Portfolio Preferences of Foreign Institutional Investors

Transcription

1 Portfolio Preferences of Foreign Institutional Investors Reena Aggarwal McDonough School of Business Georgetown University Washington D.C (202) Leora Klapper The World Bank 1818 H Street, NW Washington D.C (202) lklapper@worldbank.org Peter D. Wysocki Sloan School of Management Massachusetts Institute of Technology Cambridge, MA (617) wysockip@mit.edu (February 2003)

2 Portfolio Preferences of Foreign Institutional Investors Abstract Countries with higher levels of economic development tend to have greater ability to obtain foreign capital. However, our results suggest that the availability of foreign capital also depends on factors other than the country s economic development and the firm s financial attributes. After controlling for the country s level of economic development, we find that firms in countries with stronger shareholder rights and legal framework attract more foreign capital. We also find that foreign institutions allocate more of their assets to firms with better accounting quality. Our results suggest that steps can be taken both at the country and the firm level to create an environment conducive to foreign portfolio investment. The analysis is based on a unique dataset consisting of equity positions of U.S. mutual funds in emerging markets. 1

3 Portfolio Preferences of Foreign Institutional Investors This paper examines the investment allocations of U.S. mutual funds in emerging market equities. During the 1990 s, there was a dramatic increase in institutional ownership of and trading in U.S. equities. Recent research has also shown that rising institutional ownership has contributed to both increased trading liquidity and corporate governance reforms in U.S. companies. (for example, Gompers and Metrick, 1999). However, our understanding of investment in emerging markets by U.S. institutional investors is quite limited. The objective of this paper is to provide a better understanding of the preferences of U.S. institutional investors as revealed by their portfolio holdings in emerging market equities. We analyze portfolio holdings of mutual funds with respect to countries, equity markets, and firms that they invest in. The research attempts to examine three major issues. First, we study the relationship between country-level characteristics, for example, macroeconomic factors and legal system, and institutional investment. Second, the relationship between market level factors, such as market size and investor protection laws, and investment by funds is examined. Finally, we analyze the characteristics of firms that are part of the holdings of U.S. mutual funds. The firm characteristics studied include stock attributes, financial attributes and accounting disclosure attributes. The analysis for the first time helps us to understand the investment preferences of foreign institutional investors in emerging markets. Countries and firms are interested in attracting foreign capital because the additional demand helps to create liquidity for both the firm s stock and the stock market in general. This additional demand also leads to lower cost of capital for the firm and allows firms to compete more effectively in the global marketplace. This directly benefits the economy of the country. Recent research has demonstrated a strong linkage between capital market development and 2

4 economic development. 1 Countries with higher levels of economic development do tend to have more developed capital markets and are believed to have greater ability to obtain foreign capital. However, our results suggest that the availability of foreign capital also depends on factors other than economic development of the country. At the country level we find that while size of the economy is positively related to the level of foreign institutional investment, it does not appear to be sufficient to explain patterns in foreign capital allocation. Shareholder rights in the country and legal framework are also important attributes at the country-level. At the firm-level, we find that funds invest in larger firms and firms with a large analysts coverage. A more detailed analysis shows that funds allocate a larger proportion of their assets in firms that have better accounting quality. High quality financial reporting appears to be more important than other attributes such as liquidity. We also examine the extent to which funds invest in the same set of stocks as a major market index. While there is considerable overlap between the emerging markets holdings of U.S. funds and the MSCI Emerging Markets index, funds do not limit their investments to stocks tracked by this major market index. The index has 679 securities but our sample of firms invests in almost twice that number of securities. In addition, some of the securities included in the index are not part of the fund holdings. Equities held by U.S. funds that are not in the index tend to be smaller, have less analyst coverage, and are less profitable. Further analysis is done based on examining the mean percentage allocation of funds relative to the MSCI weight. Funds are found to have higher relative allocation in larger firms, firms with large analyst following, and firms with better accounting quality. Our findings on emerging markets extend the growing literature on the determinants of global investment flows and allocations. Prior research focuses on international portfolio flows 1 For example see Levine (1997), King and Levine (1993a, 1993b) and Rajan and Zingales (1998). 3

5 and examines the relationship between portfolio flows and stock returns. These studies have analyzed the global, regional and local factors that influence portfolio flows. 2 A few studies have also examined allocations but they have generally focused only on a specific country. We extend this analysis and undertake a comprehensive analysis of all emerging markets and provide more detailed analysis of country and firm-level factors that influence investment allocations by U.S. funds. Dahlquist and Robertsson (2001) undertake a detailed analysis of foreign ownership and firm characteristics for the Swedish market. They find that foreigners have a preference for large firms, firms paying low dividends, and firms with large cash holdings. The finding on firm size is driven by liquidity and international presence as measured by foreign listings and export sales. Foreigners tend to underweight firms with a dominant owner. Brennan and Cao (1997) develop a theoretical model that accounts for information asymmetry between domestic and foreign investors. Their empirical analysis shows that domestic investors have informational advantages. Covrig, Lau, and Ng (2002) also conclude that foreign fund managers have less information about domestic stocks than do domestic fund managers. They find that ownership by foreign funds is related to size of foreign sales, index memberships, and stocks with foreign listing. These findings are attributed to information asymmetries between foreign and domestic investors. Similarly, Kang and Stulz (1994) report that foreign investment in Japan is concentrated in large firms and in firms that have a larger proportion of export sales. The findings suggest that foreigners invest in firms that they are better informed about. For the U.S. markets, Falkenstein (1996) investigates the preferences of U.S. open-end mutual funds and 2 For example, see Kaminski, Lyons, and Schmukler (2001), Froot, O Connell and Seasholes (2001), Borensztein and Gelos (2001) and Brennan and Cao (1997). 4

6 reaches similar conclusions. Even in the U.S. there is a preference towards stocks with high visibility and liquidity. The rest of the paper is organized as follows. Section I discusses the data used in the empirical analysis. The empirical results on the relationship between fund holdings and firm characteristics are presented in Section II. Section III reports results on fund holdings relative to the market index. A summary and conclusions are provided in Section IV. I. Data and Methodology A. Mutual Funds in Sample We obtain portfolio holdings data for 576 U.S. mutual funds with stated objectives of investing in foreign markets that include emerging market countries. We include all actively managed equity mutual funds with a stated objective of diversifying their investments into emerging markets. Consistent with the MSCI Emerging Markets Index, we define the following countries as emerging markets: Argentina, Brazil, Chile, China, Colombia, Czech Republic, Egypt, Greece, Hungary, India, Israel, Jordan, Malaysia, Mexico, Morocco, Pakistan, Peru, Philippines, Poland, Russia, South Africa, South Korea, Sri Lanka, Taiwan, Thailand, Turkey and Venezuela. The mutual fund portfolio holdings are obtained from the February 2002 version of the Morningstar database. 3 Our sample of funds is classified as: Diversified Emerging Markets, Diversified Pacific/Asia, Pacific/Asia excluding Japan, Latin America, and Europe. Diversified Pacific/Asia mutual funds are excluded because they invest a significant portion of their assets in countries that are not considered emerging markets, such as, Japan, Hong Kong, and Singapore. Similarly European funds are also excluded because they invest a large portion 3 Because different funds report their portfolio holdings at different times, the February 2002 version of the Morningstar database contains funds portfolio holdings for either October 2001, December 2001, or January

7 of their assets in developed European markets. From Pacific/Asia excluding Japan funds we exclude 11 funds that invest primarily in less than three countries. Many funds have multiple classes that have exactly the same portfolio holdings but their fee structures are different. For example, one fund may have an exit fee but the other does not. We do not include multiple classes of the same fund. The sample also excludes exchange-traded funds. After excluding multiple classes of the same fund and exchange-traded funds the sample consists of 114 funds. Out of the total sample of 114 funds, 74 are classified as Diversified Emerging, 25 as Pacific/Asia excluding Japan and, 15 as Latin America. All of these funds are primarily equity funds with more than 90 percent of their investment in equities. The countries that attract the most portfolio investment from our sample of U.S. mutual funds are Brazil, Mexico, South Africa, and Malaysia. Diversified Emerging funds are the most diversified (and do not focus on any one region) and have the largest allocations in Mexico, Taiwan, Brazil, South Africa, and India. At the fund level, we collect detailed information on each fund s portfolio holdings, the proportion of the fund s assets invested in a stock, percentage invested in stocks, bonds, and cash, regional investments, and investments by industry sectors. Performance measures for the funds include various holding period returns, alpha, beta, price to book, price to earnings, and Morningstar rankings. B. Country Characteristics The country characteristics examined include the following macroeconomic variables: GDP per capita (GDPPC), GDP per capita growth (GDPPCG), and size of the stock market measured as market capitalization/gdp (MKTCAPGDP). At the country level we also introduce two measures of corporate governance: Shareholder rights (SHRIGHTS) and legality 6

8 (LEGALITY) as an indicator of investor protection (Shleifer, et al., 1997). The SHRIGHTS measure is the sum of dummies identifying one-share/one vote, proxy by mail, unblocked shares, cumulative vote/proportional representation, preemptive rights, oppressed minority, and % of shares needed to call a shareholders meeting. Klapper and Love (2001) find that firm-level corporate governance matters more in countries with weaker investor protection. In the next section, we examine fund allocation relative to a major market index. Country-level data for the year 2000 on market capitalization, GDP per capita, and growth in GDP per capita is obtained from the IFS database. Firm-specific information from Worldscope and the Emerging Markets Database and on accounting attributes is merged with the Morningstar fund portfolio holdings data. This data matching procedure posed a challenge because the Morningstar database uses non-standard company identifiers. C. Firm-Specific Attributes Firm-specific data for the year 2000 is obtained from Worldscope. These include balance sheet, income statement, and stock market data, the number of closely held shares, and measures of corporate governance. The source for share volume turnover and dollar volume turnover for each firm is the Emerging Markets Database. After examining the firm s financial characteristics we also analyze the impact of accounting and disclosure quality on fund allocations. The motivation is to examine the issue of whether accounting plays any role in the investment decisions of foreign mutual funds. The three accounting-related independent variables are the quality of the auditor, consolidated reporting and use of internationally-recognized accounting standards. First, auditor quality, AUDITOR, is a dummy equal to1 if the firm uses a international Big-5 accounting firm. Second, CONSOL is equal to 1 if the firm presents consolidated financial statements then 7

9 outsiders should have a complete performance picture of all the firm s investments and subsidiaries. Finally, ACCSTD is equal to 1 for firms that use internationally recognized accounting standards (GAAP or International Accounting Standards). These firm-level accounting data are obtained from the Worldscope database that is collected from companies annual financial statements. Firms that use internationally recognized accounting standards are viewed as having financial statements that present a true and fair picture of firm performance. This dummy variable (ACCTSTD) has a positive and statistically significant coefficient in each of the models considered. Therefore it can be concluded that funds allocate more of their capital to emerging market companies that have better accounting standards (after controlling for other firm characteristics). The listing requirements of ADRs often require firms to adopt or reconcile to U.S. GAAP for their financial statements. Therefore, it is expected that firms that have ADRs will follow higher quality accounting standards. The ADR coefficient is not significant in our models with or without the accounting variables. A firm is categorized as having better quality financial statements if it is audited by an international Big-5 auditor. Prior research has show that audit quality tends to be higher for Big-5 auditors. In one of the models this dummy variable (AUDITOR) is significant and negative, however, as we discuss later in the full model that controls for country-specific factors auditor quality is positive and significant. D. MSCI Index Benchmark We use the MSCI Emerging Markets Free Index as the market benchmark. This is a free float-adjusted market capitalization weighted index that is designed to measure equity market performance in global emerging markets. The weight of each security in the MSCI index is obtained from Morgan Stanley. These weights at the broadest level also allow us to build an 8

10 index at the regional and country levels. A total of 704 securities were included in the MSCI Emerging Markets Index at the end of We find that the portfolio holdings of U.S. registered mutual funds investing in emerging markets include many firms that are not part of the MSCI index. As discussed in detail later, we also find that there are firms included in the MSCI index that are not held by any of the mutual funds in our sample. We carefully examine the characteristics of firms that are included in the MSCI Index and are also included in the mutual funds portfolio holdings; firms that are included in the MSCI Index but are not included in mutual fund portfolio holdings; firms that are not included in the MSCI index but are included in the mutual fund portfolio holdings. II. Fund Characteristics and Allocations There are three different types of funds analyzed that invest in emerging markets. 4 The first category is Diversified Emerging market funds that invest primarily in emerging markets around the world and generally do not concentrate their investments in any one region. The second is Latin American funds have if at least 75% of assets are invested in the region. Finally, the third is Pacific/Asia excluding Japan funds that invest at least 75% of their assets in Pacific countries, with less than 10% of stocks invested in Japan. We refer to the last category as Asian funds. A. Fund Characteristics The Diversified Emerging Market funds are on average much larger in size than any of the other fund categories as seen in Table 1 with mean net assets of $ million under 4 All markets except the following developed markets: Australia, Austria, Belgium, Canada, Denmark, Finland, France, Germany, Ireland, Italy, Japan, the Netherlands, New Zealand, Norway, Singapore, Spain, Sweden, Switzerland, the United Kingdom, and the United States are considered emerging markets for our sample period according to Morningstar but we do not classify Hong Kong as an emerging market consistent with the MSCI definition. 9

11 management. Latin America funds are the smallest with mean assets of $33.81 million under management. On average, funds hold positions in 80 firms. The Diversified Emerging category is the most diversified and also the largest in terms of assets under management and these funds have 108 holdings on average. This is followed by positions in 38 firms for Latin America funds and 31 for Pacific/Asia excluding Japan funds. The actual number of positions taken by funds is larger than those reported here because a fund may hold several different classes of the same firm s equity. Mean asset turnover is close to 100 percent for all fund categories except Diversified Pacific that has the highest turnover at 104 percent. The sensitivity of the fund portfolios relative to the U.S. market is captured by beta. The three fund categories have a much higher risk ( ) as captured by beta. The Latin America category has the highest beta at These are all equity funds therefore percent of their assets are on average invested in equities and only small amounts are invested in bonds or held as cash. The portfolio holdings of the funds are concentrated as captured by the proportion of a fund s portfolio invested in top-ten holdings. The concentration is lowest for Diversified funds with 28 percent invested in top-ten holdings and is highest for Latin American funds that have 54 percent of their assets invested in top-ten holdings. B. Allocation and Ownership We next analyze the size of investments by U.S. mutual funds in foreign firms. As reported in Table 2, the sample consists of 10,688 positions in 1,456 firms. As expected, several funds from the same fund category might invest in the same firm; therefore the total number of positions is larger than the number of firms. Also, the same firm may show up in several fund categories. For example, the Indian firm VSNL is included in the portfolio holdings of mutual 10

12 funds categorized as Diversified Emerging and also those categorized as Diversified Pacific/Asia. Similarly, there are funds that do not invest in a particular firm hence the minimum ownership is zero in each case. The largest number of equity positions is held by the Diversified Emerging market funds at 7,866. This is to be expected because these funds are most diversified and as noted earlier they are also much larger in size as measured by assets under management. Additionally, these funds do not concentrate their holdings in any particular region. Diversified Pacific/Asia funds hold 611 equity positions, European funds 115, Latin American 543, and Pacific/Asia excluding Japan 1,553. European funds, relative to their size, hold a smaller number of equity positions than some other fund categories such as Latin America. This is possible because there are a number of large stocks in Europe that are extremely liquid but this liquidity might not be available for Latin American stocks and therefore the fund manager needs to diversify into a larger number of stocks. The percentage of a fund s assets allocated to a single position is reported in Panel A of Table 2. The Diversified Emerging funds are most diversified with mean and median allocation in each position of 0.68% and 0.43%, respectively. European funds allocate more than 2% of their assets to any single position. The largest single position held by the Diversified Emerging funds is 7.21%, Diversified Pacific/Asia funds is 6.05%, Europe is 9.68%, Latin America is 14.73%, and Pacific/Asia excluding Japan is 21.41%. These findings indicate that funds do hold positions that are large relative to their fund s total assets. These positions are largest for the Latin America and Pacific/Asia excluding Japan funds. Diversified Emerging funds invest in a total of 1,339 firms, Diversified Pacific/Asia in 224, European in 65, Latin American in 117 and Pacific/Asia excluding Japan in

13 Panel B of Table 2 shows the percentage of a firm owned by the funds based on each position. The mean and median ownership in each of the 10,688 positions is small. Maximum ownership based on any single position varies from region to region being highest for Latin America (5.13%) followed by Diversified Emerging (2.65%), Pacific/Asia excluding Japan (1.70%), Europe (1.59%) and Diversified Pacific/Asia (0.30%). Portfolios that do not include any developed markets have larger maximum ownership. These results reflect the fact that firms in emerging markets tend to be smaller therefore maximum ownership by mutual funds is expected to be larger. U.S. mutual funds are only one group of foreign institutional investors. Non-U.S. mutual funds, pension funds and hedge funds are other examples of foreign institutional investors that hold positions therefore making the foreign institutional holdings much larger than that reported here. However, our focus is to study the investment behavior of U.S. mutual funds and not all foreign institutions. Borensztein and Gelos (1999) discuss the possibility that even though mutual fund investment in firms might be modest but their trading activity may play a much larger role. This is because turnover is much lower in emerging markets than in mature markets partially because of concentrated ownership. We also estimate ownership based on float instead of total shares outstanding. This method results in slightly higher percentages for ownership but does not change the conclusion and the results are not reported here. III. Fund Holdings and Country Analysis A. Portfolio Holdings at Country-Level The first part of our analysis examines fund investments at the country-level and also analyzes fund investments in a country relative to the weight of the country in the MSCI Index. 12

14 The MSCI Index is constructed at three levels: 1)Asia, 2)Latin America, and 3)Diversified Emerging in order to make the appropriate comparisons with the three fund categories examined in this study. Asian funds invest in nine countries: China, Indonesia, India, Korea, Malaysia, Philippines, Sri Lanka, Taiwan and Thailand. Of the 25 funds, 24 have investments in Korea and Taiwan as shown in Table 3. Only one fund invests in Sri Lanka. The MSCI Index includes firms from all the nine countries except Sri Lanka. The MSCI Index also invests in firms in Pakistan but the funds have no investment in Pakistan. The MSCI Index includes 88 firms from Taiwan, 66 firms from Malaysia, and 58 firms from India. All 15 Latin American funds invest in Brazil and Mexico. Only one fund invests in Venezuela. The MSCI Index for Latin America invests in 34 firms in Brazil, 22 in Chile and Mexico, and only 5 firms in Venezuela. All 74 Diversified Emerging funds invest in Korea and Mexico. Other countries frequently included in fund portfolios are India (71 funds), Taiwan (71 funds), Thailand (69 funds), Israel (68 funds) and South Africa (68 funds). The countries with the smallest number of funds are Jordan (1 fund), Slovenia (2 funds), Zimbabwe (3 funds), Ghana (3 funds), Venezuela (3 funds), and Morocco (4 funds). 5 Table 4 reports the proportion of a fund s assets invested in each country. Also reported are the weights of the country in the MSCI Index. The countries with the largest weights in the portfolio of Asian funds are Korea (40.78%) and Taiwan (27.77%). These two countries also have the largest weights in the MSCI Index with Korea at 31.73% and Taiwan at 28.04%. Next, we analyze the proportion of a fund s assets invested in a country relative to the country s weight in the MSCI Index. The % Relative Spread is defined as: % Relative Spread = ((Fund s Portfolio % - MSCI Weight %) / MSCI Weight %)*100; 5 The following countries are not included in Table 3: Jordan, Slovenia, Zimbabwe, Ghana, and Morocco. 13

15 Therefore, relative to the Index the funds over-invest in Korea by 28.54% and underinvest in Taiwan by 0.95% as measured by % Relative Spread in Table 4. Fund portfolios underweight firms in China, Indonesia, India, Malaysia, Philippines, and Taiwan relative to the MSCI weights. Latin American funds, on average, invest 47.78% and 40.98% of their assets in Mexico and Brazil, respectively. Mexican firms constitute 40.14% of the MSCI Index and Brazilian firms constitute 40.14%. Therefore, Mexico is over-weighted by 19.03% in fund portfolios and Brazil is over-weighted by 1.34%. Fund portfolios under-weight firms in Chile, Colombia, Peru and Venezuela relative to the Index as measured by % Relative Spread in Table 4. Diversified Emerging funds on average invest a smaller proportion of their assets in firms from a single country as compared to the regional Asian and Latin American funds discussed above. They invest in a total of 30 countries. On average, Korean firms receive the largest proportion of the funds assets (18.86%), followed by Mexico (13.50%), Taiwan (11.17%), and Brazil (10.17%). These same countries also have the largest weights in the MSCI Index. However, the fund s over/under-investment relative to the Index varies from 254% to -94%. B. Portfolio Holdings and Country-Level Attributes We next relate fund s portfolio holdings to country attributes. The objective is to examine the country-level factors that may influence a fund s portfolio holdings in a country. A fund s portfolio holdings are measured in two ways: 1) Percentage of a Fund s Portfolio invested in a country, and 2) Percentage Relative Spread that measures the over/under-investment by a fund relative to the MSCI Index weight for the country. These two variables are used as dependant variables. The independent variables are proxies for 1) the size of the economy: log of GDP per capita (LGDPPC); 2) the size of the market: market capitalization/gdp 14

16 (Mktcap/GDP); and 3) corporate governance: legal framework (LEGALITY) and shareholder rights (SHRIGHTS). A number of other variables in each of the three categories were also explored. For example, growth in GDP per capita as a proxy for size of the economy; market turnover as a proxy for market size; quality of accounting, mail proxy, as proxies for corporate governance. However, the correlation among some of these variables is high and therefore only a limited number of regression models are reported. We report correlations for attributes at the country-level in Table 5. Log of GDP per capita is included in all the regression models to control for size of the country s economy. This variable is not significantly correlated with any of the other variables used in the regression models except LEGALITY. However, it is positively related with market capitalization to GDP with a correlation of Mktcap/GDP is significantly and positively correlated with accounting standards in the country (0.60) and also with LEGALITY (0.53). Because of the high correlations, separate regression models are estimated. Table 6 presents regression results using % Fund Portfolio Holdings as the dependant variable. The models are estimated for each fund category: Asia, Latin America, and Diversified Emerging. The five independent variables in the models are LGDPPC, Mktcap/GDP, LEGALITY, SHRIGHTS, and the interaction between LEGALITY and SHRIGHTS. The number of observations for each fund category equals the number of funds times the number of countries. For example, there are 15 Latin American funds that can invest in seven Latin American countries. Therefore, the maximum number of observations for Latin America is 105 (15*7) but due to missing values for some variables only 104 observations are reported. Based on the regression models in Table 6 it can be concluded that funds invest more in countries with larger economies as measured by log of GDP per capita (LGDPPC). After 15

17 controlling for the size of the economy, the coefficient of market capitalization/gdp (MKTCAPGDP) is significant and negative for the two regional categories, Asia and Latin America and is significant and positive for Diversified Emerging funds. Two proxies are used for corporate governance, LEGALITY is positive and significant for two of the categories. SHRIGHTS is positive and significant for Asia, negative and significant for Latin America, and not significant for Diversified Emerging. Next, we examine the relationship between % Relative Spread and country attributes. 6 The objective is to examine factors that determine over and under investment by funds relative to the MSCI Index and country attributes. Three conclusions can be reached based on the regression results presented in Table 7. Funds invest more of their assets relative to the MSCI Index weights in countries with less developed equity markets as measured by market capitalization to GDP. Funds also invest relatively more in countries with weaker corporate governance. The coefficients for both LEGALITY and SHRIGHTS are negative and statistically significant at the 5% level. However, including LEGALITY, SHRIGHTS, and the interaction between these two variables, in the same model results in each coefficient being positive and significant and the interaction coefficient being negative for Asia and Diversified fund categories. This suggests that funds are particularly concerned about countries that have weak shareholder rights and a weak legal framework. IV. Fund Holdings Relative to Market Next, we examine the characteristics of emerging market firms included in the portfolio holdings of U.S. mutual funds and compare them to all firms in the Worldscope database. 6 % Relative Spread was also measured simply as the difference between % Fund Portfolio Holding and % MSCI Index weight. 16

18 Worldscope covers a total of 4,136 firms for the countries in our sample. We assume this is the population of publicly traded firms in emerging markets available to U.S. mutual funds to make investment allocations. Table 8 shows the median values for firm characteristics classified in three groups. Group 1 consists of all 4,136 Worldscope firms; Group 2 consists of 3015 firms that are in Worldscope but are not included in the mutual fund portfolio holdings; and Group 3 consists of 1121 firms that are in Worldscope and in the mutual funds portfolio holdings. The firm-characteristics examined here are similar to those examined by Dahlquist and Robertson (2000) for Sweden, Kang and Stulz (1997) for Japan, and Falkenstein (1996) for the United States. These include market value of equity in U.S. dollars, annual dividend yield in percentage, ROE, and total debt to total capitalization. Market value is considered a proxy for size and total debt to total capitalization is a proxy for leverage. We again include market float, and number of analysts that follow the firm. Dahlquist and Robertsson (2001) find that foreign ownership is higher for Swedish stocks that have lower dividend yields and they conclude that this might indicate a preference for growth stocks. The correlation between some of these variables is significant at the 5% level therefore we estimate separate regression models. The firm characteristics examined are proxies for the size of the firm - market capitalization (in US dollars); leverage total debt/total capitalization; and profitability return on equity (ROE), and return on assets (ROA). Dividend yield, percentage float, number of analysts that follow the firm, and accounting quality are also examined. Accounting quality is the sum of four accounting attributes (auditor quality, accounting standard, consolidated statements, accounting opinion). A comparison of the three columns shows that U.S. mutual funds invest in firms that are larger as measured by market capitalization (MVALUE); have higher leverage as measured by 17

19 total debt to total capital (TDTCAP); have better profitability as measured by net income margin, ROE, ROA; and also have higher dividend yield relative to the average firm in Worldscope. Fund firms also have a higher float. On average, four analysts follow firms that are in fund portfolios compared to only one for all firms in Worldscope. The accounting quality is the same at two for each of the three categories. Table 9 presents correlation coefficients for some of the firm-level attributes. The correlation between market value and number of analysts; between market value and accounting standards; and between number of analysts and accounting standards is positive and significant at the 5% significance level. Therefore, in our multivariate framework we are careful about not including all of these attributes as explanatory variables in the same model. Next, we estimate a logit model to analyze differences between firms from Worldscope that are included in the portfolio holdings and those firms that are not included in the portfolio. In our models the dependent variable equals one if the Worldscope firm is included in the funds portfolio holding and zero otherwise. The explanatory variables are divided into three categories. The first category consists of firm characteristics, log of market capitalization (LMVALUEUSD), dividend yield (DYIELD), total debt to total capitalization (TDTCAP) and return on equity (ROE). At the firm-level we also include percentage market float (FLOAT), number of analysts that follow the company (NANALYSTS), and accounting standards (ACQUALITY). The country-level variables include market capitalization to GDP of country (MKTCAPGDP) and log of GDP per capita (LGDPPC). Industry dummies indicating 1-digit SIC industrial codes are included in the model. Three different models are estimated that include log of market capitalization in US dollars, number of analysts, and accounting standards separately in each model. These models are estimated separately because of the high correlation 18

20 between the three variables, market value, number of analysts that follow the firm, and accounting standards. We do not report results that include ROE and FLOAT because a number of observations are lost due to missing data on these two variables in Worldscope. Four main conclusions can be reached from the logistic regression results reported in Table 10. Funds invest more in firms from countries with better economic development as measured by log of GDP per capita (LGDPPC). However, funds invest less frequently in firms (relative to Worldscope firms in general) in firms from countries that have more developed equity markets as measured by market capitalization to GDP (MKTCAPGDP). The size and visibility of the firm as proxied by market value and number of analysts following the firm are both significant explanatory variables in determining fund investment. The coefficient of the firm s accounting standards is also significant. These results lead us to the conclusion that the size of the economy, size and visibility of the firm, and firm-level accounting standards are all important attributes that help firms to decide whether to decide in a firm or not. V. Fund Holdings Relative to MSCI Index In this section, the relationship between mean percentage mutual fund allocation to a firm and firm characteristics is analyzed. We examine the decision-making process from a fund manager s perspective. The manger has funds to invest and must first select the group of firms in which to invest as seen in the previous section. The next decision for the manager is to decide the percentage of the mutual fund s assets to allocate to a particular security holding. The allocation decision depends on the characteristics of the individual firm, the market and the country. 19

21 We address the question as to what extent fund holdings mimic a market index, because fund managers are often compensated based on their performance relative to a benchmark. One of the most widely used benchmarks for U.S. managers is the MSCI Emerging Markets Free Index, which is a free float-adjusted market capitalization weighted index that is designed to measure equity market performance in global emerging markets. These tests allow us to identify firm characteristics that attract institutional investment to firms that are and that are not included in the benchmark. The MSCI Emerging Markets index consisted of 708 securities at the end of The Index includes several securities of the same firm therefore the number of firms represented in the index is only 679. A. Univariate Analysis Table 10 provides a comparison of three groups. Group 1 consists of 791 firms that are included in out funds portfolio holdings but are nor in MSCI; Group 2 consists of 523 firms that are both in MSCI and are also included in our sample of funds portfolio holding; and Group 3 consists of 156 firms that are in MSCI but are not included in any of the fund s holdings. There is considerable overlap between firms included in MSCI and in portfolio holdings. However, funds invest in many other firms that are not part of the index. This is not surprising because these are actively managed funds and copying the MSCI index would not be considered active management. This highlights that funds are not simply mimicking the MSCI portfolio, but choose to invest in additional firms. Our tests discuss how funds choose to diversify their portfolios beyond the MSCI benchmark and the characteristics of firms that are not included in the MSCI index. The mean and median market capitalization of firms in the index is US$1.25 billion and US$366 million, respectively. The minimum and maximum values are US$9 million and US$40.65 billion, respectively. The index consists of both small and some very large firms. 20

22 The mean and median weight of the firms in the index is 0.124% and 0.04%, respectively. The minimum weight can be quite small and almost equal to zero percent and the largest weight in the index is 4.03%. The three firms with the largest weights in the index are Samsung Electronics, Taiwan Semiconductor and China Mobile. We first focus on analyzing the differences between characteristics of two groups of firms 1) the 791 firms in Group 1 that are included in the funds portfolio holdings but not included in the MSCI index, and 2) the 523 firms in Group2 that are included in the funds portfolio holdings and also included in the MSCI index. Table 10 shows that firms in the funds portfolio holdings that overlap with the MSCI index are larger in size based on market capitalization, have higher dividend yield and lower leverage. These firms perform better as measured by Net Income Margin, ROE, and ROA. The median number of analysts following firms that are included in both the funds holdings and in MSCI (Group 2) is 10, much higher than Group 1 and Group 3 firms. After examining the firm s financial characteristics we also analyze the impact of accounting reporting and corporate governance practices on fund allocations. The motivation is to examine the issue of whether accounting and corporate governance plays any role in the investment decisions of foreign mutual funds. Do foreign institutional investors care about these practices when deciding to invest in firms and countries? The four accounting-related independent variables- the quality of the auditor, consolidated reporting, accounting opinion, and the use of internationally recognized accounting standards have been combined to derive the measure of the firm s overall accounting standards. There is no difference in the median accounting quality of the three groups. B. Multivariate Analysis 21

23 We estimate regression models in order to further analyze the relationship between firm characteristics and ownership in a multivariate framework. The dependent variable used in the regression is mean proportion of the fund s assets invested in the firm expressed in percentage. The fund manager must decide how to allocate the funds that are available. The number of observations corresponds to the number of firms in our sample for which data is available. Table 12 reports results for several different regression models. The first set of regression includes the country-specific variables: log of GDP per capita (LGDPPC) and market capitalization to GDP (MKTCAPGDP). Industry dummies are also included in all the models. Firm-specific independent variables included in all models are dividend yield (DYIELD) and total debt to total capital (TDTCAP). Market value of the firm in US dollars and the number of analysts that follow the firm are each included separately because they are highly correlated. The coefficients of the industry dummies are not reported. The regression results are presented in Table 12. Funds are found to have a larger proportion of their assets invested in firms that are in countries with higher levels of economic development (LGDPPC) but that have less developed equity markets (MKTCAPGDPPC). Funds invest a larger proportion of their assets in larger firms (LMVAUEUSD) and firms that are covered by more analysts (NANALYSTS). Dividend yield is not significant in any of the models. Funds do invest more in firms that have lower leverage (TDTCAP). We next analyze the relationship between fund investment and shareholder rights at the country-level and accounting standards at the firm-level. We estimate three models: 1) includes only accounting standards, 2) includes only shareholder rights, 3) includes accounting standards, shareholder rights, and the interaction between them. The coefficient on firm-level accounting standards is positive and significant in the first model. The coefficient on shareholder rights 22

24 when included alone is positive but not significant. In the third model that includes both these variables and the interaction, both variables are positive and significant and the interaction is negative and significant. Therefore, we can conclude that funds invest a larger proportion of their assets in firms that have better accounting standards and this becomes particularly important in countries that do not have strong shareholder rights. Finally, we also examine a fund s investment relative to the weight of the firm in the MSCI index. The weights of firms in the MSCI index as discussed above are at the level of all emerging markets but they can be reconstructed at the regional and country level. A measure of relative fund allocation is calculated as follows for each fund category: %Relative Fund Allocation in Firm = %Fund Allocation - %MSCI Weight; Therefore, relative fund allocation can be positive, negative, or zero. It will be positive if the fund allocates a larger percentage of its assets to the firm than the market index. This will certainly be the case for the 791 firms that are part of the fund portfolio holdings but are not included in the MSCI index. Relative fund allocation will be negative if the fund allocates a smaller percentage of its assets to a firm than the index does. This will be true of at least the 156 firms that are part of the MSCI index but are not included in the fund portfolios. Funds allocate even more than the MSCI weight to larger firms and firms with more analysts following leverage as shown in Table 13. A number of conclusions can be reached from the regression results. First, size of the firm is an important consideration for portfolio managers. Size as measured by log of market value is statistically significant in all the models estimated. Size could be picking up other firm characteristics such as visibility and/or liquidity. It may also be a proxy for firms that are in major market indexes. Managers may prefer firms with which they are more familiar and about 23

25 which they have more knowledge. In this section we introduce four variables to capture visibility and liquidity and in the next section we will also examine the impact of being included in a major market index. The ADR dummy has a value equal to one if the firm has an ADR and it is hypothesized that U.S. fund managers will have more information about firms that have ADRs. If an ADR exists then the manager has the option to invest in the firm by investing in the stock in the local market or by investing in the ADR or by investing in both at the same time. Some managers may have a preference for holding ADRs rather than stock in the local market for transactions costs and settlement reasons. In addition to the ADR dummy, we also use volume turnover ratio, value turnover, and the float percentage as proxies for liquidity (not shown). Surprisingly, the ADR dummy, share volume turnover and value turnover are not statistically significant in any of the models and the results are not reported. 7 This result is not consistent with findings for other countries. In order to examine the effect of liquidity we also tried the variable float as an independent variable, which is not found to be significant. Float is measured based on shares closely held as reported by Worldscope. Float has a mean and median value of approximately 50% for the emerging market firms in our sample, although the standard deviation is equal to about 25%. The problem could be that there is a lot of missing data for the float variable so the sample size in the regressions goes down considerably. Accounting standards and shareholder rights are not significant explanatory variables when included in separate regressions. However, in the model that includes both attributes together along with the interaction term, both coefficients are positive and significant. In addition, the interaction term has a negative and significant coefficient. This implies that both the firm-level accounting standards and the country-level shareholder rights are important 7 Volume turnover and value turnover are obtained from the Emerging Markets Database and is missing for many firms. 24

26 determinants of fund allocations. The negative coefficient on the interaction term implies that firm-level accounting standards and country-level shareholder rights are particularly important when both tend to be weak. V. Summary and Conclusions We have examined attributes of countries and firms that are part of the holdings of U.S. emerging market funds. Both financial and corporate governance attributes at the country and firm-level are analyzed. Portfolio holdings are examined relative to the population of emerging market firms covered in Worldscope and also relative to a major market benchmark, MSCI Emerging Markets index. In addition to financial characteristics, such as firm size, leverage and returns, we find that accounting disclosure quality is equally important. However, liquidity and float are not found to be significant. Funds are found to invest in many more firms than a market benchmark. Fund allocation is higher than the MSCI weight in larger firms, firms with greater analyst coverage, firms with lower leverage and higher returns. A larger proportion of the funds assets are invested in firms that have higher quality accounting disclosures. Our results suggest that investment by foreign institutions depends not only on the firm s financial attributes but also on accounting standards and corporate governance. 25

27 References Amihud, Yakov, Haim Mendelson, and Jun Uno, 1999, Number of shareholders and stock prices: Evidence from Japan, Journal of Finance 54, Bailey, Warren, Y. Peter Chung, and Jun-Koo Kang, 1999, Foreign ownership restrictions and equity price premiums: What drives the demand for cross-border investments? Journal of Financial Quantitative Analysis 34, Bailey, Warren, and Julapa Jagtiani, 1994, Foreign ownership restrictions and stock prices in the Thai capital market, Journal of Financial Economics 36, Brennan, Michael, and H.Henry Cao, 1997, International portfolio investment flows, Journal of Finance 52, Covrig, Lau, and Lilian Ng, 2002, Do domestic and foreign fund managers have similar preferences for stock characteristics? A cross-country analysis, Working Paper, Nanyang Technological University. Dahlquist, Magnus, and Goran Robertsson, 2001, Direct foreign ownership, institutional investors, and firm characteristics, Journal of Financial Economics 59, Drogt, Emily L., 2002, Mutual fund investment decisions: the interaction of accounting quality and legal enforcement, working paper, University of Michigan. Falkenstein, Eric G., 1996, Preferences for stock characteristics as revealed by mutual fund portfolio holdings, Journal of Finance 51, Gompers, Paul and Andrew Metrick, 1999, Institutional investors and equity prices, Quarterly Journal of Economics, forthcoming. Jordan, Cally and Mike Lubrano, 2002, How effective are capital markets in exerting governance on corporates?: Lessons of recent experience with private and public legal rules in emerging markets, working paper. Kaminsky, Graciela L., Richard K. Lyons and Sergio Schmukler, 2001, Mutual fund investment in emerging markets: An overview, World Bank Economic Review 15, Kang, J.K. and Rene Stulz, 1997, Why is there a home-bias? An analysis of foreign portfolio equity ownership in Japan, Journal of Financial Economics, 46, King, Robert and Ross Levine, 1993a, Finance and growth: Schumpeter may be right, Quarterly Journal of Economics, 108, King, Robert and Ross Levine, 1993b, Finance, entrepreneurship, and growth: Theory and evidence, Journal of Monetary Economics, 32,