Scholarly article on topic 'Foreign institutional investment: Is governance quality at home important?'

Foreign institutional investment: Is governance quality at home important? Academic research paper on "Economics and business"

CC BY
0
0
Share paper
OECD Field of science
Keywords
{G11 / G20 / G38 / "Foreign institutional investment" / "Home country effect" / "Flight to quality" / "Familiarity argument"}

Abstract of research paper on Economics and business, author of scientific article — Nida Abdioglu, Arif Khurshed, Konstantinos Stathopoulos

Abstract This paper examines the investment preferences of foreign institutional investors investing in the U.S. market. We analyse both firm and country-level determinants that influence the foreign institutional investors' allocation choices. At the country level, we find that the governance quality in a foreign institutional investor's home country is a determinant of their decision to invest in the U.S. market. Our findings indicate that investors who come from countries with governance setups similar to that of the U.S. invest more in the United States. The investment levels though, are more pronounced for countries with governance setups just below that of the U.S. Our results are consistent with both the ‘flight to quality’ and ‘familiarity’ arguments, and help reconcile prior contradictory empirical evidence. At the firm level, we present unequivocal evidence in favour of the familiarity argument. Foreign institutional investors domiciled in countries with high governance quality prefer to invest in U.S. firms with high corporate governance quality. This effect is primarily driven by grey (non-monitoring) institutional investors.

Academic research paper on topic "Foreign institutional investment: Is governance quality at home important?"

Contents lists available at SciVerse ScienceDirect

Journal of International Money and Finance

journal homepage: www.elsevier.com/locate/jimf

INTERNATIONAL MONEY and FINANCE

Foreign institutional investment: Is governance quality at home important?

Nida Abdioglu3,*, Arif Khurshed b, Konstantinos Stathopoulosb

a Bandirma Faculty of Economics and Administrative Sciences, Balikesir University, Bandirma, Balikesir 10200, Turkey b Manchester Business School, University of Manchester, Booth Street East, Manchester M15 6PB, UK

ABSTRACT

JEL classification:

Keywords:

Foreign institutional investment Home country effect Flight to quality Familiarity argument

This paper examines the investment preferences of foreign institutional investors investing in the U.S. market. We analyse both firm and country-level determinants that influence the foreign institutional investors' allocation choices. At the country level, we find that the governance quality in a foreign institutional investor's home country is a determinant of their decision to invest in the U.S. market. Our findings indicate that investors who come from countries with governance setups similar to that of the U.S. invest more in the United States. The investment levels though, are more pronounced for countries with governance setups just below that of the U.S. Our results are consistent with both the 'flight to quality' and 'familiarity' arguments, and help reconcile prior contradictory empirical evidence. At the firm level, we present unequivocal evidence in favour of the familiarity argument. Foreign institutional investors domiciled in countries with high governance quality prefer to invest in U.S. firms with high corporate governance quality. This effect is primarily driven by grey (non-monitoring) institutional investors.

© 2012 Elsevier Ltd. All rights reserved.

1. Introduction

Recent studies highlight the importance of home country characteristics in determining the target country investment preferences of foreign investors. Kim et al. (2011) show that foreign investors

* Corresponding author. Tel.: +90 266 7380945; fax: +90 266 7380946. E-mail addresses: nidaabdioglu@yahoo.com (N. Abdioglu), Arif.khurshed@mbs.ac.uk (A. Khurshed), k.stathopoulos@mbs.ac. uk (K. Stathopoulos).

0261-5606/$ - see front matter © 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.jimonfin.2012.08.001

domiciled in countries with good corporate governance quality prefer good corporate governance firms, that is, they prefer familiar stocks. They also argue that foreign investors from low corporate governance quality home countries are indifferent to the level of corporate governance in the target country. In contrast, Forbes (2010) argues that a primary factor of foreign investment in the U.S. is the desire of foreign investors from less developed financial markets to invest in a developed market. That is, her conclusions support the flight to quality argument. Using a unique research design, which utilizes the distance of governance quality between home and target countries, this study reconciles prior evidence by illustrating that both the flight to quality and familiarity arguments can hold. Our country-level results on foreign investment in the U.S. capital markets suggest that the two effects, i.e., familiarity and flight to quality, are complimentary and not substitutes. We also investigate the firmlevel investment preferences of foreign investors in the U.S. and relate them to home country governance. Our firm-level evidence supports the familiarity effect; we show that this is mainly due to the information asymmetry costs of foreign institutional investors. Overall, this paper provides new evidence on the interplay between the flight to quality and familiarity effects and extends the existing literature on the importance of home country governance as a determinant of foreign investment allocations.

Investors allocate only a small fraction of their wealth to foreign markets, a practice typically referred to as the 'home bias puzzle' (Tesar and Werner, 1995). Kang and Stulz (1997) identify some explicit and implicit barriers that explain the home bias puzzle.1 Information asymmetry is one of the important barriers that foreign investors encounter when they invest abroad. Since the foreign investors have an informational disadvantage relative to local investors, the cost of investing abroad is high for them and it is this cost that prevents them from investing in foreign markets.2 However, recent trends document an increase in foreign investment as a result of globalisation. With globalisation, foreign capital has become an important source of finance in many capital markets and foreign investors have started to allocate more of their money abroad (Leuz et al., 2009 and Bekaert et al., 2002).

Over the last decade, equity ownership by foreign institutional investors in U.S. firms has almost doubled in size. In 1999, nearly 4% of the equity in S&P 1500 firms3 was in the hands of foreign institutional investors. This had increased to 8% by 2008. According to the 2008 U.S. Treasury report on institutional holdings, $2.969 trillion worth of capital is in the hands of foreign institutional investors. This substantial and growing foreign investment in the U.S. motivates us to investigate the determinants of foreign investment. The fact that most of it is channelled through institutional investors justifies the research focus of this paper.

Although foreign institutional investors have grown in size and importance in the U.S. market, most of the academic studies on institutional ownership focus on aspects of U.S. institutional investment, either within the U.S. or abroad with almost no attention being paid to foreign institutional investment in the United States. Further, studies that focus on the determinants of institutional ownership have always considered firm-specific variables, such as financial performance, liquidity, size, volatility and corporate governance setups (for example see Dahlquist and Robertsson, 2001; Almazan et al., 2005 and Li et al., 2006). It is only recently that the role of country-level governance quality has been the focus of attention.4

In this paper, we examine the investment preferences of foreign institutional investors investing in U.S. firms and investigate the role of country-level governance quality on their investment preferences. Specifically, we address two important research questions: First, does the governance quality of the home country5 play any role in the foreign institutional investor's decision to invest in the U.S? Second,

1 One explicit barrier is the constraint on foreign exchange transactions. Political risk and information asymmetry differences between foreign and domestic investors are examples of implicit barriers.

2 Several studies, such as Brennan and Cao (1997), Kang and Stulz (1997), report the information disadvantage of foreign investors relative to domestic investors. Dvorak (2003) also discusses the information asymmetry between foreign and domestic investors.

3 S&P Composite 1500 covers approximately 85% of the U.S. market capitalisation.

4 Notable examples are the studies by Kim et al. (2011) and Forbes (2010); we refer to these papers in detail below.

5 Home country is the foreign institutional investors' country of domicile.

does country-level governance quality affect the foreign institutional investor's preferences for specific U.S. firms?

So far, the existing literature finds that, in order to avoid high levels of investment costs, foreign institutional investors place importance on the country-level governance quality of the countries they invest in (Li et al., 2006 and Leuz et al., 2009). However, in this paper we argue that the governance quality of the foreign institutional investors' own countries affects their investment preferences as well. If they have low-quality governance at home, their costs when investing at home might be much higher than the costs they bear when investing abroad. This trades off the information asymmetry costs associated with foreign investment. In line with the 'good country bias' theory of Giannetti and Koskinen (2010), we argue that weak investor protection, low information disclosure and, therefore, high information asymmetry result from low-quality governance in a country. Thus, institutional investors who are domiciled in a country with weak investor protection are inclined to invest in foreign countries that provide higher-quality corporate governance than their home country. As a result, a good country bias is seen in the investment preferences of foreign institutional investors. Putting this in a different way, we expect to see a pronounced effect of flight to quality in the investment behaviour of foreign institutional investors who have lower governance quality at home than exists in the United States. Still, one should acknowledge the possibility of higher institutional investment levels from countries with similar (therefore, even just above that of the U.S.) levels of governance quality. Based on the familiarity argument (Chan et al., 2005), institutional investors prefer the familiar to the unfamiliar, since the former allows them to reduce the costs associated with investment uncertainty and 'prudent man rule' mandates. Therefore, we argue that both biases, flight to quality and familiarity, affect foreign institutional investment preferences in the U.S. market. Our argument allows for heterogeneity in the investment preferences of foreign institutional investors based on the relative distance in governance quality between the U.S. and the home country.

Furthermore, although prior studies (i.e., Ferreira and Matos, 2008; Kang and Stulz, 1997) have examined the relation between foreign institutional investment and firm-specific characteristics, they have not investigated whether this relation depends on the home country's governance quality. We argue that foreigners do not have homogeneous preferences when choosing the firms they invest in. Foreign institutional investors who have low information asymmetry at home (high governance quality) invest in low information asymmetry U.S. stocks (high corporate governance quality stocks). This prediction is In line with the familiarity argument. Foreign institutional investors look for familiarity in stock characteristics when they invest abroad, because familiarity will reduce the information costs and, as a result, the home bias will decrease. We proxy familiarity with the similarity between the level of information asymmetry a foreign investor is exposed to at home and the level they experience in the target firm.6

We find that foreign institutional investors who have low governance quality at home invest more in the U.S. market. To our knowledge, our study is among the first in the literature to focus on home country governance quality as a determinant of foreign institutional investment preferences. In addition, we examine differences in the choices of institutional investors, based on the different levels of governance quality they experience at home. We split our sampled countries into those with governance quality above and below that of the U.S. (the Above-U.S. and Below-U.S. groups) and thus investigate separately foreign institutional investment from countries with higher and lower governance quality than the United States. Foreign institutional investors from countries with similar (just above/below) governance quality to that of the U.S. invest more in the U.S. market. Thus, we also find support for the familiarity argument at the country level: the closer the foreign institutional investors are to the U.S., in terms of the governance quality they experience at home, the more they invest in the U.S. These findings suggest a complementary effect (instead of a substitution one) between the familiarity and flight to quality arguments. This is the first study that reports this complementary effect, thus extending recent papers in this area (Kim et al., 2011; Forbes, 2010).

6 The literature uses several proxies for familiarity. We present some of them in Section 3.

At the firm level, our results show that foreign institutional investors who come from countries with high governance quality invest in U.S. firms with good corporate governance systems. Thus, the home country's level of governance affects the investor's portfolio choices abroad. We argue that the familiarity bias holds even for firm-level preferences. Our results also indicate that this effect is driven by grey investors, who do not actively monitor the management teams of the firms they invest in, and therefore have the most to gain from a reduction in information asymmetry. The distinction of this effect between grey (non-monitoring) and independent (monitoring) investors is more pronounced for non-U.K. foreign investors.7

The rest of the paper is structured as follows. In Section 2, we provide an overview of the literature on the investment preferences of foreign institutional investors and highlight our contributions. Section 3 presents our hypotheses. Our data and methodology follow in Section 4. Section 5 discusses the empirical results. Section 6 presents our robustness tests. Section 7 provides our conclusions.

2. Literature review

Prior empirical work that examines the investment preferences of foreign institutional investors typically investigates the preferences of U.S. institutional investors when investing abroad. Leuz et al. (2009) analyse the foreign holdings of U.S. investors and document that the typical U.S. investor invests less in countries with weak legal institutions and poor information frameworks. Aggarwal et al. (2005) examine the investment choices of U.S. mutual funds in emerging markets. They find that strong accounting standards, shareholder rights and legal frameworks attract more U.S. investment. Ferreira and Matos (2008) also find that foreign institutional investors invest more in countries that have strong governance systems. Analysing the equity holdings of mutual funds from 26 developed countries, Chan et al. (2005) report that high levels of foreign institutional investment are expected in countries with low expropriation risk. Li et al. (2006) use the degree of enforcement of shareholder rights as a proxy for a country's governance quality. They find a positive association between the degree of enforcement and foreign institutional investment. Finally, Gelos and Wei (2005) find that international institutional investors invest in more transparent markets.8

All the above papers examine the effect of the governance quality of the countries in which the institutional investors are investing (target country) and establish a significant link with foreign investment levels (capital inflows). However, we argue that foreign institutional investment preferences are not homogeneous with regard to country-level governance quality. By examining the differences between the home country's governance quality and a benchmark, i.e., the level of governance in the U.S., we study the sources and impact of this heterogeneity. Our research design allows us to examine the impact of each country's governance quality (home country) on the level of foreign investment in the U.S. by institutional investors domiciled in that country (capital outflows). Two recent papers investigate foreign investment in the U.S. market. Cai and Warnock (2006) study the security-level investment preferences of foreign and domestic institutional investors in the United States. They identify a preference of institutional investors for domestic multinationals. They argue that this is a safe way of achieving international diversification. The scope and focus of our paper are different. In particular, we relate investment preferences to home governance quality and investigate the investment patterns of foreign institutional investors both at the country and firm levels. We find that the heterogeneity in countries' governance quality around the globe explains the investment choices in the U.S. market.

The paper closest to this study is that of Forbes (2010). She studies the level of foreign investment in the U.S. and concludes that foreign investors hold a greater amount of their wealth in the United States, if they have a less developed financial market at home. There are significant differences between our study and that of Forbes (2010). Forbes (2010) studies the holdings of all types of

7 Our study documents that the majority of foreign investment in S&P 1500 firms comes from U.K. institutional investors. Throughout our analysis we report the sensitivity of our results to the exclusion of holdings by foreign investors domiciled in the United Kingdom.

8 La Porta et al. (1997,1998, 2000) also examine the effect of country-level governance quality on investment decisions.

foreign investors, including government agencies and other official institutions. In contrast, we focus on private institutional investment; we expect market-based considerations, e.g., governance quality, to be the main driver of the decision making of this group of informed investors. Political, and other non-market based, influences are expected to carry less weight in the investment choices of this group. Also, Forbes (2010) investigates foreign holdings in both the equity and debt markets. There are substantial differences in the profiles of investments in equities and liabilities, e.g., risk characteristics and exposure, investment horizon, investor types. To avoid capturing systematic differences in the investor profiles, which might affect our results in unpredictable ways, we focus only on the equity markets. In addition, our research design helps us to explore the heterogeneity in the governance quality of countries with substantially different profiles to that of the United States. Our findings on the co-existence of the flight to quality and familiarity biases extend those provided by Forbes (2010) and help us to reconcile the conflicting evidence she offers with that of other studies (e.g., Aggarwal et al., 2005). Finally, but equally importantly, we also provide security-level analyses, which highlight the importance of governance quality at home for the firm-level investment preferences of foreign investors, which Forbes (2010) does not consider; we turn to this literature below.

The literature examines the effect of firm-level governance quality on foreign institutional investment, without taking into account the effect of the home country's governance quality. The definition of 'good governance' varies between studies.9 Kang and Stulz (1997), Ahearne et al. (2004), Edison and Warnock (2004) and Aggarwal et al. (2005) use the ADR (American Depository Receipt) issuance as a proxy for strong shareholder rights, or a reduction in asymmetric information, for foreign firms and find that firms with ADR issuance have high levels of U.S. ownership. Leuz et al. (2009) also conclude that U.S. institutional investors invest less in foreign stocks with high information asymmetry and high monitoring costs. Furthermore, Giannetti and Simonov (2006) show, for the Swedish market, that foreigners invest less in firms with high levels of outside investor expropriation.10

A recent study by Kim et al. (2011) finds that if an investor's country has a low level of disparity between ownership and control, the investor does not invest in high-disparity firms in the Korean market. Our study is different to that of Kim et al. (2011) in the following ways: First, Kim et al. (2011) use the disparity between control and ownership as their proxy for corporate governance risk, hence concentrating on the risk of tunnelling by controlling shareholders. We capture the country-level governance quality using the World Bank KKM (Kaufmann et al., 2007) indicators. These six key indicators provide a more holistic representation of a country's governance level.11 Second, Kim et al. examine the behaviour of all investors, retail and institutional. We concentrate on institutional investors since they are informed, highly-skilled traders, and therefore our conjectures on their investment preferences, based on the governance quality at home and abroad (the processing of which information requires expert knowledge and tools), appear more robust. Third, Kim et al. examine the investment preferences of foreign investors in an emerging market, i.e., Korea; it is unclear what the foreign investors' perceptions of the overall governance quality of Korea are (thus it is unclear whether this factor affects the heterogeneity of foreign investors investing in this country). However, we study foreign investment in the U.S., which undoubtedly has a strong corporate governance system, thus allowing the maximum possible degree of investor heterogeneity. Finally, our research design helps us to reconcile the results reported in Kim et al. (2011) with the conflicting evidence provided by prior studies.

9 McKinsey and Company (2002) report the following ranking in the investors' governance priorities: strong shareholder rights (for 33% of investors), good accounting standards (for 32% of investors), more effective disclosure (for 31% of investors), and stronger enforcement (for 27% of investors).

10 Other studies examine the effect of firm-level accounting variables on institutional investment. For example, Dahlquist and Robertsson (2001), Kang and Stulz (1997), Gompers and Metrick (2001), Falkenstein (1996), Ferreira and Matos (2008) and Khurshed et al. (2011).

11 Beltratti and Stulz (2009) and Caprio et al. (2011) use the KKM measures as proxies for country-level governance quality.

3. Hypotheses

Domestic investors prefer to invest more in their local market than abroad. This 'home bias puzzle' (Lewis, 1999) is mainly attributed to the higher information asymmetry costs associated with investing abroad. Given that domestic investors have more information about the domestic economy, foreign investors suffer from adverse selection problems and require a premium in order to compete with domestic investors. However, there is a considerable and growing amount of foreign institutional investment in the United States (see Table 1). This indicates that the U.S. market offers investment opportunities to foreign investors that can give them the required premium to compensate them for their adverse selection problems. Given the recent high levels of foreign investment in the U.S., we explore the factors that attract high level of foreign institutional investment to the U.S. market. We argue that the level of investor protection is one of the most important factors in drawing foreign investment to the United States.

Investor protection is defined as "... the extent of the laws that protect investors' rights and the strength of the legal institutions that facilitate law enforcement" (Defond and Hung, 2004; p. 269). The level of investor protection depends on the governance quality of a country. It is well-documented in the literature that investors prefer to invest in countries with high-quality governance, and thus a high level of investor protection, for example Aggarwal et al. (2011) and Leuz et al. (2009). Even if a firm has good internal corporate governance, if it is domiciled in a country with weak governance quality, it benefits less from the domestic capital markets. It cannot reduce its cost of capital to the level it would obtain if it was domiciled in a country with good governance (Doidge et al., 2007). Firms cannot compensate for the absence of a high-quality governance environment. They can improve the quality of their internal corporate governance quality but this cannot substitute for the quality of legal rules (Klapper and Love, 2004). In line with this argument, we predict that one of the most important determinants of institutional investment is the quality of the governance environment. We expect the level of governance at home to affect the level of institutional investment abroad, as investors try to identify better investment opportunities in a globalised environment.

Table 1

Institutional ownership.

Variable 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008

Panel A: Foreign vs. Domestic ownership

TI 52% 54% 56% 59% 61% 65% 66% 69% 72% 70%

FI 4% 4% 5% 5% 3% 6% 7% 7% 8% 8%

DI 48% 46% 51% 54% 58% 58% 59% 61% 63% 61%

UNC - 4% - - - 1% - 1% 1% 1%

Panel B: Ownership by investor type

BANKS 13% 13% 13% 15% 14% 16% 16% 17% 16% 15%

INS 5% 5% 6% 5% 4% 5% 4% 4% 6% 4%

IA 31% 32% 33% 34% 37% 37% 39% 42% 42% 45%

PPS 2% 3% 3% 3% 3% 4% 3% 3% 4% 2%

OTHERS 1% 1% 1% 2% 3% 3% 4% 3% 4% 4%

Total 52% 54% 56% 59% 61% 65% 66% 69% 72% 70%

This table presents descriptive statistics on institutional ownership. Panel A reports the fraction of shares held by foreign and domestic institutional investors investing in the U.S. between 1999 and 2008. The investment levels are reported for the constituent firms of the S&P 1500 index. Institutional ownership at the firm level is defined as the sum of the institutional investors' holdings at fiscal year-end divided by total shares outstanding. TI is the level of total institutional investment in the United States. FI is the investment level of foreign institutional investors investing in the United States. DI is the investment level of domestic (U.S.) institutional investors. UNC is the investment level of institutional investors for which there is no information on origin (cannot be classified). Panel B reports the level of investment in the U.S. by different types of institutional investors for the period 1999-2008. BANKS is ownership by banks. INS is insurance companies. IA is investment advisors, i.e., investment companies and independent investment advisors. PPS is public pension funds. OTHERS includes university and foundation endowments, corporate pension funds and miscellaneous investors.

The good country bias theory of Giannetti and Koskinen (2010) helps us to build our first hypothesis. We argue that institutional investors domiciled in countries with low levels of investor protection, are inclined to seek investment opportunities in foreign countries that provide better investor protection. This allows them to take advantage of the high-quality governance in better protected economies and identify more cost-efficient investments. The extent of this strategy depends on how bad the home governance is compared to the level of governance in the target country. In other words, we expect the flight to quality of foreign institutional investors to vary systematically with respect to the relative distance in governance quality between the home and target countries.

The investigation of the relative distance between countries also allows us to test an alternative argument based on the familiarity bias. This postulates that foreign investors prefer to invest more in familiar environments since this helps them reduce their information asymmetry costs. Therefore, one would expect foreign investors to invest more in countries with a similar level of governance quality to that of their home country. The literature so far presents the familiarity and flight to quality biases as substitutes. The evidence produced supports either but not both. In this paper, we hypothesise that both effects can co-exist; it is an empirical question whether the magnitude of either effect is more pronounced.

Hypothesis 1a. If the governance quality of an investor's country is lower than that of the U.S., the investor will invest in the United States.

Hypothesis 1b. If the governance quality of an investor's country is similar to that of the U.S., the investor will invest in the United States.

We use the United States as the "benchmark" in order to investigate the importance of the difference between countries' governance regimes. The U.S. economy provides a high standard of investor protection and information disclosure, and thus can operate as an ideal benchmark, since it should allow the greatest possible heterogeneity in foreign investor origins, i.e., several countries will be represented.

As we mentioned above, one way of reducing the information asymmetry costs associated with a foreign investment is by investing in familiar stocks abroad i.e., stocks that share characteristics with firms at home. Merton (1987) documents that investors prefer to invest in stocks they have information about. Chan et al. (2005) argue that foreign institutional investors aim at investing in familiar stocks, since they can better process the available information for these firms. Also, given that institutional investors are typically bound by "prudent man rules" (Del Guercio, 1996), they need to invest in reliable assets. According to these rules, institutional investors have incentives to protect themselves from liability by tilting their portfolios toward those assets that are easy to defend in court. Therefore, understanding the structure, organisation, business activities, etc., of a firm is vital to them. In line with the familiarity argument, foreign investors who experience high governance quality at home are expected to invest more in U.S. stocks with high-quality corporate governance.

Prior literature uses several definitions for familiarity. According to Kang and Stulz (1997), larger and more internationally-known stocks are familiar stocks to foreign investors. Coval and Moskowitz (1999) find that geographical proximity to be an important factor for foreign mutual funds in identifying familiar stocks. Grinblatt and Keloharju (2000) show common language and cultural background as a reason for foreign investors to invest in Finland. We expect a firm to be familiar to a foreign investor when the level of information asymmetry within the firm resembles the information asymmetry level in the investor's country of domicile.

We exploit the cross-sectional variation in the corporate governance quality of U.S. firms. We expect foreign institutional investors who have a high quality of country-level governance quality at home to invest in U.S. firms with high internal governance scores. Thus, we expect that the governance quality in investors' home countries affects their portfolio choices abroad.

Hypothesis 2. Foreign institutional investors who experience high governance quality in their home countries prefer to invest in U.S. firms with high corporate governance quality.

4. Data and methodology

4.1. Data

We use the Thomson-Reuters Institutional Holdings database to collect data on foreign institutional ownership from the 13F filings of U.S. firms for the period 1999-2008.12,13 We include all stocks listed in the S&P1500 index, i.e., constituent firms of the S&P 500, S&P MidCap 400 and S&P SmallCap 600 indices; we exclude ADRs and foreign stocks. We identify more than three million investor-level observations; foreign institutional investors originate from 18 countries.14 For every firm in our sample, we aggregate the holdings of investors from the same countries, creating 77,697 country-level observations.

For the accounting variables we use the Compustat North America database. Data for market turnover is collected from CRSP. Corporate governance information, i.e., the directors' index (DINDEX), is collected from the director files of the Investor Responsibility Research Centre (IRRC), available through the RiskMetrics database. Compustat, IRRC and CRSP provide the related firm-level data at each fiscal year-end. The 13F filings report institutional investor holdings on a quarterly basis. We use the institutional holdings reported for the last quarter of each fiscal year and merge these with the other accounting and market data. Missing data on accounting variables and firm-level governance indicators reduces our sample size to 59,491 observations.

We follow Beltratti and Stulz (2009) in using the average of the Kaufmann, Kraay and Mastruzzi (KKM) governance indicators as a proxy for a country's governance quality. The World Bank website provides these indicators for each country in our sample for all related years.15 We collect additional country-level turnover data from Datastream.

4.2. Variables

4.2.1. Country-level governance quality

Our main proxies for a country's governance quality are the KKM governance quality indicators as defined by Kaufmann et al. (2007). We take the average of the six KKM indicators to create a variable which captures the annual average governance quality of a country. These indicators cover several dimensions of a country's governance, related to the level of accountability and freedom of speech, the efficiency and stability of the political system, the quality and independence

12 The availability of information on the country of origin of institutional investors dictates our choice on start date for our sample period, i.e., 1999.

13 There are two databases that provide large sample, firm-level institutional investor ownership details for U.S. listed firms. These are the Thomson-Reuters 13F and Thomson One Banker databases. We follow the extensive literature on institutional investment and use the 13F database as our data source. Thomson-Reuters 13F collects all the information contained in Form 13F filed with the Securities and Exchange Commission (SEC). U.S. law requires all institutional investors with $100m or more in assets under management to file a 13F form with the SEC. Therefore, if a country is not represented in our sample, then this is because of the following scenarios: (a) there is no institutional investment in S&P 1500 firms originating from this country; (b) institutional investors originating from this country manage very small portfolios (less than $100m) so do not have to comply with 13F rules. In that case their holdings are expected to be minimal and limited to a very small number of firms; (c) these countries invest in U.S. equities through other special investment vehicles and not through institutional investors. We explain in the main text why we believe that the investigation of institutional investment is ideal in our context. As a robustness check, we have also collected institutional investor ownership details for a random sample of 300 firms listed in the S&P1500 index from Thomson One Banker. We find a similar number of countries represented in the sample as well as similar levels of foreign investment across the years.

14 The majority of foreign investment in the U.S. originates from the United Kingdom, i.e., more than 50% in each year. We provide annual information on country-level holdings in Appendix Table A.1.

15 http://info.worldbank.org/governance/wgi/index.asp.

of public services, the regulatory quality, the rule of law, and the level of corruption.16 All these indicators capture, to varying degrees, the level of information asymmetry and uncertainty under which investors operate in a country. For this reason, we attempt to isolate their joint effect by calculating their average.

In our robustness checks, we alternatively define the level of governance in a country by the turnover ratio of its main financial market. The large literature on the relationship between law and finance (e.g., La Porta et al., 1998) finds that better governed countries are associated with more developed financial markets. A proxy for the development and depth of a market is its liquidity, typically measured through the dollar volume of shares traded in the market over a given period, scaled by its market capitalisation, i.e., turnover ratio. To calculate the turnover ratio for all the countries in our sample, we collect the appropriate variables from Datastream.17 We also collect information on the holdings of U.S. institutional investors abroad and use this to classify our countries. High levels of U.S. investment are associated with more developed foreign financial markets. We retrieve the relevant data from the Bureau of Economic Analysis reports.

4.2.2. Firm-level variables

We follow the extant literature, in particular Ferreira and Matos (2008), Dahlquist and Robertsson (2001) and Bushee et al. (2010), to identify and define the relevant firm-level variables:

(i) Firm Size (SIZE): We take the logarithm of the fiscal year-end market value of equity as a proxy for firm size (Compustat item: MKVLT). Prior studies mainly report a positive relation between firm size and institutional ownership. For example, Dahlquist and Robertsson (2001) agree with Merton (1987) and Huberman (2001), who argue that investors prefer firms which are more well-known and with which they are familiar, and find a significant positive relationship between firm size and investor ownership. This positive relationship between size and institutional ownership is consistent with the results of Badrinath et al. (1989), Cready (1994), Falkenstein (1996), and Bennett et al. (2003) for U.S. firms. In contrast, Hussain (2000) finds that institutional investors in the U.K. prefer to invest in smaller and riskier firms. In addition, Bennett et al. (2003) find that institutional investors appear to have recently changed their preferences in favour of smaller and riskier securities since these stocks provide an opportunity for institutional investors to exploit their informational advantages.

(ii) Book-to-Market Ratio (BM): We use the logarithm of the ratio of the book value of common equity outstanding (Compustat item: CEQ) to the market value of equity, to calculate BM. BM is used to differentiate between value and growth stocks. High and low values represent value and growth firms, respectively. According to Lakonishok et al. (1994) and Dahlquist and Robertsson (2001), institutional investors prefer growth firms. Ferreira and Matos (2008), however, find that U.S. (foreign) institutional investors prefer value (growth) stocks.

(iii) Firm Market Turnover (TURN): We first calculate the monthly turnover of a stock, which is equal to its monthly volume (CRSP item: VOL) scaled by the firm's shares outstanding for that month (CRSP item: SHROUT). The annual figure is the average of the twelve monthly observations.

16 According to Kaufmann et al. (2007, p. 3-4): "1. Voice and Accountability (VA), measures the extent to which a country's citizens are able to participate in selecting their government, as well as freedom of expression, freedom of association, and a free media. 2. Political Stability and Absence of Violence (PV), measures perceptions of the likelihood that the government will be destabilized or overthrown by unconstitutional or violent means, including domestic violence and terrorism. 3. Government Effectiveness (GE), measures the quality of public services, the quality of the civil service and the degree of its independence from political pressures, the quality of policy formulation and implementation, and the credibility of the government's commitment to such policies. 4. Regulatory Quality (RQ), measures the ability of the government to formulate and implement sound policies and regulations that permit and promote private sector development. 5. Rule of Law (RL), measures the extent to which agents have confidence in and abide by the rules of society, and in particular the quality of contract enforcement, the police, and the courts, as well as the likelihood of crime and violence. 6. Control of Corruption (CC), measures the extent to which public power is exercised for private gain, including both petty and grand forms of corruption, as well as "capture" of the state by elites and private interests."

17 Turnover is defined as the average daily turnover by value in a year (Datastream item VA) divided by the average daily market value of the country in that year.

Turnover measures the market liquidity of a firm's shares (Dahlquist and Robertsson, 2001). According to Almazan et al. (2005), liquid stocks are characterised by greater information flows, i.e., less information asymmetry, which allows institutional investors to better identify and replace poor managers. Gompers and Metrick (2001) use size, price and turnover as determinants of liquidity and also report a positive relation between institutional holdings and market liquidity.

(iv) Dividend Yield (DY): Dividend yield is the dividend per share (Compustat item: DVPSX-F) divided by the fiscal year-end share price (Compustat item: PRCC-F). According to Dahlquist and Robertsson (2001), foreign institutional investors invest in low dividend paying stocks because of tax advantages.

(v) Return on Equity (ROE): Return on equity proxies for the profitability of a firm and is calculated as the ratio of net income (Compustat item: NI) to common equity (Compustat item: CEQ). Aggarwal et al. (2005) find that institutional holdings are positively related to the return on equity of the firms in which institutions invest.

(vi) Leverage (LEV): Leverage is the ratio of debt in current liabilities (Compustat item: DLC) plus long-term total debt (Compustat item: DLTT) to total assets (Compustat item: AT). Leverage determines the investment choices of institutional investors in the following ways. Jensen and Meckling (1976) argue that debt reduces a firm's agency costs through increased monitoring by the bondholders. If institutional ownership acts as a monitoring mechanism as well, one would expect to see a negative relationship between leverage and institutional ownership due to the substitution effect. Bathala et al. (1994) examine the role of institutional ownership on managerial ownership and debt policy. They indeed find a negative relation between debt levels and institutional ownership. In addition, Dahlquist and Robertsson (2001) argue that leverage is a measure of a firm's long-term financial distress and find that foreign investors avoid investing in firms with high leverage.

(vii) Cash (CASH): Cash is the ratio of cash and short-term investments (Compustat item: CHE) to total assets (Compustat item: AT). According to Dahlquist and Robertsson (2001), institutional investors prefer to invest in firms with more cash, since these firms are associated with greater financial strength.

(viii) Directors Index (DINDEX): We follow Bushee et al. (2010) to create this index. DINDEX includes five different dummy variables for: board size, percentage of independent directors, CEO-chairman duality, presence of board interlocks, and attendance of board meetings. If the CEO and chairman positions are combined, there are less checks and balances for the CEO, and therefore less monitoring of his actions. We create a dummy variable (CEO) which is equal to one if the positions are combined and zero otherwise. Interlocked directors are defined as directors who serve on each other's boards and their presence on a board is considered an indicator of weaker governance. Bushee et al. (2010, p. 11) explain the reasoning behind this idea as follows: "Interlocked directors are considered indicative of weaker governance because such directors have reciprocating relationships that create incentives to vote in ways that benefit their counterparts and, hence, themselves." We create a variable (DLOCK) which is equal to one if there are any interlocks on the board of directors and zero otherwise. If there is less attendance of board meetings, in theory the monitoring of the management team will be less successful. Therefore, a low attendance level is an indication of ineffective governance. We code a variable (DBAD) as one if any of the directors miss 75 percent or more of the board meetings and zero otherwise. The proxy for board size is the logarithm of the number of directors (LNDIR). If the board is large, it is assumed that there are communication, coordination and decision-making problems. Next, since independent directors' careers do not depend heavily on the management team, they are considered to be more effective monitors of a firm's managers. To measure ineffective governance, we calculate the percentage of directors that are dependent (PNID). DINDEX so far includes three dummy variables, CEO, DLOCK and DBAD, and indicators for whether the firm has high levels of LNDIR and PNID. We then split the distribution of LNDIR and PNID into high and low groups using k-means cluster analysis. We create dummies for these two variables which are equal to one if they are in the high group and zero otherwise. Therefore, we now have five dummy variables and the DINDEX variable (the sum of these five dummy variables) takes values between zero and five.

A value of zero (five) indicates a board with the most effective (weakest) governance structure. Bushee et al. (2010) find a negative relation between institutional ownership and DINDEX.18

4.3. Methodology

To test our hypotheses we use the following two models:

The dependent variable (FIO) is the same in both models. FIO is the percentage ownership of a firm's equity by foreign institutional investors domiciled in country c at time t. It is defined as the ratio of the shares held by the group of investors from that country, as reported at the last quarterly filing before the fiscal year-end, to the firm's shares outstanding at fiscal year-end. GQ is the governance quality at time t of the country (c) in which the investors are domiciled, measured as the average of the six KKM governance indicators. D is the directors' index (DINDEX) of firm f at time t, and GQd is a dummy variable which allows us to split our sampled countries in each year into the Above-/Below-U.S. groups in terms of governance quality (GQ). To confirm our hypotheses, we expect the coefficients a1 and b1 to be negative and statistically significant.

In both models, we use the same vector of firm-level control variables (X). These are firm size, book-to-market ratio, dividend yield, turnover ratio, return on equity, leverage and cash level.

We use random-effect Tobit panel regressions to run our analyses. We opt for this model specification since our dependent variable is censored between zero and one. Tobit regressions allow us to control for the substantial observed clustering of the dependent variable at values close to zero (left clustering at 0.001% is pronounced). All specifications account for random unobserved firm effects. We also add country fixed effects into our models, to account for omitted time invariant country effects.19 In addition, we include year dummies to control for cross-sectional dependence, i.e., market-wide effects that could influence the level of foreign institutional investment in the U.S. market.

In our robustness section, we also present alternative model specifications that control for the influence of specific market-wide influences, such as the enactment of the Sarbanes-Oxley Act (SOX) in 2002, as well as firm fixed effects.

5. Empirical results

5.1. Descriptive analyses

Table 1, Panel A, presents the fraction of shares held by foreign (FI) and domestic (DI) institutional investors as well as overall institutional investment (TI) in the U.S. between the years 1999 and 2008. We define institutional investment (II) as the fraction of a firm's stocks that are owned by institutional investors (Gompers and Metrick, 2001). Panel A illustrates that TI increases almost mono-tonically over our sample period, reaching 70% in 2008. DI follows a similar pattern. Apart from a decline in 2003, foreign institutional ownership levels increase during the sample period. By the end of 2008, FI levels were twice of those that existed in 1999 (8% compared to 4%). This large increase in FI confirms that the level of foreign institutional holdings in the U.S. has grown significantly in recent years.

Table 1, Panel B, gives a breakdown of the different types of institutional investors who invest in the U.S. The Thomson-Reuters database identifies five types of institutional investors: Banks, Insurance Companies, Investment Companies and their managers, Independent Investment Advisors and Others.

18 Appendix Table A.3 provides descriptive statistics of the individual indicators, as well as the time-series changes in DINDEX.

19 Leuz et al. (2009) present several additional macroeconomic factors that might affect the level of foreign institutional investment, for example the degree of market integration, transaction costs, language and restrictions on capital flows. Even though some of these factors change over time, arguably these changes are going to be small in our short time-series.

FIOct = a0 + ai GQct + azXft + c

FIOc,t = b0 + bi GQdc,t * Df^t + bzXf;t + £

Others includes Public Pension Funds, University Endowments and Foundations. However, ThomsonReuters has serious classification errors in the S34 file (widely reported in the literature) with many banks and independent investment advisors classified as Others from 1998 onwards. Because of this classification problem, we categorise institutional investors according to the Bushee et al. (2010) classification.20 Following Bushee et al., we are able to identify Public Pension Funds and therefore separate them from the Others group. We also classify investment companies and independent investment advisors into one group, which we call investment advisors (IA). We end up with five different types of institutional investors: Banks, Insurance Companies, Investment Advisors, Public Pension Funds and Others. According to Panel B, out of the five types of institutional investors, Investment Advisors invest the most in U.S. firms. Their investment levels increase over the sample period to reach 45% in 2008. Banks are the next largest investors after Investment Advisors, in terms of holdings, reaching 15% in 2008. Insurance Companies and Public Pension Funds do not show substantial increases in their holdings during the sample period. The holdings of Others quadruple over our sample period.

Table 2, Panel A, presents summary statistics of the governance variables. The governance indicators take values between 2.5 and -2.5. A value of 2.5 (-2.5), indicates the highest (lowest) level of governance quality. Most governance indicators have an average of one or above, indicating a high quality of governance at the global level. Nevertheless, the standard deviation shows significant heterogeneity, both cross-sectionally and along the time-series.21 Table 2, Panel B, reports the descriptive statistics for the firm-specific characteristics. We winsorise leverage, dividend yield and return on equity at 1% (two tail) because these variables are highly skewed. The average firm in our sample has a market value of $5.7 billion and a return on equity of 9%. The average log book-to-market ratio is -86%, leverage is at 23%, and turnover is almost 19%. Cash holdings account for 14% of the value of total assets and the average directors' index is 1.7, indicating an above average (i.e., 2.5) level of internal governance quality. Ferreira and Matos (2008) report similar values for dividend yield, leverage and cash ratio but a higher (lower) average return on equity (book-to-market ratio). Bushee et al. (2010), report a similar average turnover ratio (15%).

Table 3 reports the Pearson correlation coefficients for all the variables used in this study. At a univariate level, FIO is negatively correlated with size which is consistent with prior literature (Ferreira and Matos, 2008). Moreover, we find a negative relation between foreign institutional investment and DINDEX, indicating that institutional investors prefer to invest in firms with high-quality corporate governance. Country governance (GQ) positively affects FIO in the U.S. market. This is consistent with the familiarity argument. Still, caution should be exercised in interpreting this result. We present detailed multivariate results in the next paragraphs which illustrate that the univariate result is only part of the story. Finally, FIO is also positively and significantly correlated with turnover and book-to-market ratio. In contrast, FIO is negatively correlated with return on equity and monitoring cost.

The results presented in Table 3 show that multi-collinearity is not an issue in our subsequent multivariate analyses. The only exception is the 50% correlation reported between the turnover ratio and monitoring cost (MC), which is expected given that monitoring cost is defined as 1/ Turnover (Almazan et al., 2005). So we exclude turnover from all model specifications that use MC to alleviate concerns about multi-collinearity. Our results are not sensitive to this decision (unre-ported results).

5.2. Foreign institutional ownership and country governance

We start by examining whether the governance quality of foreign institutional investors' home countries drives foreign institutional investment in the United States. Table 4, Panel A, presents the results of random-effect Tobit regressions on our full sample where the dependent variable is FIO. Consistent with our predictions, we find a negative, significant relation between GQ and foreign institutional

20 The data for this classification are available from http://acct3.wharton.upenn.edu/faculty/bushee/.

21 We provide the cross-sectional and time-series changes in GQ in Appendix Table A.2.

Table 2

Summary statistics.

Variable N Mean SD p25 p50 p75

Panel A: Descriptive statistics-Country governance variables

VA 128 1.21 0.45 1.02 1.36 1.52

PV 128 0.98 0.26 0.81 1.02 1.18

GE 128 1.65 0.37 1.34 1.74 1.94

RQ 128 1.49 0.31 1.25 1.53 1.73

RL 128 1.52 0.28 1.32 1.61 1.71

CC 128 1.71 0.41 1.35 1.77 2.07

GQ 128 1.43 0.25 1.19 1.47 1.62

Panel B: Descriptive statistics-Firm-specific characteristics

SIZE (billions) 17,011 5.736 14.325 0.500 1.255 3.928

DY 17,377 0.013 0.020 0.000 0.003 0.020

BM (log) 16,656 -0.864 0.771 -1.275 -0.788 -0.392

TURN 17,322 0.189 0.168 0.082 0.140 0.239

LEV 17,328 0.227 0.190 0.054 0.208 0.350

ROE 17,380 0.095 0.327 0.051 0.116 0.178

CASH 17,382 0.144 0.176 0.022 0.067 0.203

DINDEX 12,769 1.702 0.977 1.000 2.000 2.000

MC 17,322 9.752 8.658 4.178 7.168 12.172

Panel A presents the descriptive statistics for the KKM governance indicators used in the study. VA is Voice and Accountability; PV is Political Stability and Absence of Violence; GE is Government Effectiveness; RQ is Regulatory quality; RL is Rule of Law; CC is Control of Corruption. GQ is the average of the six KKM governance indicators. The six governance indicators are scaled from -2.5 to 2.5, with higher values corresponding to better governance outcomes. Panel B reports the descriptive statistics for the firm-level control variables used in this study. SIZE is firm size, defined as the fiscal year-end market value of equity. BM is the logarithm of the ratio of the firm's book value of equity to its market value of equity. DY is dividend yield, defined as the dividend per share, divided by the share price. LEV is leverage, which is equal to the ratio of total debt to total assets. ROE is the return on equity, defined as net income divided by common equity. CASH is cash holdings, defined as the ratio of cash and short-term investment to total assets. TURN is the firm's market turnover, which is equal to common shares traded divided by common shares outstanding. The turnover ratio is first calculated monthly. Annual turnover is defined as the average of the twelve monthly observations. DINDEX is the directors' index, defined as in Bushee et al. (2010). It is based on information regarding CEO duality, number of board directors, the percentage of directors who are not independent, interlocks on the board of directors and attendance of board meetings. A value of five (one) indicates a poorly (well)-governed firm. MC is monitoring cost, which is equal to 1/Turnover. Mean, median (p50), standard deviation (SD), 25th percentile (p25) and 75th percentile (p75) are reported. N is the number of observations. Leverage, dividend yield, firm size and return on equity are winsorised at 1% (two tails).

ownership. Institutional investors from countries where the governance is of a lower quality are associated with higher investment levels in U.S. firms. This result is In line with the flight to quality argument.

In the second model, we insert a dummy variable that splits the institutional investors' countries into the Above-/Below-U.S. groups (GQd). The coefficient of GQd is negative and highly significant, further confirming the flight to quality argument.

To test our second hypothesis, we interact GQd with each firm's DINDEX (GQd*D). We find a negative, significant relation between GQd*D and FIO. This shows, consistent with the familiarity argument, that foreign institutional investors who come from better (than the U.S.) -governed countries prefer to invest in U.S. firms with high levels of internal governance. The coefficient remains negative and significant even when we use an alternative proxy (MC) for corporate governance in our last model, i.e., foreign institutional investors from well-governed (better than the U.S.) countries invest in U.S. firms with low monitoring costs.

The above results, even though significant and robust, do not allow us to observe any heterogeneous preferences within the Above-/Below-U.S. groups.

In Table 5, we address this limitation by running the initial test (the first model specification of Table 4) separately for the investors from countries from each group. We find a negative (positive) and significant relationship between FIO and GQ in the above (below) group. Thus, among the foreign investors who experience a higher quality of governance at home than exists in the U.S., it is those with a lower GQ that invest more in U.S. firms. Meanwhile, for the below-U.S. group, foreign investors with a higher GQ invest more in U.S. firms. In other words, investors from countries that are closest to the U.S. (just above and below the U.S.), in terms of governance quality, invest the most in the U.S. market. This is clear evidence of familiarity.

Table 3

Correlation matrix.

FIO GQ DINDEX SIZE DY BM TURN LEV ROE CASH MC

GQ 0.105* 1

DINDEX -0.067* -0.021* 1

SIZE -0.130* -0.063* 0.247* 1

DY -0.001 0.032* 0.114* 0.089* 1

BM 0.048* 0.047* -0.016* -0.357* 0.238* 1

TURN 0.017* -0.019* -0.227* -0.078* -0.143* -0.011* 1

LEV -0.006 -0.003 0.084* 0.017* 0.289* 0.069* -0.080* 1

ROE -0.019* -0.020* 0.034* 0.198* -0.016* -0.404* -0.092* -0.025* 1

CASH 0.007 0.002 -0.184* -0.069* -0.293* -0.266* 0.327* -0.366* -0.021* 1

MC -0.031* 0.038* 0.260* -0.060* 0.149* 0.053* -0.504* 0.049* 0.033* -0.201* 1

This table presents the Pearson correlation coefficients of the variables used in this study. FIO is the firm-level foreign institutional ownership in the United States. It aggregates the firm-level equity holdings of all investors domiciled in each country (i.e., country-level holdings in each U.S. firm in our sample). GQ is the level of a country's governance quality calculated as the average of the six KKM governance indicators. DINDEX is the directors' index defined as in Bushee et al. (2010). It is based on information regarding CEO duality, the number of board directors, the percentage of directors who are not independent, interlocks on the board of directors and attendance of board meetings. A value of five (one) indicates a poorly (well)-governed firm. SIZE is the logarithm of firm size, which is defined as the fiscal year-end market value of equity. BM is the logarithm of the ratio of the firm's book value of equity to its market value of equity. DY is dividend yield, defined as the dividend per share divided by the share price. LEV is leverage, which is equal to the ratio of total debt to total assets. ROE is return on equity, defined as net income divided by common equity. CASH is cash holdings, defined as the ratio of cash and short-term investment to total assets. TURN is a firm's market turnover, which is equal to common shares traded divided by common shares outstanding. The turnover ratio is first calculated monthly. Annual turnover is defined as the average of the twelve monthly observations. MC is monitoring cost, which is equal to 1/Turnover. Leverage, dividend yield and return on equity are winsorised at 1% (two tails).

Fig. 1 illustrates the country-level findings from Tables 4 and 5. Investors from countries with higher (lower) governance quality than the U.S. belong to the Above-(Below-) U.S. group. Results in Table 4 provide evidence of flight to quality by foreign institutional investors. Investment levels in the average U.S. firm are higher from the Below-U.S. countries than from the Above-U.S. countries. But this result is mainly driven by those countries whose governance levels are just below that of the U.S. In other words, institutional investors from countries with significantly different governance quality to that of the U.S. do not invest in U.S. firms. We can only speculate on the reasons behind this. It seems rational that these institutional investors face only moderate pressure from their end-users to invest in better quality countries. These investors might also be faced with substantial barriers when trying to invest in the U.S. (capital controls, informational barriers, etc). Finally, this result might also be a manifestation that smaller countries, that are less likely to have institutional investors big enough to achieve internationally diversified portfolios, do not invest in their governance quality.22 In contrast, for the Above-U.S. group, it is clear that investors from countries with a significantly better level of governance quality than that of the U.S. have no incentive to invest in the U.S. For them, there are no benefits in investing in a country with a significantly lower quality of governance that would justify the additional costs associated with an investment abroad.

Appendices A.1 and A.2 show that amongst foreign institutional investors, the U.K. institutional investors are the largest investors in U.S. firms and that the U.K. has higher level of governance quality than the U.S. during our sample period. In order to test whether our results in Tables 4 and 5 are driven only by the U.K. institutional investors, we exclude institutional investors domiciled in the U.K. from our dataset and run our regressions again (Table 4, Panel B, and Table 5, model 2). Given that the U.K. has higher governance quality than the U.S. for every year in our sample, it is meaningful to re-run, in Table 5, only the Above-U.S. regression without U.K. domiciled institutional investors. The results

22 According to La Porta et al. (2006) there is a positive relationship between the size of the country and its governance quality. Thus, if a country is small in size, its governance quality should be low. These low governance quality countries have small institutional investors that cannot achieve international diversification. Because of this, we do not expect an investment from these countries to the U.S.

Table 4

Determinants of foreign institutional ownership. FIO (%)

Panel A: Full sample Panel B: Excluding U.K. investors

GQ -5.480*** [0.000] -0.097*** [0.000]

GQd*D -0.180*** [0.000] -0.003** [0.026]

GQd*MC -0.042*** [0.000] -0.001** [0.022]

GQd -0.119*** 0.189*** 0.167*** -0.021*** 0.063*** 0.003

[0.004] [0.000] [0.000] [0.000] [0.000] [0.478]

DINDEX -0.032*** -0.033*** 0.091*** -0.029*** -0.002** -0.002** 0.0003 -0.001*

[0.000] [0.000] [0.000] [0.000] [0.020] [0.022] [0.799] [0.071]

MC 0.015*** [0.000] -0.004*** [0.000]

SIZE (log) 0.025*** 0.014*** 0.015*** 0.012** 0.020*** 0.020*** 0.017*** 0.021***

[0.000] [0.007] [0.003] [0.025] [0.000] [0.000] [0.000] [0.000]

DY 0.994*** 0.849** 0.893** 1.213*** -0.182*** -0.182*** -0.213*** -0.105***

[0.004] [0.016] [0.011] [0.001] [0.000] [0.000] [0.000] [0.007]

BM (log) -0.006 -0.014 -0.013 -0.008 0.002* 0.002* 0.001 0.003**

[0.591] [0.203] [0.236] [0.473] [0.068] [0.061] [0.425] [0.026]

TURN 0.417*** 0.400*** 0.405*** 0.112*** 0.110*** 0.098***

[0.000] [0.000] [0.000] [0.000] [0.000] [0.000]

LEV 0.077* 0.058 0.060 0.055 0.017*** 0.017*** 0.013*** 0.017***

[0.065] [0.173] [0.160] [0.197] [0.000] [0.000] [0.006] [0.000]

ROE 0.090*** 0.089*** 0.092*** 0.086*** 0.005* 0.006** 0.006** 0.006**

[0.000] [0.000] [0.000] [0.000] [0.053] [0.036] [0.027] [0.029]

CASH 0.021 0.009 0.005 0.038 0.010* 0.010* 0.009 0.009*

[0.668] [0.857] [0.917] [0.424] [0.059] [0.056] [0.115] [0.088]

Constant 7.722*** -1.314*** -1.544*** -1.271*** 0.027*** -0.079*** -0.059*** -0.085***

[0.000] [0.000] [0.000] [0.000] [0.004] [0.000] [0.000] [0.000]

Year FE Yes Yes Yes Yes Yes Yes Yes Yes

Country FE Yes Yes Yes Yes Yes Yes Yes Yes

N 59,491 59,491 59,491 59,491 44,694 44,694 44,694 44,694

Chi2 82,698 76,531 76,942 77,829 10,846 10,898 6756 10,946

p-value 0 0 0 0 0 0 0 0

This table presents the results of random-effect Tobit panel regressions on the determinants of foreign institutional ownership. The dependent variable in all the regressions is the firm-level, foreign institutional ownership in the United States (FIO). It measures the firm-level equity holdings of all investors domiciled in a given country (i.e., country-level holdings in each U.S. firm in our sample). In Panel A (Panel B) we run the regressions on the full sample (excluding U.K. investors). GQ is the level of a country's governance quality, calculated as the average of the six KKM governance indicators. GQd is a dummy variable which equals one if the governance quality of a country (GQ) in year t is greater than that of the U.S. in that year and zero otherwise. DINDEX is the directors' index, defined as in Bushee et al. (2010). It is based on information regarding CEO duality, number of board directors, the percentage of directors who are not independent, interlocks on the board of directors and attendance of board meetings. A value of five (one) indicates a poorly(well)-governed firm. GQd*D is equal to the country-level governance quality dummy (GQd) multiplied by DINDEX. GQd*MC is equal to the country-level governance quality dummy (GQd) multiplied by MC. MC is monitoring cost, which is equal to 1/Turnover. SIZE is the logarithm of firm size, which is defined as the fiscal year-end market value of equity. BM is the logarithm of the ratio of the firm's book value of equity to its market value of equity. DY is dividend yield, defined as the dividend per share divided by the share price. LEV is leverage, which is equal to the ratio of total debt to total assets. ROE is return on equity defined as net income divided by common equity. CASH is cash holdings, defined as the ratio of cash and short-term investment to total assets. TURN is a firm's market turnover, which is equal to common shares traded divided by common shares outstanding. The turnover ratio is first calculated monthly. Annual turnover is defined as the average of the twelve monthly observations. Leverage, dividend yield and return on equity are winsorised at 1% (two tails). Year dummies (Year FE) and country dummies (Country FE) are included in all regressions. The numbers in brackets are p-values. * indicates 10% significance level, ** indicates 5% significance level and *** indicates 1% significance level. N is the number of observations.

remain qualitatively similar. As expected, given the large holdings of U.K. investors, the coefficients of interest in our analyses become smaller. For example, the large GQ coefficient in Table 4 Panel A (-5.480) drops to -0.097 in Panel B. Similarly, the GQ coefficient in the Above-U.S. regression of Table 5 (-11.793) drops to -1.554 when we exclude U.K. domiciled institutional investors. This highlights that

Table 5

Investment preferences for ABOVE-/BELOW-U.S. governance quality at home.

FIO (%)

ABOVE-U.S. ABOVE-U.S. (excluding U.K.) BELOW-U.S.

Model 1 Model 2 Model 3

GQ -11.793*** [0.000] -1.554*** [0.000] 0.353*** [0.000]

DINDEX -0.044*** [0.000] -0.013*** [0.010] -0.008** [0.021]

SIZE (log) 0.043*** [0.000] 0.021*** [0.000] -0.014*** [0.000]

DY 1.742*** [0.000] 0.215 [0.381] -0.15 [0.416]

BM (log) 0.008 [0.602] -0.002 [0.789] -0.007 [0.167]

TURN 0.530*** [0.000] 0.184*** [0.000] 0.107*** [0.000]

LEV 0.096* [0.092] 0.018 [0.563] 0.063*** [0.003]

ROE 0.120*** [0.000] -0.028* [0.087] 0.013 [0.273]

CASH 0.045 [0.488] 0.048 [0.180] -0.01 [0.678]

Constant 18.345*** [0.000] 2.692*** [0.000] -0.580*** [0.000]

Year FE Yes Yes Yes

Country FE Yes Yes Yes

N 41,054 28,908 18,437

Chi2 58,122 2080 1034

p-value 0 0 0

This table presents the results of random-effect Tobit panel regressions on the investment preferences of foreign institutional investors who come from countries with higher/lower governance quality than the United States. The dependent variable is still FIO, i.e., the firm-level foreign institutional investment in the U.S. by each country represented in our sample, but now, for the ABOVE-(BELOW-) U.S. models we only include foreign institutional ownership by institutional investors who come from countries with higher (lower) governance quality than that of the U.S. Given that the foreign investment coming from the U.K. accounts for the majority of foreign investment in the U.S., as well as the fact that the U.K. governance level is above that of the U.S. for all years in our sample, we also run the Above-U.S. regression excluding the investment originating by institutional investors domiciled in the U.K. This allows us to test the sensitivity of our results to this one country. GQ is the level of a country's governance quality calculated as the average of the six KKM governance indicators. DINDEX is the directors' index defined as in Bushee et al. (2010). It is based on information regarding CEO duality, number of board directors, the percentage of directors who are not independent, interlocks on the board of directors and attendance of board meetings. A value of five (one) indicates a poorly (well)-governed firm. SIZE is the logarithm of firm size, which is defined as the fiscal year-end market value of equity. BM is the logarithm of the ratio of the firm's book value of equity to its market value of equity. DY is the dividend yield, defined as the dividend per share divided by the share price. LEV is leverage, which is equal to the ratio of total debt to total assets. ROE is return on equity, defined as net income divided by common equity. CASH is cash holdings, defined as the ratio of cash and short-term investment to total assets. TURN is a firm's market turnover, which is equal to common shares traded divided by common shares outstanding. The turnover ratio is first calculated monthly. Annual turnover is defined as the average of the twelve monthly observations. Leverage, dividend yield and return on equity are winsorised at 1% (two tails). Year dummies (Year FE) and country dummies (Country FE) are included in all regressions. The numbers in brackets are p-values. * indicates 10% significance level, ** indicates 5% significance level and *** indicates 1% significance level. N is the number of observations.

a large part of the economic significance of our results is attributable to the U.K. investors. The fact that all the coefficients remain statistically significant and with the predicted sign is an indication that our results do not merely reflect the investment preferences of U.K. institutional investors.

6. Robustness tests

We run a series of tests to examine the robustness of our results. In Table 6 we show the results of allowing heterogeneity in the monitoring activity of the foreign institutional investors. Foreign investment levels in the U.S. could be driven by investors who cannot reduce their portfolios' information asymmetry at home since they do not monitor the investee firms' management (non-monitoring investors). These investors could try to reduce their portfolios' information asymmetry costs by investing in a country with lower information asymmetry levels. If there are systematic differences in the country of origin of these investors we could observe different results between monitoring and non-monitoring institutional investors. Following Bushee et al. (2010), we classify banks, insurance companies, corporate pension funds, university foundation endowments and others as grey (nonmonitoring) investors. In contrast, investment companies, independent investment advisors and

Fig. 1. Illustration of the main results reported in Tables 4 and 5.

public pension funds are classified as independent (monitoring) investors. Table 6, Panel A, shows that there are no differences in the investment preferences of grey and independent foreign institutional investors, with regard to the governance of the country in which they are investing. The flight to quality argument is valid for both types of investors. Interestingly, there are differences with respect to firmlevel investment. The familiarity argument holds only for the grey foreign institutional investors. This is not a surprising result: given that grey investors do not engage in active monitoring of a firm's management team, they are likely to prefer to invest in more familiar firms so as to reduce information asymmetry costs. As we mention in the previous Section, U.K. institutional investors are the largest foreign investors in U.S. firms. So, it is important we test the robustness of this result after excluding U.K. investors. In Table 6, Panel B, we report this analysis. The result for the independent institutional investors is sensitive to the inclusion of U.K. institutional investors; once we exclude these investors it becomes insignificant. Still, the coefficients for grey institutional investors remain significant throughout. Thus, consistent with prior studies, our results appear to support the argument that information asymmetry problems are strongly related to the decision to invest abroad.

In Table 7, we use alternative proxies for a country's governance quality. Following the law and finance literature (La Porta et al., 1998) we assume a positive relationship between a country's governance quality and the development of its main financial market. We use market liquidity (market turnover ratio) as a proxy for the development and depth of the financial market. We introduce the variable Td*D in the first column. Td*D is an interaction term that includes a country-level turnover dummy (Td) and DINDEX. Td is equal to one if the turnover ratio of a country's market is greater than the turnover ratio of the U.S. market in a given year and zero otherwise. We test whether foreign institutional investors who have developed markets at home (high governance quality or less information asymmetry) prefer to invest in high governance quality stocks (stocks with less information asymmetry) in the U.S. Table 7 reports a negative relationship between Td*D and foreign institutional investment. Thus, foreign institutional investors who experience less information asymmetry (high governance quality or a developed market) in their home countries prefer to invest in stocks involving less information asymmetry (those with high governance quality) in the U.S. This result is consistent with the familiarity argument. The result becomes insignificant if, instead of DINDEX, we use monitoring cost (MC) as the proxy for a firm's information asymmetry. However caution should be exercised when interpreting the turnover results. Even though, as we mention above, the law and finance

Table 6

Investment preferences of grey and independent foreign institutional investors.

FIO (%)

Panel A: Full sample

Panel B: Excluding U.K. investors

Grey Ind. Grey Ind. Grey Ind. Grey Ind. Grey Ind. Grey Ind.

GQ -5.821*** [0.000] -0.225*** [0.000] -0.079*** [0.000] 0.001 [0.815]

GQd*D -0.193*** [0.000] -0.006 [0.461] -0.003* [0.054] 0.001 [0.328]

GQd*MC -0.054*** [0.000] -0.001 [0.556] -0.002*** [0.000] -0.001 [0.105]

GQd 0.447*** 0.027 0.435*** 0.021 0.026*** -0.015*** 0.030*** -0.008*

[0.000] [0.308] [0.000] [0.400] [0.000] [0.001] [0.000] [0.067]

MC 0.031*** [0.000] -0.005*** [0.000] 0.001*** [0.004] -0.004*** [0.000]

DINDEX -0.034*** -0.016*** 0.123*** -0.013** -0.029*** -0.016*** -0.001* -0.001** 0.001 -0.002** -0.001 -0.001

[0.000] [0.000] [0.000] [0.037] [0.001] [0.004] [0.088] [0.044] [0.473] [0.024] [0.162] [0.114]

SIZE -0.024*** 0.003 -0.018*** 0.003 -0.022*** 0.001 0.004*** 0.012*** 0.002*** 0.013*** 0.003*** 0.013***

[0.000] [0.303] [0.003] [0.402] [0.000] [0.760] [0.000] [0.000] [0.000] [0.000] [0.000] [0.000]

DY 2.188*** -0.584*** 2.057*** -0.583*** 2.497*** -0.517** -0.136*** -0.111*** -0.141*** -0.085** -0.135*** -0.02

[0.000] [0.007] [0.000] [0.007] [0.000] [0.017] [0.000] [0.003] [0.000] [0.026] [0.000] [0.598]

BM 0.008 -0.012* -0.009 -0.012* -0.003 -0.012* 0.001 0.001 0.001 0.0004 0.001 0.001

[0.547] [0.069] [0.494] [0.066] [0.823] [0.070] [0.578] [0.630] [0.466] [0.691] [0.472] [0.300]

TURN 0.190*** 0.251*** 0.224*** 0.250*** 0.012 0.116*** -0.002 0.115***

[0.000] [0.000] [0.000] [0.000] [0.103] [0.000] [0.790] [0.000]

LEV 0.011 0.081*** -0.003 0.080*** -0.027 0.083*** -0.004 0.018*** -0.004 0.018*** -0.005 0.019***

[0.822] [0.001] [0.949] [0.002] [0.594] [0.001] [0.300] [0.000] [0.354] [0.000] [0.288] [0.000]

ROE 0.132*** 0.01 0.118*** 0.011 0.125*** 0.002 0 0.007*** 0.001 0.007*** 0.001 0.007***

[0.000] [0.466] [0.000] [0.446] [0.000] [0.871] [0.869] [0.007] [0.685] [0.005] [0.666] [0.003]

CASH 0.063 -0.027 0.048 -0.028 0.028 0.013 0.002 0.001 0.003 0.002 0.001 0.001

[0.268] [0.348] [0.408] [0.339] [0.623] [0.654] [0.766] [0.814] [0.606] [0.738] [0.785] [0.812]

Constant 5.402*** 0.340*** -1.683*** -0.055 -1.377** 0.071 0.132*** -0.040*** 0.159*** -0.035*** 0.156*** -0.085***

[0.000] [0.005] [0.006] [0.399] [0.025] [0.273] [0.000] [0.000] [0.000] [0.000] [0.000] [0.000]

Year FE Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes

(continued on next page)

Table 6 (continued)

FIO (%)

Panel A: Full sample

Panel B: Excluding U.K. investors

Country FE N

Chi2 p-value

38,712 71,974 0

42,932 2834 0

38,712 67,519 0

42,932 2822 0

38,712 68,222 0

42,932 2792 0

25,323 12,161 0

31,493 5502 0

25,323 8294 0

31,493 5158 0

25,323 8317 0

31,493 6201 0

This table presents the results of random-effect Tobit panel regressions on the determinants of grey/independent foreign institutional ownership. The dependent variable in the regressions is the firm-level foreign ownership by grey (independent) institutional investors in the United States. We classify banks, insurance companies and others (investment advisors/companies and public pension funds) as grey (independent) investors. This measures firm-level equity holdings of grey/independent investors domiciled in each country (i.e., country-level holdings in each U.S. firm in our sample). In Panel A (Panel B) we run the regressions on the full sample (excluding U.K. investors). GQ is the level of a country's governance quality calculated as the average of the six KKM governance indicators. GQd is a dummy variable which equals one if the governance quality of a country (GQ) in year t is greater than that of the U.S. in that year or zero otherwise. DINDEX is the directors' index defined as in Bushee et al. (2010). It is based on information regarding CEO duality, number of board directors, the percentage of directors who are not independent, interlocks on the board of directors and attendance of board meetings. A value of five (one) indicates a poorly (well)-governed firm. GQd*D is equal to the country-level governance quality dummy (GQd) multiplied by DINDEX. GQd*MC is equal to the country-level governance quality dummy (GQd) multiplied by MC. MC is monitoring cost, which is equal to 1/Turnover. SIZE is the logarithm of firm size which is defined as the fiscal year-end market value of equity. BM is the logarithm of the ratio of the firm's book value of equity to its market value of equity. DY is dividend yield, defined as the dividend per share divided by the share price. LEV is leverage, which is equal to the ratio of total debt to total assets. ROE is return on equity, defined as net income divided by common equity. CASH is cash holdings, defined as the ratio of cash and short-term investment to total assets. TURN is a firm's market turnover, which is equal to common shares traded divided by common shares outstanding. The turnover ratio is first calculated monthly. Annual turnover is defined as the average of the twelve monthly observations. Leverage, dividend yield and return on equity are winsorised at 1% (two tails). Year dummies (Year FE) and country dummies (Country FE) are included in all regressions. The numbers in brackets are p-values. * indicates 10% significance level, ** indicates 5% significance level and *** indicates 1% significance level. N is the number of observations.

Table 7

Alternative proxies for governance quality at home.

FIO (%)

Td*D -0.136*** [0.000]

Td*MC -0.002 [0.661]

Hd*D -0.106*** [0.000]

Hd*MC -0.034*** [0.000]

Td -0.439*** [0.000] -0.673*** [0.000]

MC -0.015*** [0.000] 0.016*** [0.000]

Hd 0.526*** [0.000] 0.594*** [0.000]

SIZE 0.024*** [0.000] 0.022*** [0.000] 0.017*** [0.001] 0.012** [0.028]

DY 1.094*** [0.004] 1.481*** [0.000] 0.835** [0.019] 1.218*** [0.001]

BM -0.015 [0.209] -0.011 [0.339] -0.013 [0.246] -0.007 [0.503]

TURN 0.408*** [0.000] 0.406*** [0.000]

LEV 0.06 [0.197] 0.057 [0.226] 0.063 [0.147] 0.062 [0.151]

ROE 0.097*** [0.000] 0.092*** [0.000] 0.088*** [0.000] 0.085*** [0.000]

CASH 0.012 [0.827] 0.034 [0.523] 0.01 [0.837] 0.049 [0.311]

DINDEX -0.027*** [0.001] -0.032*** [0.000] 0.052*** [0.000] -0.030*** [0.000]

Constant -1.642*** [0.000] -1.354*** [0.000] -1.638*** [0.000] -1.361*** [0.000]

Year FE Yes Yes Yes Yes

Country FE Yes Yes Yes Yes

N 52,767 52,767 58,665 58,665

Chi2 69,748 70,068 75,315 75,842

p-value 0 0 0 0

This table presents the results of random-effect Tobit panel regressions on the determinants of foreign institutional ownership using alternative proxies for a country's governance quality. The dependent variable in all the regressions is the firm-level foreign institutional ownership in the U.S. (FIO). This measures firm-level equity holdings of all investors domiciled in a given country (i.e., country-level holdings in each U.S. firm in our sample). Td*D is an interaction term using the country-level turnover dummy (Td) and DINDEX. Td is equal to one if the turnover ratio of a country in year t is greater than that of the U.S. in that year and zero otherwise. DINDEX is the directors' index, defined as in Bushee et al. (2010). It is based on information regarding CEO duality, number of board directors, the percentage of directors who are not independent, interlocks on the board of directors and attendance of board meetings. A value of five (one) indicates a poorly (well)-governed firm. Td*MC is an interaction term based on the product of the country-level turnover dummy (Td) and monitoring cost (MC). MC is equal to 1/Turnover. Hd*D is the interaction term based on the product of the U.S. holding dummy (Hd) and DINDEX. Hd*MC is the interaction term based on the product of the U.S. holding dummy (Hd) and MC. Hd is equal to one if the level of U.S. holdings in a country is greater than the median U.S. holdings for the cross-section of sampled countries in that year and zero otherwise. SIZE is the logarithm of firm size, which is defined as the fiscal year-end market value of equity. BM is the logarithm of the ratio of the firm's book value of equity to its market value of equity. DY is dividend yield, defined as the dividend per share divided by the share price. LEV is leverage, which is equal to the ratio of total debt to total assets. ROE is return on equity, defined as net income divided by common equity. CASH is cash holdings, defined as the ratio of cash and short-term investment to total assets. TURN is a firm's market turnover, which is equal to common shares traded divided by common shares outstanding. The turnover ratio is first calculated monthly. Annual turnover is defined as the average of the twelve monthly observations. Leverage, dividend yield and return on equity are winsorised at 1% (two tails). Year dummies (Year FE) and country dummies (Country FE) are included in all regressions. The numbers in brackets are p-values. * indicates 10% significance level, ** indicates 5% significance level and *** indicates 1% significance level. N is the number of observations.

literature has routinely used market liquidity as a proxy for market development, our research design limits its usability. This is because the U.S. has one of the most developed markets and hence one of the highest turnover ratios. This leaves us with a heavily unbalanced split between the Above- and Below-U.S. turnover groups, with only six percent of our observations classified into the Above-U.S. group.

To address this issue, in the last two columns of Table 7, we show the results from following prior studies on U.S. institutional investment abroad and classifying countries as more (less) developed depending on whether they have above (below) median U.S. investment levels. Hd is equal to one if the level of U.S. holdings in a country is greater than the median U.S. holdings in the cross-section of sampled countries in that year and zero otherwise. Hd*D is the interaction term based on the product of the U.S. holdings dummy (Hd) and DINDEX. In order to create Hd, we obtain the level of U.S. portfolio holdings of each country.23 We find a negative relation between the interaction term and FIO further

23 We retrieve the data on US holdings abroad from the Bureau of Economic Analysis reports, available at http://www.bea.gov/ international/dilusdbal.htm.

Table 8

Alternative specifications: SOX and firm fixed effects.

F1O (%)

RE Tobit RE Tobit RE Tobit FE OLS FE OLS FE OLS

GQ —5.499* ** [0.000] —4.317*** [0.000]

GQd*D —0.196* ** [0.000] —0.156*** [0.000]

GQd*MC —0.044* ** [0.000] —0.044*** [0.000]

GQd 0.236*** [0.000] 0.201*** [0.000] 0.239*** [0.000] 0.280*** [0.000]

MC 0.005*** [0.010] 0.021*** [0.000]

SIZE 0.036*** [0.000] 0.029*** [0.000] 0.022*** [0.000] 0.060* [0.100] 0.056 [0.138] 0.058 [0.127]

DY 2.189*** [0.000] 2.652*** [0.000] 3.113*** [0.000] —0.212 [0.836] —0.358 [0.733] —0.062 [0.952]

BM 0.020* [0.064] 0.029*** [0.007] 0.032*** [0.003] —0.019 [0.565] —0.037 [0.274] —0.025 [0.460]

TURN 0.691*** [0.000] 0.790*** [0.000] 0.492*** [0.000] 0.520*** [0.000]

LEV 0.022 [0.606] —0.016 [0.717] —0.016 [0.712] —0.226 [0.181] —0.227 [0.195] —0.2 [0.251]

ROE 0.125*** [0.000] 0.138*** [0.000] 0.124*** [0.000] 0.047 [0.323] 0.035 [0.472] 0.021 [0.677]

CASH —0.085* [0.080] —0.125* * [0.012] —0.053 [0.274] 0.174 [0.327] 0.155 [0.396] 0.212 [0.251]

DINDEX —0.064* ** [0.000] 0.053*** [0.000] —0.072* ** [0.000] —0.015 [0.287] 0.086*** [0.000] —0.013 [0.354]

SOX 0.339*** [0.000] 0.543*** [0.000] 0.482*** [0.000]

Constant 7.965*** [0.000] —1.190* ** [0.000] —0.739* ** [0.000] 5.342*** [0.000] — 1.408*** [0.000] —0.764** [0.020]

Firm effect RE RE RE FE FE FE

Year FE No No No Yes Yes Yes

Country FE Yes Yes Yes Yes Yes Yes

N 59,491 59,491 59,491 59,491 59,491 59,491

Chi2 77,665 71,366 72,518 - - -

p-value 0 0 0 - - -

R2 - - - 0.132 0.101 0.105

F-stat - - - 49.48 162.48 54.35

p-value - - - 0 0 0

This table presents the results of random-effect Tobit and fixed effect OLS panel regressions. The dependent variable in all the regressions is the firm-level foreign institutional ownership in the U.S. (FIO). This measures firm-level equity holdings of all investors domiciled in a given country (i.e., country-level holdings in each U.S. firm in our sample). GQ is the level of a country's governance quality, calculated as the average of the six KKM governance indicators. GQd is a dummy variable which equals one if the governance quality of a country (GQ) in year t is greater than that of the U.S. in that year and zero otherwise. DINDEX is the directors' index, defined as in Bushee et al. (2010). It is based on information regarding CEO duality, number of board directors, the percentage of directors who are not independent, interlocks on the board of directors and attendance of board meetings. A value of five (one) indicates a poorly (well)-governed firm. GQd*D is equal to the country-level governance quality dummy (GQd) multiplied by DINDEX. GQd*MC is equal to the country-level governance quality dummy (GQd) multiplied by MC. MC is monitoring cost, which is equal to 1/Turnover. SIZE is the logarithm of firm size, which is defined as the fiscal year-end market value of equity. BM is the logarithm of the ratio of the firm's book value of equity to its market value of equity. DY is dividend yield, defined as the dividend per share divided by the share price. LEV is leverage, which is equal to the ratio of total debt to total assets. ROE is return on equity, defined as net income divided by common equity. CASH is cash holdings, defined as the ratio of cash and short-term investment to total assets. TURN is a firm's market turnover, which is equal to common shares traded divided by common shares outstanding. The turnover ratio is first calculated monthly. Annual turnover is defined as the average of the twelve monthly observations. Leverage, dividend yield and return on equity are winsorised at 1% (two tails). SOX is a dummy variable taking the value of one for the period after 2002 and zero otherwise. Year dummies (Year FE) and country dummies (Country FE) are included in all regressions. The numbers in brackets are p-values. * indicates 10% significance level, ** indicates 5% significance level and *** indicates 1% significance level. N is the number of observations.

confirming our results, i.e., investors from more developed countries invest in well-governed U.S. firms. The result remains even if we use MC, instead of DINDEX, as our proxy for information asymmetry in a firm.

We would like to highlight that in all our analyses we use country fixed effects. These fixed effects capture the impact of omitted time invariant country characteristics, such as language, educational and cultural bonds, closeness/familiarity, etc. However, there are macroeconomic factors that might change over time, in which case their impact won't be captured by our country fixed effects. Still, our time-series is relatively short, compared to the typical periods examined in macroeconomics. As we argue in footnote 19, most of the macro factors will change little over the 10 years we examine. In order to alleviate concerns that our results are driven by omitted macro effects, we construct some of

the variables used in the extant literature and re-estimate our main models presented in Tables 4 and 5 (untabulated results). In particular, instead of the country dummies we use 3 country-specific variables in all specifications: GDPPC is the gross domestic product per capita (Chan et al., 2005); DISTANCE is a measure of geographical distance which captures the bilateral distance of capital cities of countries, and is used as a proxy for closeness (Frankel et al., 1995); MCAPGDP is the relative size of the stock market of each country, measured by the stock market capitalization as a percentage of the country's GDP and proxies for stock market development (Forbes, 2010).24 Our results remain largely unchanged.

Finally, we present our results using alternative model specifications. First, instead of controlling for time trends (market-wide effects) using year dummies, we control for the effect of the enactment of the Sarbanes-Oxley Act (SOX) in 2002. SOX imposes higher disclosure and more accountability on firms listed in the U.S. exchanges. Therefore, it is expected to increase the internal corporate governance quality of U.S. firms. We re-run our results using a SOX dummy that takes the value of one (zero) post-(pre-) SOX, instead of the year dummies. Table 8, columns 1-3, reports the results of this specification. We obtain results consistent with the prior analyses. Second, the use of Tobit models forces us to control for random unobserved firm effects. To test the impact of the random firm effects assumption, we run our analysis using OLS regressions. This allows us to assume fixed firm effects, but at the cost of using a suboptimal regression technique (given the censoring of the dependent variable, which makes it left-clustered). The results remain unchanged.

7. Conclusion

The main contribution of this paper lies in the examination of the influence of the foreigners' country-level governance quality on their investment preferences when they invest in the United States. We reconcile prior evidence in the literature by presenting results consistent with both the familiarity and flight to quality arguments. Institutional investors from countries with governance quality similar to that of the U.S. invest more in U.S. firms. But investors from countries with governance quality just below (just above) that of the U.S. invest more (less) in comparison.

We also present evidence that governance quality at home also affects firm-level investment preferences of foreign investors. Our results indicate that foreign institutional investors from high governance quality (low information asymmetry) countries invest in U.S. firms with high corporate governance quality (less information asymmetry). Our evidence supports the familiarity argument, and highlights information asymmetry costs as a cause of familiarity, particularly for non-U.K. investors. Our findings, both at the country and firm level are robust to the use of different variable and model specifications.

In sum, we contribute to the foreign institutional investment literature by further investigating the role of home governance. Our research design allows us to explore the heterogeneous preferences of foreign institutional investors depending on the distance of their countries' governance quality from that of the United States. This helps us to reconcile contradictory evidence presented in prior studies on the validity of the flight to quality and familiarity arguments.

Acknowledgements

We would like to thank an anonymous referee, the participants of the FMA European Conference (Portugal 2011) and our colleagues at Manchester Business School for their valuable comments and suggestions on earlier drafts of the paper. All errors are our responsibility.

24 In additional tests, we have also used LANGUAGE, that is a dummy variable taking the value of 1 for English speaking countries, zero otherwise, and CAPITAL CONTROL, which is a measure of "freedom" investors enjoy in a capital market. These two variables are highly correlated with GDPPC and DISTANCE, so we cannot use them in the same model.

Appendix

Table A.1

Ownership by foreign institutional investor's country of domicile.

Country 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008

United Kingdom 2.848% 2.950% 3.506% 3.571% 1.608% 5.187% 5.329% 5.317% 5.891% 6.307%

Australia 0.000% 0.000% 0.000% 0.000% 0.000% 0.000% 0.000% 0.000% 0.001% 0.007%

Bahamas 0.000% 0.000% 0.000% 0.000% 0.000% 0.000% 0.000% 0.001% 0.002% 0.008%

Barbados 0.000% 0.000% 0.000% 0.000% 0.000% 0.000% 0.000% 0.000% 0.000% 0.000%

Belgium 0.000% 0.000% 0.000% 0.000% 0.000% 0.000% 0.000% 0.000% 0.000% 0.024%

Bermuda 0.062% 0.085% 0.110% 0.137% 0.143% 0.130% 0.137% 0.158% 0.138% 0.055%

Canada 0.879% 0.573% 0.651% 0.406% 0.458% 0.355% 0.407% 0.520% 0.641% 0.701%

Cayman islands 0.002% 0.002% 0.003% 0.003% 0.001% 0.001% 0.002% 0.008% 0.013% 0.003%

Denmark 0.000% 0.000% 0.000% 0.028% 0.056% 0.033% 0.040% 0.044% 0.040% 0.037%

France 0.015% 0.016% 0.021% 0.019% 0.020% 0.022% 0.074% 0.151% 0.195% 0.152%

Germany 0.006% 0.000% 0.000% 0.000% 0.000% 0.000% 0.135% 0.153% 0.044% 0.015%

Ireland (Republic of) 0.023% 0.031% 0.038% 0.050% 0.068% 0.073% 0.054% 0.028% 0.044% 0.051%

Japan 0.105% 0.142% 0.123% 0.150% 0.140% 0.148% 0.178% 0.196% 0.139% 0.167%

Netherlands 0.028% 0.033% 0.092% 0.125% 0.179% 0.196% 0.179% 0.196% 0.246% 0.381%

Norway 0.000% 0.000% 0.000% 0.000% 0.000% 0.000% 0.000% 0.093% 0.255% 0.439%

Singapore 0.013% 0.019% 0.030% 0.067% 0.073% 0.068% 0.063% 0.064% 0.057% 0.043%

Switzerland 0.033% 0.037% 0.051% 0.066% 0.062% 0.049% 0.064% 0.071% 0.048% 0.032%

Taiwan 0.000% 0.062% 0.079% 0.006% 0.000% 0.000% 0.000% 0.000% 0.000% 0.000%

Total 4% 4% 5% 5% 3% 6% 7% 7% 8% 8%

This table shows the foreign institutional investments in the U.S. between 1999 and 2008. We report the percentage of shares held by all foreign institutional investors domiciled in each country. The investment levels are reported for the constituent firms of the S&P1500 index.

Table A.2

Time-series changes in governance quality (GQ) by country.

Country 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008

Australia 1.60 1.63 1.63 1.54 1.57 1.67 1.56 1.55 1.58 1.62

Bahamas 1.17 1.22 1.22 1.23 1.15 1.15 1.14 1.10 1.10 1.12

Barbados 1.30 1.30 1.30 1.26 1.21 1.18 1.20 1.11 1.17 1.18

Belgium 1.30 1.36 1.36 1.44 1.40 1.36 1.28 1.27 1.27 1.21

Bermuda 1.19 1.13 1.13 1.14 1.14 1.14 1.07 1.06 1.05 1.07

Canada 1.64 1.62 1.62 1.64 1.63 1.61 1.53 1.59 1.57 1.60

Cayman islands 1.36 1.38 1.38 1.39 1.15 1.20 1.16 1.17 1.17 1.07

Denmark 1.74 1.75 1.75 1.81 1.80 1.85 1.77 1.82 1.84 1.82

France 1.15 1.22 1.22 1.22 1.23 1.31 1.26 1.25 1.22 1.24

Germany 1.55 1.60 1.60 1.56 1.44 1.45 1.46 1.50 1.49 1.43

Ireland (Republic of) 1.53 1.53 1.53 1.50 1.46 1.45 1.52 1.56 1.56 1.57

Japan 1.04 1.06 1.06 0.96 1.12 1.17 1.16 1.24 1.18 1.16

Netherlands 1.81 1.82 1.82 1.74 1.69 1.70 1.63 1.60 1.62 1.61

Norway 1.73 1.61 1.61 1.69 1.65 1.72 1.64 1.66 1.64 1.64

Singapore 1.44 1.44 1.44 1.44 1.41 1.49 1.43 1.40 1.45 1.51

Switzerland 1.71 1.75 1.75 1.75 1.69 1.74 1.64 1.68 1.71 1.70

Taiwan 0.80 0.80 0.80 0.84 0.88 0.91 0.88 0.78 0.73 0.79

United kingdom 1.64 1.60 1.60 1.55 1.48 1.52 1.42 1.50 1.45 1.40

United states 1.44 1.52 1.52 1.40 1.34 1.33 1.24 1.25 1.22 1.28

This table reports the time-series changes in GQ for all countries examined in this study. GQ is the average of the six KKM governance indicators. The six governance indicators are scaled from -2.5 to 2.5, with higher values corresponding to better governance outcomes.

Table A.3

Details of the directors' index (DINDEX).

DINDEX scores

0 1 2 3 4 5

Panel A: Average values per governance indicator Variable

CEO 0 0.67 0.91 0.95 0.99 1

LNDIR 0 0.13 0.6 0.82 0.95 1

PNID 0 0.16 0.4 0.71 0.93 1

DLOCK 0 0.00 0.02 0.14 0.47 1

DBAD 0 0.02 0.08 0.38 0.66 1

Obs. 1216 4427 4940 1928 438 68

DINDEX 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008

Panel B: Time-series changes in DINDEX

0 55 52 63 61 92 115 138 154 261 225

4% 4% 5% 5% 7% 8% 10% 11% 22% 18%

1 286 332 364 404 442 486 527 507 532 547

22% 26% 29% 31% 33% 36% 38% 38% 45% 43%

2 510 500 530 517 532 536 559 496 336 424

40% 39% 42% 40% 40% 40% 41% 37% 28% 33%

3 297 295 306 229 207 173 122 162 57 80

23% 23% 24% 18% 16% 13% 9% 12% 5% 6%

4 103 77 58 68 46 39 21 17 7 2

8% 6% 5% 5% 3% 3% 2% 1% 1% 0%

5 20 15 12 8 2 4 3 4 0 0

2% 1% 1% 1% 0% 0% 0% 0% 0% 0%

Obs. 1271 1271 1270 1287 1321 1353 1370 1340 1193 1278

This table presents the descriptive statistics for the directors' index (DINDEX). DINDEX is defined as in Bushee et al. (2010). It is based on information regarding CEO duality, number of board directors, the percentage of directors who are not independent, interlocks on the board of directors and attendance of board meetings. A value of five (one) indicates a poorly(well)-governed firm. Panel A shows the average value of each governance indicator for each level of DINDEX. Panel B shows the time-series changes in DINDEX. The numbers show the number of firms with each score. The percentages are calculated by scaling the number of firms by the total observations per year. CEO measures CEO duality. It is equal to one if the positions of CEO and Chairman are combined and zero otherwise. LNDIR is the logarithm of the number of directors. PNID is the percentage of directors that are not independent. To form these indicators, we split the distribution of LNDIR and PNID into high and low groups using k-means cluster analysis. For the high (low) group the variable equals one (zero). DLOCK is equal to one if there are any interlocks on the board of directors and zero otherwise. DBAD signifies poor attendance and equals one if any director misses 75% or more of the board meetings and zero otherwise.

References

Aggarwal, R., Klapper, L., Wysocki, P.D., 2005. Portfolio preferences of foreign institutional investors. Journal of Banking and Finance 29, 2919-2946.

Aggarwal, R., Erel, I., Ferreira, M., Matos, P., 2011. Does governance travel around the world? Evidence from institutional investors. Journal of Financial Economics 100 (1), 154-181.

Ahearne, A.G., Griever, W.L., Warnock, F.E., 2004. Information costs and home bias: an analysis of US holdings of foreign equities. Journal of International Economics 62 (2), 313-336.

Almazan, A., Hartzell, J.C., Starks, L.T., 2005. Active institutional shareholders and costs of monitoring: evidence from executive compensation. Financial Management 34 (4), 5-34.

Badrinath, S.G., Gay, G.D., Kale, J.R., 1989. Patterns of institutional investment, prudence and the managerial 'safety-net' hypothesis. Journal of Risk and Insurance 56 (4), 605-629.

Bathala, C.T., Moon, K.P., Rao, R.P., 1994. Managerial ownership, debt policy, and the impact of institutional holdings: an agency perspective. Financial Management 23 (3), 38-50.

Bekaert, G., Harvey, C.R., Lumsdaine, R.L., 2002. Dating the integration of world capital markets. Journal of Financial Economics 65 (2), 203-247.

Beltratti, A., Stulz, R.M., 2009. Why Did Some Banks Perform Better during the Credit Crisis? A Cross-country Study of the Impact of Governance and Regulation. NBER Working Paper No 15180.

Bennett, J.A., Sias, R.W., Starks, L.T., 2003. Greener pastures and the impact of dynamic institutional preferences. Review of Financial Studies 16 (4), 1203-1238.

Brennan, M.J., Cao, H.H., 1997. International portfolio investment flows. Journal of Finance 52 (5), 1851-1880.

Bushee, B.J., Carter, M.E., Gerakos, J., 2010. Institutional Investor Preferences for Corporate Governance Mechanisms. The Wharton School of the University of Pennsylvania Working Paper.

Cai, F., Warnock, F.E., 2006. International Diversification at Home and Abroad. NBER Working Paper No 12220.

Caprio, L., Faccio, M., McConnell, J.J., forthcoming. Sheltering corporate assets from political extraction. The Journal of Law, Economics and Organization.

Chan, K., Covrig, V., Ng, L., 2005. What determines the domestic bias and foreign bias? Evidence from mutual fund equity allocations worldwide. Journal of Finance 60 (3), 1495-1534.

Coval, J.D., Moskowitz, T.J., 1999. Home bias at home: local equity preference in domestic portfolios. Journal of Finance 54 (6), 2045-2073.

Cready, W., 1994. Determinants of relative investor demand for common stocks. Journal of Accounting, Auditing, & Finance 9 (3), 487-507.

Dahlquist, M., Robertsson, G., 2001. Direct foreign ownership, institutional investors and firm characteristics. Journal of Financial Economics 59, 413-440.

Defond, M.L., Hung, M., 2004. Investor protection and corporate governance: evidence from worldwide CEO turnover. Journal of Accounting Research 42 (2), 269-312.

Del Guercio, D., 1996. The distorting effect of the prudent man laws on institutional equity investments. Journal of Financial Economics 40 (1), 31-62.

Doidge, C., Karolyi, G.A., Stulz, R.M., 2007. Why do countries matter so much for corporate governance? Journal of Financial Economics 86, 1 -39.

Dvorak, T., 2003. Gross capital flows and asymmetric information. Journal of International Money and Finance 22, 835-864.

Edison, H., Warnock, F., 2004. U.S. investors' emerging market equity portfolios: a security-level analysis. Review of Economics and Statistics 84 (3), 691 -704.

Falkenstein, E.G., 1996. Preferences for stock characteristics as revealed by mutual fund portfolio holdings. Journal of Finance 51 (1), 111-135.

Ferreira, M.A., Matos, P., 2008. The colors of investors' money: the role of institutional investors around the world. Journal of Financial Economics 88 (3), 499-533.

Forbes, K.J., 2010. Why do foreigners invest in the United States? Journal of International Economics 80 (1), 3-21.

Frankel, J., Stein, E., Wei, S., 1995. Trading blocs and the Americas: the natural, the unnatural, and the super-natural? Journal of Development Economics 47, 61 -95.

Gelos, R.G., Wei, S.-J., 2005. Transparency and international portfolio holdings. Journal of Finance 60 (6), 2987-3020.

Giannetti, M., Koskinen, Y., 2010. Investor protection, equity returns and financial globalization. Journal of Financial and Quantitative Analysis 45 (1), 135-168.

Giannetti, M., Simonov, A., 2006. Which investors fear expropriation? Evidence from investors' portfolio choices. Journal of Finance 61 (3), 1507-1547.

Gompers, P.A., Metrick, A., 2001. Institutional investors and equity prices. Quarterly Journal of Economics 116 (1), 229-259.

Grinblatt, M., Keloharju, M., 2000. The investment behavior and performance of various investor types: a study of Finland's unique data set. Journal of Financial Economics 55, 43-67.

Huberman, G., 2001. Familiarity breeds investment. Review of Financial Studies 14 (3), 659-680.

Hussain, S., 2000. Simultaneous determination of UK analyst following and institutional ownership. Accounting and Business Research 30 (2), 111-124.

Jensen, M.C., Meckling, W.H., 1976. Theory of the firm: managerial behavior, agency costs, and ownership structure. Journal of Financial Economics 3 (4), 305-360.

Kang, J.-K., Stulz, R.M., 1997. Why is there a home bias? An analysis of foreign portfolio equity ownership in Japan. Journal of Financial Economics 46 (1), 3-28.

Kaufmann, D., Kraay, A., Mastruzzi, M., 2007. Governance Matters VI: Aggregate and Individual Governance Indicators 19962006. World Bank Policy Research. Working Paper. No 4280.

Khurshed, A., Lin, S., Wang, M., 2011. Institutional block-holdings in UK firms: do corporate control mechanisms matter? European Journal of Finance 17 (2), 133-152.

Kim, W., Sung, T., Wei, S.J., 2011. How does corporate governance risk at home affect investment choices abroad? Journal of International Economics 85 (1), 25-41.

Klapper, L.F., Love, I., 2004. Corporate governance, investor protection, and performance in emerging markets. Journal of Corporate Finance 10 (5), 703-728.

La Porta, R., Lopez-de-Silanes, F., Shleifer, A., Vishny, R.W., 1997. Legal determinants of external finance. Journal of Finance 52 (3), 1131-1150.

La Porta, R., Lopez-de-Silanes, F., Shleifer, A., Vishny, R.W., 1998. Law and finance. Journal of Political Economy 106 (6), 11131155.

La Porta, R., Lopez-de-Silanes, F., Shleifer, A., Vishny, R.W., 2000. Agency problems and dividend policies around the world. Journal of Finance 55 (1), 1-33.

La Porta, R., Lopez-de-Silanes, F., Shleifer, A., 2006. What works in securities laws? Journal of Finance 61 (1), 1-32.

Lakonishok, J., Shleifer, A., Vishny, R.W., 1994. Contrarian investment, extrapolation and risk. Journal of Finance 49 (5), 1541-1578.

Leuz, C., Lins, K.V., Warnock, F.E., 2009. Do foreigners invest less in poorly governed firms? Review of Financial Studies 22 (8), 3245-3285.

Lewis, K.K., 1999. Trying to explain home bias in equities and consumption. Journal of Economic Literature 37 (2), 571-608.

Li, D., Moshirian, F., Pham, P.K., Zein, J., 2006. When financial institutions are large shareholders: the role of macro corporate governance environments. Journal of Finance 61 (6), 2975-3007.

McKinsey & Company, 2002. Global Investor Opinion Survey: Key Findings.

Merton, R.C., 1987. A simple model of capital market equilibrium with incomplete information. Journal of Finance 42 (3), 483-510.

Tesar, L.L., Werner, I.M., 1995. Home bias and high turnover. Journal of International Money and Finance 14 (4), 467-492.