Investor Protection and the Role of Firm-Level Financial Transparency. in Attracting Foreign Investment

Transcription

1 Investor Protection and the Role of Firm-Level Financial Transparency in Attracting Foreign Investment Bowe Hansen a, Mihail K. Miletkov b*, M. Babajide Wintoki c a Pamplin College of Business, Virginia Tech University b Paul College of Business and Economics, University of New Hampshire c School of Business, University of Kansas, Lawrence Abstract We ask if companies can attract foreign equity capital by improving the transparency of their financial statements. Using a large panel of firms across fifty-one countries outside the U.S., we show that the answer is yes, but only in countries with relatively high levels of investor protection. In countries with poor investor protection, unilaterally increasing firm-level transparency has no effect on foreign ownership. Furthermore, our results indicate that in countries with higher levels of investor protection the positive association between transparency and foreign ownership is stronger following a country s adoption of the International Financial Reporting Standards. Keywords: foreign ownership, home bias, financial transparency, institutional development, investor protection JEL Classifications: G32, G34 *Corresponding author: Mihail K. Miletkov, Paul College of Business and Economics, University of New Hampshire, Durham, NH 03824, USA; Phone: (603) ; mihail.miletkov@unh.edu. We are grateful for the research support of Virginia Tech University, the University of New Hampshire, and the University of Kansas. We appreciate comments from Valaria Vendrzyk (the AAA discussant), David Volkman (the Southern Finance Association discussant), Bryan Cloyd, Marc Lipson, Mitch Oler, and workshop participants at Virginia Tech University, the 2012 AAA Annual Meeting, and the 2013 Virginia Accounting Research Colloquium, as well as from two anonymous referees. 1

2 1. Introduction Access to foreign equity capital provides benefits for individual firms and for entire countries. At the firm-level, foreign investors expand the firm s shareholder base and reduce its cost of capital, and at the country-level, foreign investment increases aggregate investment and promotes economic growth. 1 Despite the fact that over the past three decades, many of the direct and indirect barriers to foreign portfolio investment have been lifted, the observed levels of foreign investment are still significantly below what is considered optimal under contemporary finance theory (French and Poterba, 1991; Tesar and Werner, 1995; Ahearne, Griever and Warnock., 2004; Chan, Covrig and Ng, 2005). One of the main explanations for this home bias in investors portfolio allocation decisions is the higher level of information asymmetry that investors face when investing abroad. We examine whether firms can attract foreign investors by implementing more transparent financial reporting practices, and the effect that the level of investor protection in the firm s home country has on the relationship between financial transparency and foreign ownership. Prior literature suggests that a country s overall level of investor protection, mandatory adoption of International Financial Reporting Standards (IFRS), and firm-level financial transparency are important determinants of aggregate and firm-level foreign investment (Aggarwal, Klapper, and Wysocki, 2005; Leuz, Lins, and Warnock, 2009; Khurana and Michas, 2011; Florou and Pope, 2012). 2 However, our study is the first to examine the interaction between country-level investor protection and changes in firm-level financial transparency, and 1 A partial list of studies examining the effects of foreign equity investment on firm and country-level outcomes includes: Bekaert and Harvey (2000); Blair Henry (2000a, 2000b, 2003); Chari and Blair Henry (2004); Bekaert, Harvey and Lundblad (2005, 2006); Moshirian (2008); Huang and Shiu (2009), among others. 2 One potential limitation of these prior studies is that most of them are based exclusively on foreign investments by U.S. institutional investors. In contrast, our study is based on different types of foreign investors (institutional investors, governments, individuals etc.) from multiple countries. 2

3 the effect this interaction has on a firm s ability to attract foreign investment. In other words, our study seeks to answer the question of what happens to foreign ownership, within an individual firm, when the firm s level of financial transparency changes, given the prevailing level of overall investor protection in the firm s home country. This is an important research question because any differences in the association between firm-level transparency and foreign ownership in countries with strong investor protection (as compared to that in countries with poor investor protection) have direct implications for firm management. Indeed, the biggest contribution of our paper is to demonstrate that the ability of individual firms to use financial reporting transparency as a mechanism for attracting foreign equity investment can be significantly constrained (or enhanced) by the level of investor protection in the firm s home country. Using three different proxies for investor protection, we show that the positive relationship between firm-level transparency and foreign ownership is driven by firms located in countries with higher levels of investor protection. In countries where investor rights are poorly protected and the risk of minority shareholder expropriation is greater, firm-level financial transparency does not significantly affect the portfolio allocation decisions of foreign investors. These results hold across different model specifications and estimation methods, and are robust to the use of alternative measures of financial transparency. Firm-level transparency and countrylevel investor protection turn out to be complements rather than substitutes. If a firm is located in a strong investor protection country, it can increase foreign ownership by increasing the transparency of its financial statements. If the firm is in a weak investor protection country, increasing transparency has no effect of foreign ownership. 3

4 There are several potential explanations for our findings. One possibility is that in countries with weak investor protection, foreign investors question the credibility of financial reporting of all firms regardless of their level of transparency. Another possibility is that, in countries with lower levels of investor protection, financial transparency may not necessarily reduce the risk of minority shareholder expropriation, which may be the main deterrent to foreign equity investment. Johnson, La Porta, Lopez-de-Silanes, and Shleifer (2000) show that the various self-dealing transactions that company insiders and controlling shareholders engage in (such as transfer pricing, excessive compensation, loan guarantees, asset sales, and dilution) are not illegal in many countries. Therefore, in these countries, even if foreign investors are better able to identify potential cases of minority shareholder expropriation in firms with more transparent financial reporting practices, they may still be reluctant to provide capital to these firms due to the lack of legal recourse available to them. Our findings indicate that in their quest for foreign capital, firms are captive to the investor protection environment in their home countries. In countries where investor rights are better protected, increasing firm-level financial transparency is an effective mechanism for attracting foreign equity investors. On the other hand, in countries where investor rights are not adequately protected, firms are unable to attract foreign shareholders by unilaterally committing to higher levels of financial transparency. Therefore, in order to attract foreign equity investment in their firms, countries need to make more substantial investments in the legal and extra-legal institutions protecting investor rights. Our results also support and extend the theoretical and empirical arguments in Doidge, Karolyi and Stulz (2007). While their study does not explicitly consider foreign ownership, they argue that below a threshold level of investor protection and economic development, individual firm characteristics explain very little of investor perception 4

5 of a firm s governance quality, and that only when the country has a high level of investor protection can an individual firm take affirmative steps to improve its internal governance in a way that signals its quality to investors. Many of the countries in our sample adopted International Financial Reporting Standards (IFRS) during our sample period. We specifically investigate the impact that this change in the regulation of financial reporting had on the relationship between firm-level transparency and foreign ownership. Previous literature suggests that IFRS provides more comprehensive disclosure requirements than most domestic GAAP (Ding, Hope, Jeanjean, and Stolowy, 2007; Bae, Han, and Welker, 2008), and is associated with an increase in foreign investment, especially in countries with higher levels of investor protection (Khurana and Michas, 2011; Florou and Pope, 2012). IFRS adoption can affect the association between firm-level transparency and foreign ownership in at least one of two ways. On one hand, to the extent that IFRS adoption improves the accounting quality of all firms within a country, it is possible that it reduces (weakens) the importance to foreign investors of firm-level variations in transparency. On the other hand, Byard, Li, and Yu (2011) find that IFRS adoption actually magnifies the value of firm-level transparency to analysts. If this is true for foreign investors as well, we expect that national IFRS adoption will strengthen the association between foreign ownership and firm-level transparency. Our findings suggest that the latter effect appears to be the case in countries with higher levels of investor protection; in these countries, the positive effect of firm-level financial transparency on foreign ownership is magnified following national adoption of IFRS. However, for firms in countries with weak investor protection, even in conjunction with a country s adoption of IFRS, a firm s transparency remains unassociated with foreign ownership. 2. Literature Review and Hypotheses Development 5

6 Prior literature suggests that information asymmetries between foreign and domestic investors represent a major impediment to foreign investment. One method by which firms can reduce information asymmetries, and thus potentially increase foreign ownership, is by preparing more transparent financial reports (Young and Guenther, 2003). The evidence in prior literature also suggests that foreign investment is directly impacted by the level of investor protection in a country (Gelos and Wei, 2005; Aggarwal, Klapper, and Wysocki, 2005; Ferreira and Matos, 2008; Giannetti and Koskinen, 2010; Megginson, You and Han, 2013). However, it is less clear how the interaction of country-level investor protection and firm-level reporting transparency impacts foreign investment. In other words, it is not clear by how much an individual firm can increase its foreign ownership by increasing its own individual level of financial transparency, given the institutional constraints it faces in its home country. This is the central question we seek to address in this study. The existing evidence on the interaction between firm-level transparency and countrylevel investor protection is mixed. In a sample of firms in thirty emerging market countries, Aggarwal, Klapper, and Wysocki (2005) find a relation between accounting quality and U.S. institutional ownership, but only in firms with unlisted ADRs. In contrast, Leuz, Lins, and Warnock (2009), using a sample of firms in twenty-nine countries (weighted more towards higher investor protection countries than Aggarwal, Klapper, and Wysocki, 2005), find an association between financial transparency and foreign ownership. However, even the mixed inference from both of these studies is limited in several ways. Neither paper actually examines what happens to foreign ownership within a firm when it changes its level of financial transparency relative to other firms within that country. This is partly as a result of the data limitations in those studies; both Aggarwal, Klapper, and Wysocki 6

7 (2005) and Leuz, Lins, and Warnock (2009) use a single snapshot of foreign ownership (from 2001 and 1997, respectively) and cannot measure within firm variation in foreign ownership in response to changes in firm-level transparency This would require a panel of firms over several years. In addition, both papers focus exclusively on US institutional ownership. While the U.S. is the biggest single source of foreign investment around the world (about 37% in our sample), it is not the majority source, and other sources of foreign investment have become increasingly important in recent years. Finally, these studies focus on a relatively small number of countries around the world, potentially limiting the cross-sectional variation in investor protection in their studies. There are two possible ways in which firm-level transparency and country-level investor protection could interact, and we view our study as exploring these alternatives. On one hand, firm-level transparency could be a substitute for the poor quality of investor protection in a particular country. Adopting more transparent financial statements could offer a way for managers to signal their quality to foreign investors and separate themselves from the low quality institutional environment of their country similar to the way firms might signal their commitment to higher governance standards by cross-listing in countries with better securities regulations (Pagano, Roell and Zechner 2002; Siegel, 2005). On the other hand, firm-level transparency and country-level investor protection could be complements. Besides the occasional risk to life and limb that may come from having to physically monitor investments in places with weak legal institutions, the possible risk of minority shareholder expropriation is a major deterrent to foreign equity investment in such countries. In these countries, even if foreign investors are better able to identify potential expropriation in firms with more transparent financial reporting practices, they may still be 7

8 reluctant to provide capital to these firms due to the lack of legal recourse available to them should the firms controlling insiders attempt to expropriate minority shareholders. Furthermore, Johnson La Porta, Lopez-de-Silanes, and Shleifer (2000) show that in many countries, especially those with weak legal institutions, the various self-dealing transactions that company insiders and controlling shareholders engage in (such as transfer pricing, excessive compensation, loan guarantees, asset sales, and dilution) are not illegal (i.e., they are consistent with both statutes and principles followed by judges). The hypothesis that firm-level transparency and country-level investor protection could be complements rather than substitutes at certain levels of investor protection is further underscored by arguments made by Doidge, Karolyi and Stulz (2007). They present a model, and evidence, suggesting that after controlling for country level characteristics, individual firm characteristics have almost no explanatory power for the governance ratings of firms in less developed countries. They argue that while companies in high investor protection countries may receive a value premium for better governance, companies in low investor protection countries do not, since the low overall quality of investor protection means that investors significantly discount any effort by firms in those countries to improve their governance quality. Recent work by Megginson, You, and Han (2013) further underlines the importance of country characteristics in determining foreign investment; they find that, regardless of origin, sovereign wealth fund investments tend to flow to countries with higher levels of investor protection. Thus, we have two alternative hypotheses with respect to the intersection between country-level investor protection, firm level transparency, and foreign ownership. If countrylevel investor protection and firm-level transparency are complements, we expect a stronger (positive) relationship between transparency and foreign ownership in high investor protection 8

9 countries than in low investor protection countries. If they are substitutes, we expect the opposite. 3. Data and Methodology 3.1. Descriptions of Key Variables Foreign Ownership We measure our dependent variable, Foreign Ownership, as the percentage of the firm s total shares outstanding held by foreign investors as of year t. 3 We obtain firm-level ownership data from the OSIRIS database, which is a product of Bureau van Dijk Electronic Publishing. The shareholder information provided in OSIRIS includes the investors name, type, percent of shares owned, and nationality. Bureau van Dijk lists the following information sources, which are used to compile the ownership data in OSIRIS: company annual reports, information from official registers/regulatory bodies and associated information providers, private correspondence between Bureau van Dijk and the respective companies, and information from stock exchanges and company websites. The firm-level ownership data from OSIRIS (and its European counterpart Amadeus, which is also a product of Bureau van Dijk Electronic Publishing) has been recently used by Li, Moshirian, Pham, and Zein (2006), John, Litov, and Yeung (2008), Faccio, Marchica, and Mura (2011), and Lins, Volpin, and Wagner (2013) among others. The OSIRIS database is unique when compared to the information sources used in most prior studies on foreign ownership in that it covers all types of foreign owners (i.e., institutions, individuals, governments, etc.). The vast majority of prior literature in the area has used holdings by foreign mutual funds collected from Thomson Financial Services as a proxy for total foreign 3 We classify shareholders as foreign if they are domiciled in a country other than the firm s country of incorporation. If the firm s main country of operation is different from the firm s country of incorporation (often due to tax considerations), we also require that the foreign shareholders are domiciled outside of the firm s main country of operation. 9

10 holdings. Thus, our data source vastly increases the number, type, and nationality of foreign owners when compared to prior literature Transparency We measure observable financial reporting transparency at the firm level by adapting the reporting behavior score measure from Daske, Hail, Leuz, and Verdi (2013). The measure is based on attributes of the firm s reporting observable to investors prior to their investment decision. Specifically, we first measure the absolute value of accruals divided by the absolute value of cash flows multiplied by negative 1. Multiplying by negative 1 increases the intuitiveness of the result as larger (less negative) numbers can be interpreted as indicative of stronger reporting incentives. Consistent with Daske, Hail, Leuz, and Verdi we calculate this measure over the three-year time period: t-3, t-2, and t-1, and use the average over the three year period to reduce the impact of noise in individual year measures. We modify the Daske, Hail, Leuz, and Verdi measure as follows: each year, we rank each firm on the measure and divide the firms into 10 equally sized deciles to form our measure of reporting incentives (Transparency). We use the deciles because the raw measure is truncated at zero by definition and has a significant number of outliers in the left tail, resulting in a highly skewed distribution. Ranking the observations removes the skewness and increases the interpretability of our results. 4 The use of this measure as a proxy for transparency is based on evidence that accruals and cash flows provide different information (Sloan, 1996), and the argument that large accruals could indicate that managers are using their discretion to provide misleading financial statements (Leuz, Nanda, and Wysocki, 2003). Wysocki (2004) shows that this measure is associated with several other earnings management proxies used in international settings. Since an implicit goal 4 We repeat all of our analysis using quartiles, percentiles, and the raw data after separately truncating, or winsorizing, the bottom 1% of the data and obtain qualitatively similar results. 10

11 of earnings management is to conceal information from financial statement users, evidence of less earnings management indicates more transparent financial information Investor Protection Our primary interest is in evaluating the impact that country-level investor protection has on the relationship between firm-level financial transparency and foreign ownership. We use three alternative measures of country-level investor protection to ensure the robustness of our results. First, we use the Legal Structure and Security of Property Rights Index from Economic Freedom of the World (Gwartney and Lawson, 2007). This index is designed to reflect the overall legal environment as it relates to the protection of property rights and the quality of legal institutions. We view this as a measure of a country s overarching institutional quality which represents a valid measure of the overall risk to an investor of investing in that country. An important feature of the index is that it does not simply reflect laws on the books, but rather the overall legal environment as it relates to the protection of the property rights and the overall quality of the legal institutions. The index is assessed on a scale of 0 to 10, with 0 being the lowest and 10 being the highest. It is constructed from the following seven key elements based on data from the International Country Risk Guide (ICRG), the World Economic Forum s Global Competitiveness Report (GCR), and the World Bank s Doing Business project (DB): (1) Judicial independence: the judiciary is independent and not subject to interference by the government or parties in disputes (GCR); (2) impartial courts: A trusted legal framework exists for private businesses to challenge the legality of government actions or regulations (GCR); (3) protection of property rights (GCR); (4) military interference in rule of law and the political 11

12 process (ICRG); (5) integrity of the legal system (ICRG); (6) legal enforcement of contracts (DB); and, (7) regulatory restrictions on the sale of real property (DB). The second investor protection measure we use is the anti-self-dealing index from Djankov, La Porta, Lopez-de-Silanes, and Shleifer (2008). This index focuses on national legal institutions and private mechanisms designed to monitor and prevent self-dealing transactions. The index includes measures of requirements to disclose self-dealing transactions and related enforcement mechanisms. We include this measure because investors are less likely to invest in firms if they suspect that company insiders will expropriate wealth from the firm, and foreign investors are likely to face higher costs of monitoring such self-dealing transactions (Leuz, Lins, and Warnock, 2009). Finally, we use the country-level earnings quality measure from Leuz, Nanda, and Wysocki (2003). 5 This measure is created by ranking countries on four measures of earnings quality. Each country is assigned a ranking equal to the average of its four ranks. We include this measure because it is based on observed measures of financial reporting quality and is a de facto measure of the output of a country s financial reporting institutions. The four measures on which the accounting credibility measure is based are a country s: median ratio of the standarddeviation of operating cash earnings to the standard deviation of operating cash flows, contemporary correlation between changes in accruals and changes in operating cash flows, median of our firm-level measure of transparency, and, ratio of small losses to small profits where small is defined as after-tax earnings, or losses, of 1% or less of assets. These three measures of investor protection are based on independent sets of information, and were created by different sets of researchers in the legal, economics, finance, and accounting 5 DeFond, Hung, and Li (2011) use the earnings quality score, relabeled as Implementation Credibility, in their investigation of the role of comparability in determining the impact of IFRS adoption on foreign ownership. 12

13 disciplines. Finding consistent results across all three measures would strongly indicate that our results are not driven by any one interpretation of what constitutes strong investor protection ; instead, it would suggest that our results are driven by the core construct underlying the idea of investor protection Sample Selection and Methodology Our sample consists of all firms for which we are able to calculate a firm-level Transparency score, and for which we are able to obtain ownership data and data on the firmlevel control variables used in the subsequent regression analyses. We collect firm-level data for the years 2001 through 2011 from the OSIRIS and Worldscope databases, and country-level data from the World Bank s World Development Indicators database. The intersection of these data sets provides us with 55,239 firm-year observations from 51 countries. 6 Descriptive statistics regarding foreign ownership, transparency, and each of our measures of investor protection are provided by country in Appendix A. 7 As our research question focuses on the ability of firms to attract foreign investment by improving their financial reporting transparency, we conduct all of our analyses at the firm level. We start our analysis using a research design that estimates the effect of financial reporting transparency on the level of foreign ownership, as follows: 6 The intersection of the various databases actually yields 69,002 firm year observations. However, since all our estimation techniques are within firm regressions, we can only use firm year observations for firms that have at least two observations in the case of our fixed effects regressions, two consecutive observations in the case of first difference regressions, or three consecutive observations in the case of dynamic GMM regressions. We also drop observations from countries with less than 30 firm-year observations. Thus, our biggest regression sample is the 55,239 for the fixed effects regressions. 7 Appendix A illustrates there are several countries for which we have relatively limited coverage. To ensure that our results are not affected by cross-country differences in the degree to which firms in different markets are covered, in un-tabulated analyses we re-estimate all of our regressions using only country-years for which we have firms representing at least 50% of the country s total market capitalization, or at least 50% of all publicly traded firms in the respective country (the total market capitalization and number of listed firms for each country-year are obtained from the World Bank s World Development Indicators database). All of our results are robust to these alternative estimations. 13

14 Foreign Own i,t = α 1 + β 1 Transparency i,t + β j Control i,j,t + ε i,t (1) where Foreign Own equals the percentage of the firms shares outstanding owned by foreign shareholders as of year t; Transparency equals the decile rank (1 to 10) for firm i in year t and Controls are variables found in prior literature to be significantly related to foreign ownership. We include controls for firm visibility, performance, and risk/growth characteristics which prior literature (DeFond, Hung, and Li, 2011, Yu, 2010, among others) has found to be related to cross-border investment. We include country-level controls measuring the level of economic and financial development and the degree of financial market integration and capital account openness in the firm s home country. Finally, we include year fixed effects in all of our analyses. The control variables are defined in Appendix B. There are at least two sources of endogeneity that may bias a simple ordinary least squares (OLS) estimate of equation (1). First, there is unobserved heterogeneity that may jointly affect both the level of foreign ownership and firm-level transparency. Such unobservable heterogeneity could include country-level variables that affect all firms within a country. Second, an additional potential source of endogeneity that may arise depending on how we choose to estimate (1) is dynamic past levels of foreign ownership may affect the current level of transparency. 8 One way to potentially account for these sources of endogeneity would be to use a strictly exogenous instrument for transparency and carry out a two-stage least squares (2SLS) regression; a strictly exogenous instrument would be a strong determinant of firm-level transparency while also being exogenous with respect to foreign ownership. However, none of the observable firm characteristics or variables in our data set (such as the control variables in j 8 Aggarwal, Erel, Ferreira, and Matos (2011) suggest that increases in international institutional investment lead to subsequent improvements in firm-level governance measured using a broad governance index as well as specific measures of corporate governance such as board independence and adoption of a staggered board provision. 14

15 Appendix B) meet these twin criteria. Given that weak instruments could introduce biases in 2SLS that are even bigger than those in OLS regressions (Stock and Yogo, 2005) we use a variety of other methods to account for the potential endogeneity in the relationship between firm-level transparency and foreign ownership. To account for unobservable heterogeneity, we employ both a first-differences specification and a firm fixed-effects specification. These methods account for unobserved heterogeneity (as long as we believe the form of unobserved heterogeneity is time-invariant) but may yield inconsistent estimates if the form of endogeneity is dynamic. To account for the possibility that there is dynamic endogeneity, we also estimate (1) using the system Generalized Method of Moments (GMM) estimator developed in a series of papers by Arellano and Bond (1991), Arellano and Bover (1995), and Blundell and Bond (1998). This procedure allows us to account for the potential effect of past foreign ownership on current levels of transparency while still accounting for unobserved heterogeneity using firm fixedeffects (see, for example, Beck, Levine, and Loayza, 2000; Beck, 2002; Beck and Levine, 2004; Hoechle, Schmid, Walter, and Yermack, 2012; Wintoki, Linck, and Netter, 2012). The GMM procedure involves including two lags of (past) foreign ownership in the regression equation, and then using lags of foreign ownership and transparency (and other firm characteristics) from time t-3 and t-4 as instruments for the level of transparency at time t. The assumption is that the firm's past foreign ownership and transparency (and other historical characteristics) are important determinants of the firm s current level of transparency. The key exogeneity argument is that while the firm's historical foreign ownership and historical characteristics may affect the firm s current level of transparency, foreign ownership (and firm characteristics) from time t-3 and t-4 is exogenous with respect to current foreign ownership. The firm's history from t-3 and t-4 only 15

16 affects current foreign ownership through the two lags of foreign ownership (from t-1 and t-2) that are included in the regression equation. In other words, dynamic GMM estimation can be considered an estimation of the following system: k=2 Foreign Own i,t = α 1 + γ k Foreign Own i,t k + β 1 Transparency i,t + β j Control i,j,t + ε i,t (2) k=1 j k=2 k=4 Transparency i,t = α 2 + a k Foreign Own i,t k + b j Control i,j,t + π k Z i,t k + μ i,t (3) k=1 j k=3 k=2 k=4 Control i,j,t = α 3 + c k Foreign Own i,t k + d 1 Transparency i,t + φ k Z i,t k + ε i,t j (4) k=1 k=3 where Z is a vector of foreign ownership, transparency, and other firm characteristics from lags t- 3 and t-4, respectively. Our exogeneity assumption, as described above, is that once we have included the two lags of foreign ownership Foreign Own t 1 and Foreign Own t 2, respectively, then, current foreign ownership (Foreign Own t ) is exogenous with respect to Z t 3 and Z t 4,respectively; and, that these can be used as instruments for current transparency and other firm characteristics. In our analysis, we use system GMM estimation in which (2) is estimated in a stacked system consisting of the equation in levels and first differences, with levels and first differences from t-3 and t-4 used as instruments in the first-differences and levels equations, respectively. 9 9 The choice of how many lags of foreign ownership to include in the GMM regression is a question and our selection is driven by two factors: (1) the need to ensure that k is large enough to ensure that the (lagged) instrument set is far enough into the past to be truly exogenous, but (2) is not so far into the past as to be a weak instrument (See Wintoki, Linck, and Netter, 2012 for a discussion of this issue). In all specifications where we use the system GMM estimator we carry out (unreported) tests of the validity of our instrument set as suggested by Arellano and Bond (1991), based on a test of second order serial correlation and a Hansen test of over-identification. In all cases, we find that we cannot reject the null hypothesis that our instruments are valid. 16

17 A key aspect of our paper is that we investigate the effect of country-level investor protection on the relationship between firm-level transparency and foreign ownership. We divide the firms, at the sample median, into those in countries with strong investor protection and those in countries with weak investor protection. We estimate the regression specified in (1) or (2) separately for firms in strong and weak investor protection countries. We do this for each of the three measures of investor protection which we described in section A3: the property rights index, the anti-self-dealing index, and the country earnings quality measure Sample Description We present summary statistics for foreign ownership, our investor protection proxies, and our control variables in Table 1. The mean level of foreign ownership in our sample is 8.01%. However, the median is only 1.24%. Slightly less than 2% of our firm-year observations are cross-listed in the United States. However, 26% of them are included in a market index, and 62% use a Big N Auditor. The average firm in our sample has a small positive ROA, at both the mean and the median. [Insert Table 1 about Here] In Table 2, we present correlations among foreign ownership and the explanatory variables used in our analyses. Consistent with our expectations, there is a positive and significant correlation between transparency and foreign ownership, though the correlation coefficient is only All of our control variables are significantly correlated with foreign ownership. With regard to our proxies for investor protection, foreign ownership positively correlates with both the property rights index and the anti-self-dealing index, but negatively correlates with earnings quality. Our measure of firm-level transparency positively correlates with the property 17

18 rights index and earnings quality, but negatively correlates with the anti-self-dealing index. Within the group of institutional quality proxies, the property rights index and the anti-selfdealing index have a positive correlation of Earnings quality negatively and significantly correlates with both the property rights index and the anti-self-dealing index with correlation coefficients of , and , respectively; this is to be expected since by construction, a higher earnings quality measure suggests lower investor protection. While our three investor protection measures correlate as would be expected, they do not perfectly correlate. These correlations support our view that the three measures represent alternative measures of institutional quality, and that consistent results across the three measures indicate that our results are not driven by an idiosyncrasy of any one measure. The largest correlation between any two variables, 0.743, is between the property rights index and GDP per capita. This large correlation, however, does not create a concern about multicollinearity since we estimate our regressions by splitting firms based on the strength of investor protection in their home countries, rather than by including this variable as an additional explanatory variable in the regression model. [Insert Table 2 about Here] 4. Results 4.1. Financial transparency and US and non-us foreign ownership. We examine the relationship between transparency and foreign ownership in firms across the world. We pool all foreign investors (U.S. and non-u.s.) into one group and estimate the relationship between foreign ownership and firm-level financial transparency using the model presented in (1). We present the results of this analysis in Table 3. For all three models, we find a 18

19 positive and significant relationship between a firm s level of financial reporting transparency and foreign ownership. The coefficients range from a low of (t = 1.99) for the firstdifferences model, to a high of (t = 3.56) for the dynamic GMM model. These coefficients correspond to an increase in the level of foreign ownership of approximately 20 to 43 basis points for a 5-decile increase in Transparency (i.e., an increase equivalent to a shift from the 25th percentile to the 75th percentile). While this may seem like a modest increase in the level of foreign ownership, it is economically significant given that the mean level of foreign ownership for the full sample is only 8.01%, and the median is only 1.24%. [Insert Table 3 about Here] The results in Table 3 show that, on average, there is a positive association between foreign ownership and transparency across all foreign equity investors. However, our sample shows the average level of U.S. foreign ownership is 2.97%, or approximately 37% of all foreign investment. Therefore, it is possible that this result is driven by U.S. investors. To determine if this may be the case, we repeat the analysis from Table 3 separately using U.S. foreign ownership and non-u.s. foreign ownership as our dependent variables. We present the results of these analyses in Table 4. Under all three specifications, the coefficient on transparency is positive and significant both when we perform the analysis using only U.S. investors, and when we perform the analysis excluding U.S. investors. Our results indicate that both U.S. and non- U.S. foreign investors gravitate toward the stock of firms with more transparent financial reporting practices. [Insert Table 4 about Here] Leuz, Lins, and Warnock (2009) and Aggarwal, Klapper, and Wysocki (2005) base their results on U.S. institutional ownership only. Our data set allows us to examine whether or not the 19

20 positive relationship between financial transparency and foreign ownership applies to noninstitutional investors as well. Prior evidence indicates that institutional investors are effective and efficient users of financial reporting information (Cohen, Gompers and Vuolteenaho, 2002; and Barber and Odean, 2008). In contrast, individual investors are more likely to be subject to behavioral biases and behavioral processing constraints (Cohen, Gompers, and Vuolteenaho, 2002, and Field and Lowry, 2009). Further, investments by industrial firms are more likely to be driven by strategic motivations, such as, tax implications, diversification, or, most likely in this case, an effort to gain access to a foreign market. Similarly, government investment is likely to be influenced by strategic political goals; and, sovereign wealth fund investments may be influenced by political motivations, as well as regulations designed to limit the riskiness of their investment portfolios. Therefore, to the extent that our measure, Transparency, accurately reflects the information environment of the firm, it is likely that this variable will play a more significant role in the portfolio allocation decisions of institutional investors, and a less significant role in the investment decisions of other types of investors. [Insert Table 5 about Here] The results of this analysis, shown in Table 5, indicate the association between foreign ownership and transparency does not generalize to non-institutional investors. 10 The coefficient on Transparency is positive and significant for all three specifications when we use foreign institutional ownership as the dependent variable, but is insignificant in all cases when foreign non-institutional investment is used as the dependent variable. 11 To the extent that institutional 10 In some cases, the Osiris database does not indicate the type of the investor. We exclude firm-year observations missing this information from this test. This results in a loss of just under 7% of our sample. 11 Additional un-tabulated analyses show that the association between transparency and foreign ownership is insignificant for each subset of non-institutional investors: individuals/families, governments, and other industrial firms. In contrast, the relation is significant for all the classes of institutional investors that we can identify: mutual funds, banks, insurance companies and other financial firms (e.g. pension funds). 20

21 investors are more likely to rely on, and are better able to interpret accounting information than other investor groups, these results support our assertion that Transparency is a good proxy for the information environment of the firm. Collectively, these results indicate that reporting transparency is not associated with foreign investment by individuals, industrial firms, or governments. However, it plays an important role in a firm s ability to attract both U.S. and non-u.s. foreign institutional investors. Since foreign institutional investment makes up over 79% of total foreign investment in our sample, we conclude that reporting transparency can be a significant factor in determining a firm s ability to attract foreign equity capital in general. It is worth considering the possibility that our results are driven by large foreign investors. On one hand, an argument can be made that large shareholders, who often have a longterm commitment compared with minority shareholders, may be more influenced by a firm s level of transparency. On the other hand, an argument can be made that large shareholders are less concerned about the firm s transparency, because they can directly influence the firm s governance and may have direct access to company management, and therefore, rely less on publicly available sources of information. In additional (un-tabulated) analysis, we find the association between foreign ownership and transparency is positive and significant for both small and large investors, where large investors are defined as shareholders who own at least 5% of the firm s equity Investor protection, financial transparency, and foreign ownership As demonstrated in the previous section of the paper, on average, higher levels of reporting transparency are associated with higher levels of foreign ownership. This result implies that increasing reporting transparency is a strategy that all firms could pursue to attract foreign 21

22 equity capital. However, in this section we turn to the most crucial aspect of our study, and investigate if the relationship between firm-level transparency and foreign ownership is influenced by the prevailing level of investor protection in the firm s home country. We divide the sample at the sample median of investor protection, and classify firms with investor protection at or above the median as being in strong investor protection environments, and firms with investor protection below the median as being in weak investor protection environments. 12 We perform our analysis separately for the strong and weak investor protection subsets using our three alternative measures of investor protection. Table 6 shows the results of these analyses. Panel A of Table 6 presents the results using the Property Rights index from Gwartney and Lawson (2007) as our proxy for investor protection. For all three models we find a positive and significant association between transparency and foreign ownership for firms in strong investor protection environments. The coefficients are (t = 2.31), (t = 2.22), and (t = 2.89) when we use the first differences, firm fixed-effects, and the dynamic GMM specifications respectively. This translates to between a 27 and 50 basis point increase in foreign ownership for a 5-decile increase in Transparency. In contrast, for firms in countries with weak investor protection, we find no significant association between foreign ownership and transparency for any of the three models. Finally, we compare the coefficients on the transparency variable in high versus low investor protection countries, and find the association between transparency and foreign ownership in high investor protection countries is significantly stronger than in low investor protection countries. [Insert Table 6 about Here] 12 We determine the median value of each investor protection proxy based on the cross section of countries (i.e., based on 51 country-level observations). This ensures that we have a similar number of countries in the strong and weak investor protection subsamples. 22

23 Panels B and C of Table 6 show the results of our analyses using the anti-self-dealing index from Djankov, La Porta, Lopez-de-Silanes, and Shleifer (2008) and the earnings quality measure from Leuz, Nanda, and Wysocki (2003) as proxies for investor protection, respectively. Consistent with the results for the Property Rights index we find a positive and significant relationship between transparency and foreign ownership only for firms in strong investor protection environments. Further, the coefficients on Transparency are generally larger using both of these measures of investor protection, reaching a high of (t = 4.19) using the dynamic GMM model with the anti-self-dealing index. Finally, we show in both panels B and C of Table 6, that the association between foreign ownership and firm-level transparency is significantly stronger in high investor protection countries than in low investor protection countries. 13 The results in Table 6 indicate that in countries with higher levels of investor protection, increasing financial transparency is an effective mechanism that firms can use to attract foreign equity investors. On the other hand, firms in countries with weak investor protection are unable to attract foreign investors by unilaterally improving their disclosure quality; they are captive to the institutional environment of their home country. Thus, it appears that firm-level transparency and investor protection are complements; firms can benefit from increasing the quality of their financial reporting only if their countries already have a relatively high level of investor protection. 5. Robustness tests and additional analysis 13 One argument that can be made is that the ownership variable is essentially truncated, since ownership cannot fall below 0%, or increase above 100%. Another argument that can be made is that splitting the sample reduces the power of the tests to meaningfully detect any association between foreign ownership and transparency in the Low investor protection sample of countries. In un-tabulated analyses we performed our estimations using a Tobit model, and using a pooled regression (including both strong and weak investor protection countries) interacting transparency with an indicator variable for strong investor protection countries. The results using both of these alternative specifications support our inferences. 23

24 5.1. Alternative measures of financial transparency. We assess the robustness of our results to our choice of proxy for financial reporting transparency using two measures suggested by Daske, Hail, Leuz, and Verdi (2013) which are intended to measure the incentives of firms to provide high quality accounting information. 14 The first measure is based on the idea that managers in larger, more international, and more profitable firms, with larger financing needs, more growth opportunities, and a more dispersed ownership base, have stronger incentives to provide more transparent financial reporting. To create this first alternative measure, we perform a factor analysis on: the natural log of the firm s market value, total liabilities divided by total assets, return on assets, book-to-market ratio, percentage of closely held shares, and foreign sales over total sales. We retain the first factor from this analysis and label it Transparency_Supply_Incentives. The second alternative measure of transparency (which does not rely on accounting information) is based on the idea that analyst scrutiny increases the demand for more transparent financial reporting. To create this measure, we calculate the log of the number of analysts following the firm (plus one) and label this proxy as Transparency_Demand_Incentives. 15 We re-estimate equation (1), replacing Transparency with each of these alternative proxies. Table 7 summarizes the results of these tests. [Insert Table 7 about Here] Our results on the relationship between firm-level transparency and foreign ownership are robust to both of these alternative proxies. The coefficient is positive and significant on each measure of reporting incentives in countries with strong investor protection under all three of our measures of investor protection and under all three of our statistical specifications (i.e., first- 14 We chose to use Transparency in our primary tests because it is based explicitly on observed accounting measures and therefore, provides the most direct measure available to investors of a firm s financial reporting transparency. 15 The analyst following data available through OSIRIS is based on the I/B/E/S database. Similar to Daske, Hail, Leuz, and Verdi (2013), we set analyst following to zero for firms without coverage. 24