an SWF, the country’s exchange rate regime, and the level of development of the country’s financial system. Analysis of the stylized facts revealed by the data confirms that non-oil-exporting countries maintained a net borrower position between 2007 and 2015. Egypt, Tunisia, Jordan, Lebanon, Morocco, and Mauritania had higher net debt positions in 2015 than in 2007. These countries had a higher propensity to borrow foreign currency than GCC countries. Saudi Arabia, Bahrain, and Oman extended more debt financing to international borrowers in 2015 than they did in 2007. Libya and Algeria closed 2015 as net lenders, albeit with lower net positions than in 2007, a reflection of growing domestic financing needs. Almost all oil-exporting countries closed 2015 with higher ratios of reserves to GDP than in 2007; exceptions were Algeria, Bahrain, and Iraq, which experienced slight declines. Non-oil-exporting countries ended up cumulating higher ratios of reserves to GDP than all oil-exporting countries except Libya and Saudi Arabia. This pile-up of reserves in non-oil-exporting countries was coupled with higher levels of total liabilities, including foreign borrowing, FDI liabilities, and excessive leveraging. It is costly, given the sources financing it. 17   The presence of an SWF is statistically significant and positively associated with FDI in Arab countries. Financial development—defined as credit to the private sector as a percent of GDP—is also statistically significant across different regressions. In theory, the more financially developed a country is, the more it should invest in risky assets, such as portfolio ones. Arab investors are more risk averse than other investors, however, and hence invest mainly in debt. These results are particularly evident in oil-exporting countries, which tend to invest more in debt assets than portfolio assets. For oil-importing countries, financial development is the most important determinant of incoming FDI. References Bouchet, Michel Henry, Charles A. Fishkin, and Amaury Goguel. “Volatility, Spillovers, and Crisis Contamination: The New Dynamics of Country Risk since the 1980s and 90s in the Globalized Market Economy.” In Managing Country Risk in an Age of Globalization, ed. Amaury Goguel, Charles A. Fishkin, and Michel Henri Bouchet, 167–206. London: Palgrave Macmillan. Cavallo, M., and F. Ghironi. 2002. “Net Foreign Assets and the Exchange Rate: Redux Revived.” Journal of Monetary Economics 49 (5): 1057–97. Debelle, G., and G. Galati. 2007. “Current Account Adjustment and Capital Flows.” Review of International Economics 15 (5) 989–1013. Eichengreen, Barry, and Raul Razo‐Garcia. 2013. “How Reliable Are de Facto Exchange Rate Regime Classifications?” International Journal of Finance and Economics 18 (3): 216–39. Elbadawi, I., R. Soto, and C. Zaki. 2018. “Sovereign Wealth Funds, Cross-Border Investment Bias and Institutions: The Case of Arab Countries.” Working Paper 1173, Economic Research Forum, Cairo. Ghironi, F., and V. Stebunovs. 2007. “The Domestic and International Effects of Financial Integration.” Working Paper, Boston College, Chestnut Hill, MA. Henderson, D.W., and K. Rogoff. 1982. “Negative Net Foreign Asset Positions and Stability in a World Portfolio Balance Model.” Journal of International Economics 13 (1–2): 85–104. IIF (Institute of International Finance). 2015. GCC: Managing Foreign Assets in the New Normal. Managing Country Risk in an Age of Globalization: A Practical Guide to Overcoming Reference. Washington, DC. Lane P., and G. Milesi-Ferretti. 2001. “Net Foreign Assets and External Market Structure, Evidence from the Middle East Countries.” Working Paper 58, Egyptian Center for Economic Studies, Cairo. ———. 2004. “The Transfer Problem Revisited: Net Foreign Assets and the Real Exchange Rate.” Review of Economics and Statistics 86: 841–57. ———. 2007. “The External Wealth of Nations: Measures of Foreign Assets and Liabilities for Industrial and Developing Countries.” Journal of International Economics 55: 263–94. ———. 2015. Mendoza, E.G., and K.A. Smith. 2006. “Quantitative Implications of a Debt-Deflation Theory of Sudden Stops and Asset Prices.” Journal of International Economics 70 (1): 82–114. Obstfeld, M., and K. Rogoff. 1995. “Exchange Rate Dynamics Redux.” Journal of Political Economy 103 (3): 624–60. 18   Reddy, K.S. 2019. “Pot the Ball? Sovereign Wealth Funds’ Outward FDI in Times of Global Financial Market Turbulence: A Yield Institutions-Based View.” Central Bank Review, August 30. https://www.sciencedirect.com/science/article/pii/S1303070119300782. Vermeulen, R., and J. de Haan. 2014. “Net Foreign Asset (Com) Position: Does Financial Development Matter?” Journal of International Money and Finance 43: 88–106. 19   Appendix A: Largest Direct Sovereign Wealth Fund Transactions in 2015 Table A1 Largest direct sovereign wealth fund transactions, 2015 (millions of US dollars) Rank Date Target name Target sector Sovereign wealth fund Country Amount 1 2009-11-23 Qatar Railways Development Company Infrastructure Qatar Investment Authority (QIA) Qatar 13,260.00 2 2011-07-01 Allied Irish Banks Financial National Pensions Reserve Fund Ireland 12,748.50 (NPRF) 3 2009-09-02 Porsche Automobil Holding SE Industrial QIA Qatar 9,983.40 4 2008-03-05 UBS AG Financial Government of Singapore Investment Singapore 9,760.00 Corporation (GIC) 5 2009-07-06 France Telecom SA Telecommunications Fonds Strategique d'Investissement France 8,099.08 (FSI) 6 2007-02-28 Telstra Corp Ltd. Telecommunications Future Fund Australia 7,576.77 7 2007-11-26 Citigroup Inc Financial Abu Dhabi Investment Authority United Arab Emirates 7,500.00 (ADIA) 8 2009-05-12 China Construction Bank Corporation Financial Temasek Singapore 7,314.55 9 2010-09-24 Petrobras Energy Fundo Soberano do Brasil (FSB) Brazil 7,076.64 10 2008-01-28 Citigroup Inc Financial GIC Singapore 6,880.00 11 2008/-10-16 Credit Suisse Group AG Financial QIA Qatar 6,000.00 11 2010-07-16 Agricultural Bank of China Financial QIA Qatar 6,000.00 12 2014-03-21 A.S. Watson and Co Discretionary Temasek Singapore 5,700.00 13 2007-12-28 Morgan Stanley Financial China Investment Corporation (CIC) China 5,579.14 14 2014-8-24 VTB Bank OAO Financial Russian National Wealth Fund Russian Federation 5,422.70 (NWF) 15 2011-02-16 Cia. Espanola de Petroleos SA Energy International Petroleum Investment United Arab Emirates 5,370.00 Company (IPIC) Source: Sovereign Wealth Fund Institute website. [https://www.swfinstitute.org/] 20