Information about Europe Europa
Chapter

IV Determinants of Financial Globalization: A Cross-Country Perspective

Author(s):
Giovanni Dell'Ariccia, Paolo Mauro, Andre Faria, Jonathan Ostry, Julian Di Giovanni, Martin Schindler, Ayhan Kose, and Marco Terrones
Published Date:
December 2008
Share
  • ShareShare
Information about Europe Europa
Show Summary Details

What determines cross-country differences in de facto financial globalization (in contrast to the evolution over time in integration discussed in the previous section)? Despite the major increase in de facto financial globalization documented in Section III, countries' relative success in attracting international investors has been broadly stable over time: comparing countries' rankings by de facto financial globalization in different years, the rank correlation is 0.4 between the rankings in 1975 and 2004, and 0.7 between the rankings in 1995 and 2004. Such stability suggests that persistent country characteristics are likely to be key drivers of a country's de facto international financial integration. This section analyzes the role of such persistent factors, as well as that of capital controls.

Cross-country differences in de facto financial globalization may be related to both foreign investors' and domestic policymakers' views on whether foreign financing will be put to productive use. For example, foreign investors are likely to prefer to hold external liabilities of countries where such financing is expected to yield higher returns, while policymakers are likely to embrace financial globalization if they believe it will lead to higher growth without engendering excessive volatility. In fact, cross-country evidence drawn from two waves of financial globalization (1870–1913 and 1970s–present) suggests that key determinants of the productivity of foreign capital—including the quality of broad institutions and, to some extent, measures of human capital—are also the main determinants of international investors' willingness to hold a country's external liabilities (Faria and others, 2006). Similar factors also seem to affect the composition of a country's external liabilities: in a cross section of emerging market and developing countries, equity-like liabilities (FDI and portfolio equity) as a share of countries' total external liabilities are positively and significantly associated with indicators of educational attainment, natural resource abundance, and especially, institutional quality (Faria and Mauro, 2004).

Controlling for the persistent factors identified above, empirical analysis suggests that domestic policies vis-à-vis the financial account also have an impact on countries' external liabilities. Table 4.1 presents estimates of the impact of capital controls, institutional quality, trade openness, and level of economic development on total external liabilities as well as their components. The effect of each of these factors is both economically and statistically significant. In particular, a one-standard-deviation increase in the index of capital controls—equivalent to moving from the average for the Latin American countries to the average for developing and emerging East Asia–Pacific countries—is associated with a 17 percent reduction in total liabilities per capita, other things equal. While a one-standard-deviation change in the capital controls index is certainly sizable, it has been undertaken by several countries, within a few years, during the sample period considered.

Table 4.1.Determinants of Gross External Liabilities Per Capita, 2004
Total Liabilities (1)FDI and Portfolio Equity (2)Debt (3)
GDP per capita (log)0.86***0.97***0.80***
(0.07)(0.10)(0.08)
Institutional quality index0.48***0.35**0.50***
(0.13)(0.17)(0.15)
Trade openness0.46***0.81***0.29
(0.16)(0.18)(0.20)
Controls on inflows−0.46***−0.36**−0.55***
(0.15)(0.17)(0.20)
Constant−0.10−1.58***−0.41***
(0.15)(0.22)(0.16)
R20.940.920.90
Sources: Liabilities and their components are from Lane and Milesi-Ferretti (2006). Debt includes portfolio debt, bank loans, and currency deposits. Total liabilities consist of the sum of debt, FDI, portfolio equity, and financial derivatives. GDP per capita is from the World Bank's World Development Indicators (WDI). The institutional quality index is the simple average of six indicators from Kaufmann, Kraay, and Mastruzzi (2005): voice and accountability, political stability and absence of violence, government effectiveness, regulatory quality, rule of law, and control of corruption. Trade openness is the sum of imports and exports, divided by GDP, also from WDI. Controls on inflows are averages of all available years between 1995 and 2004 of indices of capital controls on total inflows (1), equity inflows (2), and debt inflows(3). Capital controls index constructed by staff based on the IMF's Annual Report on Exchange Arrangements and Exchange Restrictions.Notes: The sample consists of 96 observations. Offshore financial centers are excluded. Estimated by OLS, with robust standard errors in parentheses. The symbols *, **, and *** indicate statistical significance at the 10 percent, 5 percent, and 1 percent level, respectively.
Sources: Liabilities and their components are from Lane and Milesi-Ferretti (2006). Debt includes portfolio debt, bank loans, and currency deposits. Total liabilities consist of the sum of debt, FDI, portfolio equity, and financial derivatives. GDP per capita is from the World Bank's World Development Indicators (WDI). The institutional quality index is the simple average of six indicators from Kaufmann, Kraay, and Mastruzzi (2005): voice and accountability, political stability and absence of violence, government effectiveness, regulatory quality, rule of law, and control of corruption. Trade openness is the sum of imports and exports, divided by GDP, also from WDI. Controls on inflows are averages of all available years between 1995 and 2004 of indices of capital controls on total inflows (1), equity inflows (2), and debt inflows(3). Capital controls index constructed by staff based on the IMF's Annual Report on Exchange Arrangements and Exchange Restrictions.Notes: The sample consists of 96 observations. Offshore financial centers are excluded. Estimated by OLS, with robust standard errors in parentheses. The symbols *, **, and *** indicate statistical significance at the 10 percent, 5 percent, and 1 percent level, respectively.

Empirical analysis also sheds light on how each country allocates foreign assets and liabilities across other countries. Estimating a fixed-effect “gravity model” for bilateral financial holdings of equity, FDI, bank loans, and other debt—similar to such models used to explain trade flows—country pairs characterized by historical links (common language, colonial history, and common legal systems) are found to have larger bilateral holdings (Table 4.2—see also Lane and Milesi-Ferretti, 2008). Moreover, countries that are further apart geographically and that do not share a border have significantly lower bilateral financial integration. The economic magnitude of the coefficient on geographical distance is substantial: for example, the estimates predict that bilateral equity holdings should be about 75 percent larger between France and the United States than between Australia and the United States. Moreover, geographical distance and historical linkages have a significant impact on financial asset allocation even controlling for the strong correlation between trade and financial patterns.1 The estimates also confirm that capital controls on inflows in recipient countries and on outflows in source countries (for each type of flow) are negatively correlated with bilateral holdings.

Table 4.2.Gravity Estimates for Bilateral Foreign Asset Positions, 2004
EquityDebtBank LoansFDITradeEquityDebtBank LoansFDI
(1)(2)(3)(4)(5)(6)(7)(8)(9)
In(distance)−0.77***−0.80***−1.08***−1.29***−1.65***−4.12***−1.27−7.46***−5.58***
(0.06)(0.06)(0.09)(0.11)(0.03)(0.04)(0.04)(0.05)(0.07)
Border0.64***0.14−0.100.63*0.75***0.52**0.060.000.67*
(0.22)(0.20)(0.31)(0.38)(0.12)(0.22)(0.20)(0.28)(0.38)
Common language0.49***0.231.14***0.54**0.81***0.62***0.24*1.12***0.28
(0.15)(0.14)(0.15)(0.27)(0.06)(0.14)(0.14)(0.23)(0.27)
Common colony1.08***2.33***1.01***1.08***2.24***
(0.39)(0.47)(0.09)(0.32)(0.63)
Common legal origin0.20**0.120.090.32*0.040.18**0.150.090.43**
(0.09)(0.09)(0.10)(0.17)(0.04)(0.09)(0.09)(0.12)(0.17)
Capital control—source2.52***-0.060.430.54-0.661.080.43
closed, recipient open(0.96)(0.51)(0.72)(0.72)(0.46)(1.70)(0.83)
Capital control—source2.66***0.06−0.130.313.63***0.06-0.880.53
open, recipient closed(0.59)(0.47)(1.17)(0.57)(0.62)(0.47)(1.73)(0.62)
Capital control—both open5.44***1.10**-0.471.30*3.74***0.411.20
(0.91)(0.54)(1.13)(0.68)(0.78)(0.47)(0.80)
Recipient transparency ×0.26***0.32***0.140.22*
In(distance)(0.07)(0.07)(0.10)(0.12)
Recipient GDP ×In(distance)0.07**-0.050.22***0.12*
(0.04)(0.04)(0.05)(0.07)
Observations1,5601,3461,77579315,5261,3041,040915578
R20.830.850.800.860.750.850.860.830.89
Sources: Equity and debt holdings are from the IMF's Coordinated Portfolio Investment Survey, bank loans from the Bank for International Settlements, FDI from the Organization for Economic Cooperation and Development, and trade from the IMF's Direction of Trade Statistics. The financial transparency measure is taken from issues of the World Economic Forum's Global Competitiveness Report. Capital controls data are constructed by staff based on the IMF's Annual Report on Exchange Arrangements and Exchange Restrictions.Notes: All dependent variables in natural logarithms, and measured in end-of-year 2004 U.S. dollars. The capital control variables refer to controls on the specific type of flows, and refer to inflows for the recipients and outflows for the source. Source and recipient country dummies (fixed effects) are included. The estimated capital controls coefficients are relative to the “base” case where both source and recipient countries are closed. Robust standard errors in parentheses. The symbols *, **, and *** indicate statistical significance at the 10 percent, 5 percent, and 1 percent level, respectively. The effects of “common colony,” “common legal origin,” and “capital control—both open” cannot be estimated for bank loans and FDI (columns 3, 4, 8, and 9), owing to insufficient variation across country-pairs for those cases.
Sources: Equity and debt holdings are from the IMF's Coordinated Portfolio Investment Survey, bank loans from the Bank for International Settlements, FDI from the Organization for Economic Cooperation and Development, and trade from the IMF's Direction of Trade Statistics. The financial transparency measure is taken from issues of the World Economic Forum's Global Competitiveness Report. Capital controls data are constructed by staff based on the IMF's Annual Report on Exchange Arrangements and Exchange Restrictions.Notes: All dependent variables in natural logarithms, and measured in end-of-year 2004 U.S. dollars. The capital control variables refer to controls on the specific type of flows, and refer to inflows for the recipients and outflows for the source. Source and recipient country dummies (fixed effects) are included. The estimated capital controls coefficients are relative to the “base” case where both source and recipient countries are closed. Robust standard errors in parentheses. The symbols *, **, and *** indicate statistical significance at the 10 percent, 5 percent, and 1 percent level, respectively. The effects of “common colony,” “common legal origin,” and “capital control—both open” cannot be estimated for bank loans and FDI (columns 3, 4, 8, and 9), owing to insufficient variation across country-pairs for those cases.

The finding that bilateral gross asset holdings are closely associated with factors such as distance, language, and former colonial links may seem surprising in a globalized world where information appears to flow freely. If distance is instead a proxy for residual informational frictions, it may be expected to matter less for relatively transparent recipient countries and for large markets that are well covered by financial analysts. To investigate this hypothesis, Table 4.2 also considers the impact of the interaction between distance and a survey-based indicator of the recipient country's financial market transparency, as well as market size. Both greater financial transparency and country size in the recipient country are found to dampen the negative impact of distance on bilateral equity holdings.2 The implication is that improved transparency may help persuade international investors to hold a larger stock of a country's external liabilities. Moreover, this effect seems to be greater for countries that are relatively isolated from the majority of international investors

This section was prepared by André Faria and Paolo Mauro.
1A regression specification controlling for bilateral trade flows delivered similar results.
2This result is robust to including an interaction of distance and the recipient's financial market development, as measured by stock market capitalization relative to GDP, but not to adding the interaction between distance and the recipient's per capita GDP, which might proxy for other facets of economic development beyond financial market transparency.

    Other Resources Citing This Publication