The global financial crisis and its consequences continue to preoccupy policymakers. Capital markets around the world have been volatile, and governments are facing the difficult task of financing the investment needs of their local economy. There is a vast body of literature on the positive role of finance in economic development (Goldsmith, 1969, McKinnon, 1973, Fry, 1988 and Levine, 1992). Among the different forms of finance, equity financing is an important source, and the role of foreign investors in funding the requirements of domestic economies has never been more vital. Errunza (2001) suggests that foreign equity portfolio investors have a significant positive impact on the development of local equity markets, which in turn should drive domestic economic development. Given the importance of foreign equity portfolio investment, it is imperative for policymakers to appreciate factors influencing the country allocation decision of foreign investors. This paper investigates whether the investment decisions of foreign investors are affected by the host country-specific equity market characteristics.
The benefits of international diversification of portfolio investment are well established (see Grubel, 1968, Levy and Sarnat, 1970, Solnik, 1974a, Errunza, 1977 and Errunza, 1994 among others). The International Capital Asset Pricing (ICAPM) model suggests that international investors should hold the world market as the benchmark portfolio because it provides the best mean-variance efficiency (Tesar and Werner, 1995, Solnik and McLeavey, 2004, Chan et al., 2005 and Fidora et al., 2007). Studies also document the gradual removal of capital controls by developed countries beginning in the early 1980s (French and Poterba, 1991), and by developing countries by the late 1980s and early 1990s (Errunza, 2001 and Bekaert et al., 2003). However, despite increased access to financial markets across the globe, an extensive number of investigations demonstrate the prevalence of home bias, i.e., the tendency to overweight home markets relative to the theoretical prescription of the ICAPM (see Cooper and Kaplanis, 1994, Tesar and Werner, 1995, Warnock, 2001, Chan et al., 2005 and Fidora et al., 2007).
The investigations on home bias document a number of potential barriers impeding foreign investors from holding the world market portfolio. These barriers may be direct legal restrictions due to different legal status accorded to foreign and domestic investors (Bekaert, 1995), or indirect barriers arising from differences in available information and investor protection (Bekaert and Harvey, 1995, Errunza, 2001, Bekaert et al., 2003 and Hunter, 2006). Similarly, market-specific risks, such as diversification opportunities, liquidity, transaction costs and level of host market efficiency, commonly known as stock market development factors, could also potentially impede foreign investment (Chan et al., 2005). However, to the best of our knowledge, no empirical study has specifically documented the role of country-specific equity market factors on country allocation decisions of foreign investors, except Chan et al. (2005), who use different predetermined variables, including stock market development factors, to explain the issue of home and foreign bias. The persistence of home bias indicates that, on aggregate, foreign investors allocate a relatively large fraction of their wealth to domestic assets. This suggests that if we are able to control for home bias, we should be able to explain the role of different country-specific equity market characteristics in explaining bilateral cross-country foreign equity country allocation.
Chan et al. (2005) note that a major factor limiting research on foreign equity portfolio investment is the lack of cross-border holdings data. We make use of the recently available IMF’s Co-ordinated Portfolio Investment Survey (CPIS) foreign equity portfolio holding data.2 Similarly, as Chan et al. (2005) state, most existing studies are from the perspective of U.S. investors, and they leave the question open of whether the explanations for a wide cross-section of other source countries are similar or not. Furthermore, a very small number of existing studies, which use multiple source and host countries in their sample, only investigate the investments from developed countries into other developed countries. Since the U.S. and other developed countries’ equity markets exhibit higher levels of development relative to emerging markets, it remains to be tested whether the inclusion of the latter markets as host countries along with developed markets yields similar results.
Our study makes three important contributions to the literature. First, we try to explain the role of country-specific equity market characteristics in explaining the cross-sectional and temporal variation of foreign equity country allocation. Apart from Gelos and Wei (2005), who use emerging markets only and focus on transparency measures, and Chan et al. (2005), who explain foreign bias, no study has undertaken comprehensive empirical investigation modelling cross-country allocations. Second, as noted earlier, despite the theoretical suggestions of the ICAPM, global investors do not hold the world market as their benchmark portfolio. The ICAPM makes a number of assumptions, such as that global financial markets are perfectly integrated and fully efficient, investors incur no transaction costs, purchasing power parity perfectly holds and there are no barriers to international investments. Most of the earlier studies on international investments offer abstract theoretical explanations of why foreign investors may not hold the world market portfolio (see Solnik, 1974b, Black, 1974, Sercu, 1980, Stulz, 1981a, Stulz, 1981b, Alder and Dumas, 1983, Errunza and Losq, 1985, Eun and Janakiramanan, 1986, Cooper and Kaplanis, 1986 and Cooper and Kaplanis, 1994). The majority of the equilibrium frameworks suggest that the violation of unrealistic ICAPM assumptions, which create costs/risks for global investors, should explain the under- or over-weighting of foreign countries relative to ICAPM’s weightings. However, empirical evidence modelling the violations of the underlying assumptions is scarce and limited by the unavailability of high quality data. We fill the gap by using different proxies of the underlying assumptions to model cross-country allocation.
Finally, we pool bilateral data from 36 countries, developed and developing, spanning a period of nine years (2001–2009), with more than 500 cross-section units yielding approximately 4600 observations. Such a comprehensive dataset with wide cross-sectional and temporal variation affords us the statistical confidence to test our hypotheses using panel data econometric estimations. Baltagi (1995) demonstrates that, compared to purely cross-section data, panel data set-up supplies more information, more variability, less collinearity, greater degrees of freedom and higher statistical efficiency, yielding reliable parameter estimates. Moreover, the application of fixed effect panel data model controls for individual heterogeneity; studies not controlling for unit-specific effects run the risk of producing biased estimates. Furthermore, and to the best of our knowledge, we are the first to demonstrate the impact of the 2007–2008 global financial crisis on portfolio holdings, country allocations and parameter stability.3
The findings show that country-specific equity market factors, particularly market size, liquidity (level of market efficiency) and transaction costs, are the key factors influencing the country allocation decisions of foreign equity portfolio investors. We demonstrate that country-specific equity market characteristics, predominantly stock market development factors, explain almost 54% of the total variation in foreign equity portfolio allocations. One of the key implications of our study is that by improving the preconditions necessary for a well-functioning capital market, policymakers should be able to attract higher levels of foreign equity portfolio investments.
The rest of the paper is organised as follows. Section 2 briefly explains the theoretical framework underlying our empirical analysis and provides a detailed discussion of the variables used. Section 3 reports and discusses empirical results and Section 4 concludes the paper.
