جریان سهام خارجی و "حجم جانبداری ": شواهدی از بازار سهام در حال ظهور

This paper examines foreign investors' equity-level transactions in an emerging stock market, the Istanbul Stock Exchange, for the period 1997–2008 to derive insights into the debate on information asymmetries between domestic and foreign investors and the home bias puzzle. The analysis suggests two important findings. First, foreign investors do not consider the market portfolio of domestic securities as predicted by standard theories of international portfolio diversification. Second, the firm's size and the expected return are central to explain foreign investors' equity trades. The results are consistent with models based upon the hypothesis of differential information between foreign and domestic investors.

The home bias in portfolio investments remains an actual, yet unresolved question in international economics and finance. The term basically refers to the fact that investors hold disproportionately larger shares of their wealth in domestic securities in contrast to the predictions of a large body of theoretical and empirical research (see, among others, Van Wincoop and Warnock, 2010, Sorensen et al., 2007 and Baele et al., 2007). Empirically, such a behavior is considered to be irrational since there is substantial evidence suggesting that the international diversification of portfolios is potentially more beneficial than domestic diversification ( Baele and Inghelbrecht, 2009, Chiou, 2009, De Santis and Sarno, 2008, Driessen and Laeven, 2007 and Gilmore and McManus, 2002). Although the magnitude of the home bias worldwide has slightly declined during the past decades ( Fidora et al., 2007), data on existing foreign holdings suggest that investors are still underdiversified despite the lifting of many barriers to cross-border portfolio investments. Academics have so far come up with several explanations to rationalize the home bias, 1 however, none seems to have provided a fully satisfactory account of the phenomenon. This paper contributes to this literature by analyzing foreign investors' portfolio trading patterns2 in an emerging stock market and, thus, providing insights into the asymmetric information hypothesis between local and foreign investors. Indeed, the argument of asymmetric information between local and foreign investors provides a simple and direct way to understand the reason as to why investors tilt their portfolio toward local stocks. As long as investors are less well-informed about the payoffs on foreign securities, overweighting domestic securities becomes a rational behavior. Specifically, I explore a unique dataset3 that comprises equity-level transactions (purchases and sales) issued by foreign investors on 84 firms traded in the Istanbul Stock Exchange (the “ISE” henceforth) for the period 1997–2008. Although the complete list includes on average more than 300 firms' shares over this period, the subset used in this study captures a significant portion of foreign investors' transactions in the ISE such that the year-end market capitalizations of this subset of stocks represent on average 86% of the total market capitalization over the sample period. The analysis basically consists on examining the determinants of non-residents' stock-level trades using an array of explanatory variables drawn from prior studies. As such, this paper complements a number of related studies like Aggarwal et al., 2005, Dahlquist and Robertsson, 2001 and Kang and Stulz, 1997, since these papers also examine the portfolio preferences of foreign investors in different contexts. Meanwhile, the analysis also extends other papers that focus on the country-level determinants of cross-border portfolio investments (see, among others, Lane and Milesi-Ferretti, 2008 and Portes and Rey, 2005) by analyzing firm-level determinants of foreign investors' firm-level portfolio trades within an emerging stock market. Preliminary inspections reject the well-known hypothesis that foreign investors consider the standard international CAPM framework when investing into the Turkish stock market. Neglecting the market portfolio of securities, they focus on a smaller subset of large capitalization stocks. The sectoral breakdown of the transactions issued by non-residents suggests that foreigners prefer particularly financial stocks. The share of their transactions within this group generally exceeds the share of financials in the value-weighted market capitalization of the ISE. On average, 61% of the transactions issued by foreign investors are concentrated within the financial stocks for the period 1997–2008 while the average share of financial stocks in the ISE composition is about 52% over the same period. To examine the determinants of foreign investors' firm-level trading patterns, I employ regression analyses and relate foreign investors' transactions (e.g. purchases, net purchases and traded value) in local stocks to an array of financial metrics that capture different characteristics of the firm of interest. The predictor variables include different proxies for the firm's size (market capitalization), market characteristics (expected return and beta), leverage (total debt/total capital ratio), valuation (price/book ratio), profitability (return on assets), dividend yield (dividend payout ratio), and liquidity (turnover). Estimations point out to the firm's market capitalization and the expected return on the corresponding stock as the main significant variables in explaining the preferences of foreign investors when they trade in local stocks. For other variables the results remain mostly inconclusive, both statistically and economically. The significance of the market capitalization and the expected return is robust to alternative subsamples, disaggregation across sectors and other additional controls. The results are consistent with two relevant theoretical frameworks that point out to information asymmetries between local and foreign investors as the root cause of the home bias. First, the size bias vis-à-vis large capitalization stocks fits well with Merton's (1987) argument that investors are more likely to hold shares in firms which they know more about and they are more familiar with, and that the degree of familiarity increases with the firm's size. Thus, the tendency by foreign investors to trade in a smaller subset of large firms within the ISE confirms Merton's hypothesis that foreign investors in Turkey behave as if they know on average less about local stocks than do local investors.4 More recently, Barron and Ni (2008) provide an indirect theoretical framework linking information asymmetries to size. By means of a rational expectations model with heterogeneous portfolio managers in terms of the size of the portfolio, the authors show that managers with larger portfolios under their management are more likely to collect information on foreign assets, which implies a negative relationship between size and the home bias that these portfolio managers exhibit. Second, the statistical significance of the expected return and the non-residents' trend following behavior in the ISE are consistent with Brennan and Cao's (1997) model of international portfolio flows. Based on the assumption that an average local investor is better informed about the payoff structure on local assets than is an average foreign investor, the authors test the asymmetric information hypothesis between foreign and local investors using portfolio flow data on US investors' purchases of foreign equities. They find that US investors appear to be less well informed about the foreign markets, and that US portfolio flows are positively associated with returns on national market indices. As noted by Brennan and Cao (1997), the major empirical implications […] are that purchases of foreign equities will be a linear function of returns on the domestic and foreign equity markets; and that the coefficient of the return on the foreign market index will be positive, provided that foreign investors are less well informed about the payoffs on stock than are local investors (p. 1854). Thus, in addition to the size bias, this paper also sheds lights into the asymmetric information hypothesis between locals and non-locals by capturing the positive relationship between foreign investors' purchases and the returns on Turkish stocks. The paper is organized as follows. Section 2 is devoted to a brief presentation of the development of modern capital markets in Turkey. Section 3 presents the data on foreign investors' stock-level trades in the ISE and the measures used to control for the determinants of foreign investors' trades. Estimation results obtained using two different techniques are provided in Section 4. Section 5 demonstrates that these results are robust to a number of additional analyses. Section 6 concludes.

This paper explores a unique dataset on foreign investors' firm-level transactions in an emerging stock market, the Istanbul Stock Exchange, using a sample of 84 stocks for the period 1997–2008. The analysis provides useful insights into the determinants of cross-border portfolio flows and the issue of information asymmetry between local and non-local investors. The results can be summarized as follows. First, foreign investors' transactions are concentrated across large capitalization stocks that consist mostly of financial firms, thereby discarding the well-known result that investors should invest in the market-portfolio of securities as proposed by standard theories of international portfolio diversification. Second, regressions suggest that the firm's market capitalization and the expected return on the corresponding stock are two important factors in explaining net equity flows by non-residents. The significance of size and return variables is robust to different specifications and subsets of the data (e.g. financial vs. industrial stocks), as well as to additional controls taking into account other firm characteristics such as leverage, valuation, profitability, dividends and liquidity. Third, foreign investors are likely to be trend-followers such that a high significant contemporaneous correlation exists between foreign investors' net purchases of local stocks and the returns on the host market. Putting together, evidence presented here provides empirical support to the ongoing debate as to whether foreign investors are at an information disadvantage compared to domestic investors. First, the significance of the market capitalization as a key covariate of the net equity inflows into the ISE confirms the Merton's hypothesis (Merton, 1987), where size substitutes for the extent to which investors feel themselves familiar with the firm of interest. Second, the contemporaneous correlation between the returns on the host market and foreign investors' trading patterns in local stocks is typically consistent with the empirical predictions of models assuming differential information between domestic and foreign investors (Brennan et al., 2005 and Brennan and Cao, 1997). As such, the finding of information asymmetries between locals and non-locals also shed light into the debate on the home bias puzzle since overweighting one's home market can be justified on the basis of information asymmetries from which investors suffer when they consider investing abroad. One possible direction for future research is to control the results by using data on the holdings of foreign investors, since ownership data is likely to provide further, and, to some extent, more precise information on the investment behavior of foreign investors in local stocks. Second, while the results presented here can be considered as a convincing case that foreign investors are informationally less advantageous in Turkish stocks than are domestic investors, whether the latter group materializes or not this informational advantage remains an unanswered issue, mainly due to lack of data. Analyzing the respective performances of foreign versus local investors within an emerging stock market that had not been carefully examined so far, can also contribute to the ongoing debate on the determinants of international portfolio flows and the lack of global portfolio diversification.