Foreign Institutional Ownership, Risk-Taking, and Crash ... · Crash Risk around the World Abstract...

1

Foreign Institutional Ownership, Risk-Taking, and

Crash Risk around the World

Abstract

Using a large sample of 48,548 firms in 72 countries from 2000 to 2008, we examine how foreign

institutional ownership affects corporate risk-taking and stock price crash risk. We find foreign

institutional ownership is positively associated with corporate risk-taking, while domestic institutional

ownership is negatively associated with it. Foreign institutional ownership substitutes for country-level

corporate governance in determining corporate risk-taking. Crash risk is a potential negative side effect

associated with risk-taking. We find when a firm takes more risks, foreign direct investors’ ownership

significantly reduces the firm’s crash risk. In contrast, domestic institutional investors’ ownership

significantly increases it.

Key words: Institutional ownership, risk taking, crash risk

JEL classification: G32, G34

Garland Huang

Australian Business

School

University of New South

Wales

Kensington NSW 2033 Sydney, Australia

Email:garland.huang@st

udent.unsw.edu.au

Donghui Li

Australian Business

School


Wales


Email:[email protected].

au

Sheng Xiao

Gore Business School

Westminster College

1840S 1300 S Salt Lake City, UT 84105

USA

Email:

[email protected]

Zhe An

Australian Business

School


Wales


Email:[email protected]

.au

mailto:[email protected]







2

1. Introduction

Corporate innovations are a major determinant of firm performance and economic growth.1 To

innovate, firms must take risks. On one hand, such risk-taking activities may lead to breakthroughs in

business models and technology. On the other hand, such risk-taking activities may also magnify the

crash risk of a firm’s stock price, especially when the firm manager hoards bad news. The ownership

structure of a firm, especially the firm’s institutional ownership, has significant effects on both corporate

risk-taking activities and stock price crash risk (henceforth “crash risk”).2However, not all institutional

investors play the same role in affecting firms’ risk-taking activities and crash risk. We examine how a

particular type of institutional owner, the foreign institutional owner, affects a firm’s risk-taking

behaviors and crash risk. For the purpose of comparison, we also examine how domestic institutional

owners affect a firm’s risk-taking behaviors and crash risk. Our research is inspired by cross-country

research by Ferreira and Matos (2008), who find that firms with higher foreign institutional ownership

have higher valuations and better operating performance. Our research provides an important channel

through which foreign institutional ownership leads to better firm performance: promoting corporate risk-

taking.

We focus on foreign institutional ownership for several reasons: first, foreign institutional owners

are likely to be stronger monitors, as they tend to have fewer conflicts of interest with the firm. In

contrast, domestic owners are unlikely to have the same authority over a firm due to their existing

business relationships with managers, which may prevent them from being efficient monitors (Ferreira

and Matos 2008). Second, diversification is more prominent among foreign owners due to their

1See, e.g., Hirshleifer, Hsu, and Li (2013); Mankiw, Romer, and Weil (1992).

2Papers on how institutional investors affect corporate risk-taking and innovative activities include:

Wright, Perris, Sarin, and Awasthi (1996); Faccio, Marchica, and Mura (2011); Aghion, Reenen, and

Zingales (2013), among others. Papers on how institutional investors affect crash risks include: An and

Zhang (2013); Callen and Fang (2013), among others.

3

internationally diversified portfolios. Such international diversification motivates foreign owners to push

managers to take more risks (Faccio, Marchica, and Mura 2011).Third, foreign institutional investors

bring additional funds that reduce the reliance of the invested firm on banks, which usually influence a

firm to pursue conservative investment policies (Morck and Nakamura 1999). These three factors imply

that foreign institutional investors, compared with domestic institutional investors, can better motivate

managers to take sufficient risks. Therefore, foreign institutional ownership is expected to be positively

associated with corporate risk-taking. On the other hand, foreign institutional owners may face more

severe information asymmetry (Chan, Menkveld, and Yang2008; Portes and Rey 2005). As a result, they

may be less effective monitors. This implies that foreign institutional ownership is expected to be

negatively associated with corporate risk-taking. Therefore, whether foreign institutional ownership

increases or decreases corporate risk-taking activities is an empirical question.

John, Litov, and Yeung (2008) show that country-level governance institutions such as investor-

protection mechanisms also promote corporate risk-taking. A natural question to ask is how foreign

institutional ownership and country-level governance institutions jointly affect corporate risk-taking. That

is, are they substitutes or complements? We empirically test their interactions in the determination of

corporate risk-taking.

Existing literature appears to focus on the positive effects of corporate risk-taking on corporate

growth and economic growth (John, Litov, and Yeung 2008). However, it is also important to consider its

potential negative side effects. One of the most fundamental negative side effects associated with risk-

taking is the increase in the firm’s crash risk. This is because managers who are motivated to actively

seek risky projects will inherently face increased exposure to losses as a result of these risky projects.

Managers who incur these losses tend to regularly manage earnings by withholding bad news due to

managerial incentives such as career concerns and compensation contracts (Kothari, Shu, and Wysocki

2009). As a result, bad news associated with poorly performing projects tends to stockpile within a firm.

When bad news accumulates to a certain threshold, managers are no longer able to hide the bad news

effectively, and all the negative information will be released to the market at once. This leads to an

4

extreme decline in stock price, which is the “crash risk” (Hutton, Marcus, and Tehranian 2009; Kim, Li,

and Zhang 2011a, b).

How do foreign institutional owners affect firms’ crash risk? Cornett, Marcus, and Tehranian

(2008) find that institutional investors are able to constrain firm earnings management. Therefore, we

hypothesize that if foreign institutional owners are effective monitors, they will reduce their firms’ crash

risk. This hypothesis is also based on the fact that foreign owners face more severe information

asymmetry than domestic owners. Therefore, they generally demand more corporate transparency, which

allows shareholders to discriminate between good and bad projects at an early stage, thus reducing the

crash risk associated with bad projects. For example, Bleck and Liu (2007) argue that greater financial

reporting opacity increases the crash risk of the firm’s stock price.

Considering the heterogeneity of foreign institutional owners, we further examine the effect of

foreign institutional ownership in two different forms of investments: (1) Foreign direct investment (FDI),

defined as a long-term-relationship investment that reflects a lasting interest and control of a resident

entity in an economy other than that of the foreign direct investor (UNCTAD(United Nations Conference

on Trade and Development)2002) and (2) Foreign portfolio investment (FPI), representing passive

holdings of foreign securities such as ownership through stocks and bonds. FPI investors generally have

shorter investment time horizons and do not exert control over firms. We hypothesize that FDI ownership

significantly reduces crash risk while FPI ownership has insignificant effects on crash risk. This is

because FDI investors are long-term institutional investors who tend to be stronger monitors than short-

term institutional investors (Chen, Harford, and Li 2007).

Examining 48,548 firms in 72 countries from the years 2000 to 2008, we find that foreign

institutional ownership is positively associated with corporate risk-taking. We also find that foreign

institutional ownership substitutes for country-level governance institutions (information transparency,

legal environment, and investor protection) in determining corporate risk-taking. Further, we examine

how foreign institutional ownership affects crash risk, a potential negative side effect associated with

5

additional risk-taking. We find that when a firm takes more risks, foreign direct ownership significantly

reduces the firm’s crash risk. In contrast, risk-taking by domestic owners significantly increases it.

Our study contributes to the literature in several ways. First, to the best of our knowledge, we are

the first to show that foreign institutional ownership promotes corporate risk-taking in a large sample of

48,548 firms in 72 developing and developed countries. This shows that foreign institutional owners are

indeed stronger monitors, despite their information disadvantages compared with domestic institutional

owners. We are also the first to show that foreign institutional ownership and country-level governance

institutions are substitutes. The research that most resembles ours is Boubakri, Cosset, and Saffar (2013),

who show that foreign institutional owners promote corporate risk-taking in a smaller sample of 381

newly privatized firms in 57 countries. But Boubakri, Cosset, and Saffar (2013)find foreign institutional

ownership and country-level governance institutions are complements in the determination of corporate

risk-taking while we find they are substitutes. Our research not only significantly expands their dataset

beyond 381 newly privatized firms (a special group of firms that have experienced the change of control

from the government to private owners), but also finds some results that are exactly opposite to those in

Boubakri, Cosset, and Saffar (2013).

Second, to the best of our knowledge, we are the first to show that despite the fact that foreign

institutional investors lead to more corporate risk-taking, they significantly lower the crash risk. To

further investigate which foreign institutional investors are responsible for the decrease in crash risk, we

differentiate between foreign direct investors and foreign portfolio investors. We find that foreign direct

investors significantly reduce the crash risk of a firm while foreign portfolio investors have an

insignificant effect on the crash risk. In contrast, domestic institutional owners exacerbate crash risk. Our

research contributes to the nascent literature on the determinants of crash risk.

Our research is important in the context of increasing globalization, a result of the recent wave of

capital market liberalizations around the world. According to the World Investment Report2010 and 2013,

global FDI rose from $154 billion in 1991 to $1.35 trillion in 2013, and global foreign portfolio equity

investments increased from $106 billion in 1991 to $744 billion in 2010. As foreign capital becomes an

6

increasingly important source of financing around the world (Bekaert, Harvey, and Lumsdaine2002), it is

essential to understand the effects of foreign institutional ownership on corporate decisions and

performance.

2. Literature Review and Hypothesis Development

Due to career and reputation concerns, managers tend to avoid taking risky projects even when

the investment enhances firm value (Amihud and Lev 1981; Hirshleifer and Thakor1992; Holmstrom and

Ricart I Costa 1986; Myers 1977). Further, unlike shareholders who are able to diversify their portfolios,

managers are unable to diversify their labor income. Consequently, managers tend to be risk averse to

new projects. However, corporate risk-taking by managers is of fundamental importance because it is

directly linked to corporate and economic growth (John, Litov, and Yeung 2008). As a result, motivating

managerial risk-taking has become a key concern to academia and industry practitioners. Existing

research focuses on how to align the interests of managers with those of shareholders by using various

macroeconomic mechanisms (e.g., investor protection) and microeconomic mechanisms (e.g., equity-

based compensation) so that managers are incentivized to take sufficient risks.3

In this paper, we focus on how foreign institutional owners affect corporate risk-taking activities

around the world. As discussed in the Introduction of this paper, foreign institutional ownership is

expected to have two opposite effects on corporate risk-taking. On one hand, the following features of

foreign institutional ownership are expected to promote corporate risk-taking: (1) the relative

independence of foreign institutional owners makes them better monitors; (2) they tend to be more

3For example, researchers have analyzed how corporate risk-taking is influenced by country-level

investor protection (John, Litov, and Yeung 2008), equity-based compensation (Jensen and Meckling

1976; Haugen and Senbet 1988; Smith and Stulz1985; Coles, Daniel, and Naveen, 2006; Low 2009;

Rajgopal and Shevlin 2002; Chen, Steiner, and Whyte 2006; Hagendorff and Valascas 2011), large-

shareholder diversification (Faccio, Marchica, and Mura 2011) and ownership structure (Boubakri,

Cosset, and Saffar2008; Anderson, Mansi, and Reeb 2003).

7

diversified internationally; (3) they reduce firms’ reliance on bank financing, while banks tend to push

firms to adopt more conservative investment policies. On the other hand, compared with domestic

owners, foreign institutional investors may suffer from more severe information asymmetry with

managers, which will weaken the monitoring role of foreign institutional owners and lead to less

corporate risk-taking. Our above analysis leads to the following hypotheses regarding the net effect of

foreign institutional ownership on corporate risk-taking:

Hypothesis 1a(H1a): Foreign institutional ownership is significantly and positively associated

with corporate risk-taking.

Hypothesis 1b(H1b): Foreign institutional ownership is significantly and negatively associated


In contrast, domestic institutional owners have information advantages over foreign institutional

owners which imply that they may be better monitors and would more effectively motivate managers to

take risks. On the other hand, domestic institutional owners may be less independent than foreign

institutional owners, which imply that they may be worse monitors and would less effectively motivate

managers to take risks. Therefore, we have the following hypotheses:

Hypothesis 2a(H2a): Domestic institutional ownership is significantly and positively associated


Hypothesis 2b(H2b): Domestic institutional ownership is significantly and negatively associated


Existing research indicates that good country-level corporate governance institutional

environments such as stronger investor protection and greater information transparency promote

corporate risk-taking (John, Litov, and Yeung 2008).It would be interesting to examine how foreign

institutional ownership and country-level corporate governance interact in influencing corporate risk-

taking. Due to the information disadvantage of foreign institutional owners (Brennan and Cao 1997; Kang

and Stulz 1997; Choe, Kho, and Stulz 2005; Leuz2006; Chan, Menkveld, and Yang 2008),foreign

institutional owners’ influence on managers is expected to be stronger in more transparent institutional

8

environments where information is regularly available and investors are well protected. This implies that

foreign institutional ownership and country-level institutional environments are complementary.

On the other hand, Aggarwal, Erel, Ferreira, and Matos(2011) find that good corporate

governance practices “travel around the world” through institutional investors. Therefore, foreign

investors may substitute for poor institutional environments via the corporate governance spill-over effect

(i.e., foreign owners from countries with better governance institutions bring about more substantial

improvements in the firm-level governance of local firms in countries with worse governance institutions)

(Albuquerque, Durnev, and Koskinen 2013). For example, Rossi and Volphin (2004) find that firms

based in weak legal environments are more frequently acquired by firms from stronger legal

environments. On the other hand, in a country with good governance institutions, domestic investors are

able to advance their interests successfully and more easily influence managers to adopt riskier projects.

In such environments, the presence of foreign owners no longer exerts significant effects on corporate

risk-taking. The above analysis implies that foreign institutional ownership and country-level institutional

environments are substitutes. Our above analysis leads to the following hypotheses:

Hypothesis 3a(H3a): Foreign institutional ownership and country-level corporate governance

institutional environments are complements in motivating managers to take risks.

Hypothesis 3b(H3b): Foreign institutional ownership and country-level corporate governance

institutional environments are substitutes in motivating managers to take risks.

Although corporate risk-taking has been shown to boost corporate growth and economic growth

(John, Litov, and Yeung 2008), it may lead to an undesirable consequence: the crash risk of the firm’s

stock price may rise. This is because when risk-taking activities result in losses, managers tend to try to

hoard bad news until the negative shocks pile up to a level where managers can no longer hide them.

When negative shocks accumulate above a threshold, the firm finally releases huge amount of negative

information to the market at once, causing its stock price to drop precipitously. Managers hoard bad news

due to their earnings-based compensation contracts, as well as career and reputation concerns

(Kirschenheiter and Melumad 2002; Ball 2009). Kim, Li, and Zhang (2011b) find that the sensitivity of a

9

chief financial officer’s (CFO’s) option portfolio value to stock price is significantly and positively

associated with the firms’ future crash risk. Hutton, Marcus, and Tehranian (2009) and Kim, Li, and

Zhang (2011a) find that firms that publish more opaque financial reports and actively manage earnings or

avoid taxes tend to be more prone to stock-price crashes, confirming Jin and Myers’s (2006) conjecture.

How do foreign owners affect a firm’s crash risk? Foreign owners are generally more

independent from firm managers than domestic owners, but they also face more severe information

asymmetry. Therefore, they demand more transparency than domestic owners to alleviate information

asymmetry. Foreign institutional owners’ commitment to increasing firm transparency reduces the

opaqueness of corporate financial reporting and lowers crash risk. Further, Mieno (2009) and Gurunlu and

Gursoy (2010) provide evidence that increased foreign institutional ownership reduces firm leverage. This

is because greater foreign ownership increases the capital available to the firm (Stiglitz 2000), and firms

prefer less costly internal funding over more costly debt (Myers and Majluf 1984). Meanwhile, the

presence of foreign owners increases the firm’s credit worthiness and expands its financing channels

(Csermely and Vincze2000). Hence, foreign institutional ownership reduces a firm’s likelihood of

financial distress and bankruptcy, which in turn reduces the firm’s crash risk. Based on the above

analysis, we propose the following hypothesis:

Hypothesis (H4): Foreign institutional ownership significantly lowers a firm’s crash risk.

We further classify foreign institutional investors into foreign direct investors and foreign

portfolio investors. Foreign direct investors generally have longer investment horizons and greater

commitments than foreign portfolio investors. These longer horizons and greater commitments make

foreign direct investors more active monitors than foreign portfolio investors. We therefore propose the

following hypothesis:

Hypothesis 5(H5): Foreign direct investor ownership significantly lowers a firm’s crash risk,

while foreign portfolio investor ownership has insignificant effects on a firm’s crash risk.

3. Data

10

We obtain data from several sources. Firm ownership data are from two sources: Datastream and

FactSet Ownership (LionShares). Firm accounting data are from Worldscope. Country-level control

variables are from the World Development Indicators (WDI) compiled by the World Bank. Country-level

governance institutional environment data are from existing papers. Finally, to construct our crash-risk

measure, we collect weekly return data at both firm level and market level from Datastream.

3. 1. Sample

We construct our sample using all firms available in Datastream. Following previous studies, we

exclude firms in heavily-regulated financial and utility industries. To create our risk-taking variables, we

require that each firm have at least five consecutive years of earnings data available from Worldscope. To

preserve the consistency of our crash-risk measure, we require that there be no fewer than 26 weekly

stock returns available for a firm-year. If any variable of interest is missing for a given year, we remove

the firm-year observation. The resulting sample consists of 48,548 public firms from 72 countries

between 2000 and 2008. Table 1 reports the distribution of the sample by year, industry, and region.

[Insert Table 1 about here]

3. 2. Corporate Risk-Taking Measures

The primary measure of corporate risk-taking (RISK1) used in this study is the volatility of a

firm’s earnings (ROA) over overlapping five-year periods over the entire sample period (e.g., year 0 to

year 4; year 1 to year 5; year 2 to year 6). As alternative measures of corporate risk-taking, we also use

RISK2: company earnings range (defined as the maximum return on assets (ROA) over the overlapping

five-year window less the minimum ROA over the same period), RISK3: country-adjusted company-

earnings volatility, and RISK4: country-and-industry adjusted earnings volatility.4The firm’s earnings,

measured by the firm’s ROA, are computed by dividing earnings before interest and taxes by total assets.

4More details on the data definitions are available in the Appendix.

11

The firm’s earnings volatility is measured by the sample standard deviation of ROA.5 This measure is the

same as the measure used in previous studies (Boubakri, Cosset, and Saffar 2013; John, Litov, and Yeung

2008; Acharya, Amihud, and Litov 2011; Faccio, Marchica, and Mura 2011; Hilary and Hui 2009). We

also include a market measure of risk-taking: stock return volatility (SRVOL) over a two-year period

beginning from the current fiscal year. As an alternative measure of corporate risk-taking, we also use

research and development (R&D) expenses over total assets over overlapping five-year periods. In this

paper, we focus on the first five measures: RISK1, RISK2, RISK3, RISK4 and SRVOL because a large

proportion of sample firms report zero R&D expenses over the sample period, resulting in insufficient

variations in the R&D variable.

3. 3. Crash-Risk Measures

Following previous studies (Chen, Hong, and Stein 2001; Jin and Myers 2006; Hutton, Marcus,

and Tehranian 2009; Kim, Li, and Zhang 2011a,b; An and Zhang 2013), we construct three crash-risk

measures: NCSKEW, DUVOL, and COUNT. First, following Jin and Myers (2006), we calculate the

demeaned firm-specific continuously compounded weekly returns for firm i in week t as the

demeaned natural logarithm of one plus the residual from the expanded market model regression:

where: is the stock return for firm i in week t, is the local market return for country j in week t,

is the United States (U.S.) market return in week t, is the change in country j’s exchange rate

vs. the U.S. dollar in week t.

5We filter out extremely large outliers from the sample. After filtering out these values, 98 percent of our

sample observations remain. We then winsorize ROA at the 1 percent level on both tails of the sample

distribution when constructing our risk-taking measures.

12

The expanded market model includes two lead and lag terms to control for the non-synchronous

trading that affects both the local market returns and U.S. market returns (Dimson1979). Individual stock

returns that are not explained by the local and U.S. markets are considered firm specific and are captured

by the residual term . We use such firm-specific returns to calculate the firm-specific continuously

compounded weekly returns , which will be used to compute the three crash-risk measures below.

The first crash-risk measure, NCSKEW, is a measure of stock return asymmetry, which is the

negative skewness of the firm-specific weekly return for a given year. This measure is computed by

taking the negative of the third central moment of firm-specific weekly return scaled by the sample

variance of firm-specific weekly return raised to the power of 3/2. We follow the literature by putting a

minus sign in front of the skewness so that an increase in NCSKEW corresponds to higher crash risk(i.e.,

a more negatively skewed stock return distribution)(Chen, Hong, and Stein 2001). Specifically:

The second crash-risk measure, DUVOL, is also a measure of stock return asymmetry, computed

by taking the natural logarithm of the ratio of the standard deviation on down weeks to the standard

deviation on up weeks (Chen, Hong, and Stein2001). Specifically:

A firm-week is considered an up (down)-week if the firm-specific weekly return is above (below) the

annual mean weekly return. The convention is that a higher value of DUVOL indicates a more left-

skewed distribution, thus higher crash risk.

The third crash-risk measure, COUNT, is computed in the following manner: we first detect crash

(jump), which occurs when the firm-specific weekly return is 3.09 standard deviations below (above) its

13

mean over the fiscal year,6we then compute COUNT as the number of crashes minus the number of jumps

over the fiscal year (Hutton, Marcus, and Tehranian 2009; Kim, Li, and Zhang 2011a,b; An and Zhang

2013).

3.4. Foreign Institutional Ownership and Domestic Institutional Ownership Measures

From Datastream and FactSet (LionShares), we construct four measures of foreign institutional

ownership: (1) foreign strategic institutional ownership (FSIO); (2) foreign institutional ownership (FIO);

(3) foreign direct investment (FDI) ownership; (4) foreign portfolio investment (FPI) ownership, and one

measure of domestic institutional ownership (DIO).

Datastream provides data on strategic holdings, defined as any disclosed holdings above the 5%

threshold of the total number of shares outstanding. We extract data on the year-end foreign strategic

institutional ownership, which is the aggregate percentage of the total shares outstanding by foreign

investors with disclosed holdings greater than 5%. We also utilize these data as a proxy for the level of

FDI (i.e., the percentage of the total shares outstanding owned by foreign institutional investors who exert

control over the domestic firm). Datastream has two limitations. The first is that it does not include

institutional ownership of less than 5%. Therefore, foreign portfolio investors are not included. The

second limitation is that it does not include domestic institutional ownership data. Therefore, we cannot

use it to test H2about how domestic institutional ownership affects corporate risk-taking.

To overcome the above two limitations of Datastream, we construct foreign institutional

ownership and a proxy for FPI from FactSet Ownership database which provides fund holdings

information for various funds such as mutual funds and pension funds. In contrast with Datastream,

FactSet includes ownership above, below or equal to 5% of the total shares.7Foreign institutional

6We follow Hutton, Marcus, and Tehranian (2009) and choose 3.09 to generate top and bottom 0.1

percent in the normal distribution.

7The fund positions are gathered globally from mutual fund reports, regulatory authorities, mutual fund

associations and fund management companies. Due to international differences in the regulatory

14

ownership is derived by summing up the percentage ownership by each unique fund domiciled in a

foreign country. Further, we single out foreign institutional investors with percentage ownership of less

than 5 percent, and aggregate their percentage ownership to compute FPI ownership. FPI thus represents

foreign institutional investors who gain ownership of domestic firms without gaining control.

Domestic institutional ownership is included in our models so that we are able to contrast the

effects of foreign institutional ownership and domestic institutional ownership on corporate risk-taking

and crash risk. We construct domestic institutional ownership from FactSet Ownership database.

Specifically, we aggregate the equity holdings of domestic institutions as a percentage of the total shares

outstanding at the end of the previous year. The domestic ownership used for each of the overlapping

five-year windows is again the year-end value of the first year within the window.

3.5. Country-level Corporate Governance Institution Measures

We use two groups of country-level corporate governance institution measures: (1) Information

transparency measures from La Porta, Lopez-de-Silanes, and Shleifer(2006), Djankov, McLiesh, and

Shleifer (2007), and Bushman, Piotroski, and Smith (2004); (2) Judicial system efficiency and investor

protection measures from La Porta, Lopez-de-Silanes, Shleifer, and Vishny(1998) and Djankov, La Porta,

Lopez-de-Silanes, and Shleifer(2008). We expect that managers are monitored more intensely in countries

where information is more transparent, the judicial system is more efficient, and investors are better

protected. As a result, we expect firms to take more risks in countries with greater information

transparency, a more efficient judicial system and better investor protection.

requirements and disclosure conventions, the reporting frequency of fund positions varies from a monthly

basis to an annual basis, therefore we scale the last reported holdings in a year by the total shares

outstanding at the end of the previous year to construct a percentage ownership. Using an alternative

definition, such as the year end shares outstanding, we find that our results are insensitive to the choice of

scale.

15

3.6. Control Variables

We include both firm-level and country-level controls in our risk-taking regressions. Standard

controls that have been shown previously to significantly influence corporate risk-taking are included.

Specifically, we include the following firm-level controls: (1) the annual growth rate of total sales

(SALESGROWTH)(to control for the effect of firm growth opportunities on risk-taking) (John, Litov,

and Yeung 2008); (2) firm size (SIZE)(i.e., the natural logarithm of total sales in millions of U.S. dollars);

(3) firm profitability (ROA)(i.e., earnings before interests and taxes scaled by total assets); (4) leverage

(LEVERAGE)(i.e., net debt over assets); (5) capital expenditure (CAPEX) (i.e., capital expenditure over

total assets). We include the following country-level controls: (1) growth in real gross domestic product

(GDPGROWTH); (2) market interest rates (MARKETRATES); (3) economic freedom

(ECONFREEDOM) (greater economic freedom provides investors and managers with more incentives to

implement riskier innovative projects) (Gwartney, Lawson, and Norton 2008); (4) economic development

(GDP, measured by the natural logarithm of GDP per capita). In addition to these variables, we include

year, industry and country dummy variables. To reduce the effects of outliers, we winsorize all firm-level

variables at 1% at both tails of the distribution.

In crash-risk regressions, we include firm-level control variables that have been shown to

significantly affect crash risks in Chen, Hong, and Stein (2001), Hutton, Marcus, and Tehranian (2009),

and Kim, Li, and Zhang (2011a,b). We also include country-level control variables that potentially affect

crash risks. Specifically, we include the following firm-level control variables: (1) de-trended average

monthly stock turnover (DTURN)(i.e., the average monthly turnover minus the average monthly turnover

from the previous year)—we expect stocks with higher turnover to be more negatively skewed; (2) the

standard deviation of firm-specific return (SIGMA), derived from the expanded market model (equation

1)(higher volatility of firm-specific return is expected to increase the crash risk of the firm); (3) average

firm-specific weekly return (RET)(lower average firm-specific return indicates that a firm is more likely

to experience more down weeks than up weeks, which should lead to more crash risk); (4) lagged three-

year moving sum of the absolute value of discretionary accruals (OPACITY)(this is a measure of accrual

16

manipulation and proxies for the ability of a manager to hide adverse information from the financial

markets. This should increase a firm’s crash risk, because when the threshold of bad news that the firm

can sustain is higher, more extreme crashes are more likely); (5) leverage (LEVERAGE)(i.e., net debt

over total assets. Higher leverage increases the probability of bankruptcy and crash risk); (6) corporate

risk-taking (RISK1)(the volatility of a firm’s earnings over overlapping five-year periods during the entire

sample period. If a firm takes more risks, then the firm could incur more losses. Hiding such losses from

the market could contribute to a firm’s crash risk); (7)profitability (ROA)(i.e., earnings before interest and

tax over total assets—high profitability is expected to be associated with more stability and lower crash

risk);(8) market-to-book ratio (MTB) (a higher market-to-book ratio has previously been shown to be

associated with higher distress risks (Griffin and Lemmon 2002), which are expected to lead to higher

crash risk); (9) the natural logarithm of the market value of equity in US dollar (SIZE); (10) lagged crash

risk (lagged NCSKEW) (Kim, Li, and Zhang 2011a,b).

In crash-risk regressions, we also include the following country-level controls:(1) growth in real

Gross Domestic Product (GDPGROWTH); (2) market interest rates (MARKETRATES);(3) economic

development (GDP)(i.e., the natural logarithm of GDP per capita). In addition to these variables, we

include year and industry dummy variables. We winsorize all the variables at 1% level at both tails of the

distribution, to reduce the effects of outliers. Table 2 reports the summary statistics of the variables used

in this paper.


4. Empirical Models

To test H1, we estimate the following regression equation:

where “risk-taking” is measured by five measures of corporate risk-taking, FSIO is the percentage of

shares held by foreign strategic institutional owners. CONTROLS denotes the set of control variables,

17

and denotes the error term. Subscript “i” indexes firm and subscript “t” indexes year. Robust standard

errors clustered at both firm and country levels are estimated. If is positive and significant, then H1a is

supported (i.e., foreign institutional owners promote corporate risk-taking). If is negative and

significant, then H1b is supported (i.e., foreign institutional owners lower corporate risk-taking).

To further test H1, we estimate the following regression equation:

where FIO is foreign institutional ownership, as an alternative measure to FSIO, and DIO is domestic

institutional ownership. If is positive and significant, then H1a is supported (i.e., domestic institutional

owners promote corporate risk-taking). If is negative and significant, then H1b is supported (i.e.,

domestic institutional owners lower corporate risk-taking).

To test H3 regarding the interactions between foreign institutional ownership and

country-level corporate governance institutions, we estimate the following regression equation:

Where GI is country-level governance institutions, including information transparency (IT) and institutional

environment (IE). If and are positive, then H3a is supported (i.e., foreign institutional ownership and

country-level governance institutions are complements). If and are negative, then H3b is supported

(i.e., foreign institutional ownership and country-level governance institutions are substitutes).

The crash-risk regression model is as follows:

where CRASH is our crash-risk measures (NCSKEW, DUVOL, and COUNT). FSIO is foreign strategic

institutional ownership, RISK1 is our primary measure of corporate risk-taking, CONTROLS is a vector

18

of control variables (firm characteristics and country characteristics).Subscript “i” indexes firm and

subscript “t” indexes year. Robust standard errors clustered at both firm and country levels are estimated.

If is negative and significant, then H4 is supported (i.e., foreign institutional owners reduce crash risk).

To test H5, we estimate the following regression model:

where CRASH is our crash-risk measures, FDI, FPI, and DIO are foreign direct investors’ ownership,

foreign portfolio investors’ ownership, and domestic institutional ownership in a firm, respectively.

RISK1 is our primary measure of corporate risk-taking, CONTROLS is a vector of control variables (firm

characteristics and country characteristics). If is negative and significant, and is insignificant, then

H5 is supported (i.e., foreign direct investors’ ownership significantly reduces crash risk while foreign

portfolio investors’ ownership has an insignificant effect on crash risk).

5. The Effects of Foreign Institutional Ownership on Corporate Risk-Taking

5.1. The Effects of Foreign Institutional Ownership and Domestic Institutional Ownership on

Corporate Risk-Taking

Table 3 reports the estimation results when we use foreign strategic institutional ownership as our

measure of foreign institutional ownership in equation (2). We find that foreign strategic institutional

ownership is positively and significantly associated with all four corporate risk-taking measures.

Regressions of all five risk-taking measures on foreign strategic institutional ownership show consistent

results that foreign institutional owners promote corporate risk-taking.


Table 4 presents our results using the FactSet Ownership database. When we include both foreign

and domestic institutional ownership in our model, we find that foreign and domestic institutional

19

ownership enter the regression positively and negatively, respectively, both at the 1% significance level.

These results strongly support hypotheses 1a and 2b.


We further classify foreign institutional ownership into foreign direct investment (FDI)

ownership and foreign portfolio investment (FPI) ownership and examine their respective effects on

corporate risk-taking. Table 5 reports that both FDI ownership and FPI ownership significantly promote

corporate risk-taking. In contrast, domestic institutional ownership significantly lowers corporate risk-

taking. These results support hypotheses 1a and 2b.


5.2. The Joint Effects of Foreign Institutional Ownership and Country-level Corporate Governance

Institutions on Corporate Risk-Taking

To examine whether foreign institutional owners and country-level corporate governance

institutions complement or substitute for each other, we use various country-level governance institution

measures used in previous studies. We first examine the following six information-transparency

measures: (1) AUDIT, measuring the credibility of financial accounting disclosure (an index ranging from

1 to 4 depending on the percentage of the firms in the country audited by the Big 5 Accounting Firms);

(2) ANALYST, the number of analysts following the largest 30 companies in the country in 1996; (3)

ITENF, measuring insider trading regulation, a dummy variable that equals one if a country has enforced

insider trading laws before 1995 (measures (1), (2) and (3) are from Bushman, Piotroski, and Smith

2004); (4) INFSHA03 (“information sharing dummy” from Djankov, McLiesh, and Shleifer 2007), a

dummy variable that indicates whether a public or a private credit bureau operates in the country in 2003;

(5) DISREQ, the disclosure requirement index; (6) LIASTA, the liability standard index (measures (5)

and (6) are from La Porta, Lopez-de-Silanes, and Shleifer 2006) Each of the measures we have used is

designed in such a way that higher scores reflect greater country-level information transparency.

For brevity, in Table 6, we report results only when we measure corporate risk-taking by RISK1,

but the results are qualitatively similar when we use other corporate risk-taking measures. For almost all

20

regressions in Table 6, we find that both FDI and FPI are positively and significantly related to corporate

risk-taking. The interactions between foreign institutional ownership and country-level information-

transparency measures are negative and significant. Evidence from these regressions supports H3b that

foreign institutional ownership and country-level governance institutions are substitutes.


In addition to country-level information-transparency measures, we also use the following three

country-level legal-environment and investor-protection measures: (1) EFFJUD, a measure of the

efficiency of judicial system (La Porta, Lopez-de-Silanes, Shleifer, and Vishny 1998); (2) LEGCOM, a

dummy variable for English common law origin (La Porta, Lopez-de-Silanes, Shleifer, and Vishny (1998)

find that countries with English common law origin protect investors more effectively than countries with

civil law origin); (3) ANTISELF, the anti-self-dealing index (Djankov, La Porta, Lopez-de-Silanes, and

Shleifer 2008), which ranges from zero to one and captures the ex-ante and ex-post private control of self-

dealing. The higher the values of the three variables, the more efficient is the judicial system, and the

more effectively investors are protected. In almost all regressions in Table 7, we find that the interactions

between foreign institutional ownership and country-level legal enforcement and investor protection

measures are negative and significant. Evidence from these regressions again supports H3b that foreign

institutional ownership and country-level governance institutions are substitutes.


Following our above findings we separate foreign institutional holdings into four different

groups. The groups are categorized according to the development (developed/developing) of the country

of the foreign institutional investor and firm in which they invest in8. This is beneficial because it

provides us with a more aggregate position on governance. After categorizing the institutional holdings

we run a separate regression for each group to determine the effect of foreign institutional investor

ownership under each classification.

8 We classify the development based on the classification provided by the International Monetary Fund

(IMF).

21

The results are reported in Table 8. We find that the risk-taking effects from foreign institutional

investors are almost entirely driven by the groups: (1) Developed to Developed, and (2) Developed to

Developing, since foreign institutional ownership does not enter significantly otherwise. This provides

some evidence for our hypothesis regarding the substitution effect between foreign institutional investors

and corporate governance. This is because we expect developed countries to have a more highly

developed economy and subsequently higher standards of country-level governance.


In summary, we find that there is evidence for a substitution effect between foreign institutional

ownership and corporate governance. This suggests that foreign institutional investors are able to

motivate risk-taking by motivating domestic firms to undertake improvements in corporate governance.

5.3. The Effect of Foreign Institutional Ownership from Different Institution Types on Corporate

Risk-Taking

Prior literature suggests that investor behavior depends on both nationality and institution type

due to differences in preferences and potential business ties (Brickley, Lease, and Smith 1988; Almazan,

Hartzell, and Starks 2005; Chen, Li, and Harford 2006; Ferreira and Matos 2008). While different

institution types have different preferences, the presence of business ties between institutional investors

and firm can also affect both their ability to be active in monitoring managers’ decisions and their ability

to act as independent shareholders (Ferreira and Matos 2008). Different institution types are also exposed

to distinct investment mandates and regulations. Therefore, we cannot expect all institutions to be equally

equipped or motivated to be active monitors.

Following Ferreira and Matos (2008) we classify institutions into seven different categories: (1)

Banks; (2) Insurance Companies; (3) Investment Companies (typically Mutual Fund Management

Companies); (4) Investment Advisors; (5) Pension Funds and Endowments; (6) Hedge Funds and Venture

Capital; (7) Government9. We then further classify these categories into two groups, namely, independent

and grey institutions based on both the institutions preferences and potential for business ties to a firm.

9 Note that Government and Endowment holdings does not enter into our dataset.

22

We characterize independent institutions as institutions that are subject to fewer regulatory restrictions

and potential business relationship, on the other hand grey institutions tend to have higher monitoring

costs and are more loyal to corporate management. In our study, we classify Investment Companies and

Investment Advisors as independent, meanwhile Banks and Insurance Companies as the typical grey

institutions (Brickley, Lease, and Smith 1988; Almazan, Hartzell, and Starks 2005; Chen, Li, and Harford

2006; Ferreira and Matos 2008). For example, Brickley, Lease, and Smith (1988) finds that investment

advisers and mutual funds tend to be very active monitors, while banks and insurance companies are more

supportive of management actions.

Although many also classify Pension and Hedge funds as grey institutions there are

disagreements amongst researchers on whether they can actually be considered grey. In particular, since

the 1980s we have seen an increase in the involvement of public pension funds in regards to submitting

shareholder proposals, pressuring management for corporate reforms, and using the press to target the

management and boards of poorly governed or performing companies (Gillian and Starks 2007).

Similarly, hedge funds have also seen a rise in activism since the beginning of the 21st century and have

become particularly important in their ability as monitors of corporate performance and agents of change

(Gillian and Starks 2007). Yet few empirical studies have been able to show that this has impacted the

value of the target firms especially in the long-term operating or stock-market performance (Barber 2006;

DelGuercio and Hawkins 1999). Several studies do attempt to show that the firms do end up

implementing governance reforms or even significant changes in business activities after activism by

Pension and Hedge funds (Wahal 1996; Carleton, Nelson, and Weisbach 1998; Gillian and Starks 2007;

Ertimur, Ferri, and Stubben 2009), however it is difficult to establish a causal relationship between these

changes and activism. Consequently, there is no solid empirical evidence to classify Pension or Hedge

funds as either independent or grey.

In order to determine institution type effect we divide both foreign and domestic institutional

ownership into the seven categories. We then report separate regressions for each of the categories.

Following our discussion we anticipate that more independent institutions are less likely to have business

23

relations with the firm, they will subsequently be less compelled to be loyal to management and will take

a more active stance when it comes to monitoring and firm management.

The results are reported in Table 9. We find that foreign institutional ownership is positively and

significantly associated with risk-taking for the categories (3) Investment Companies, (4) Investment

Advisors, and (5) Pension Funds & Endowments, each at the 1% significance level. On the other hand,

(1) Banks, (2) Insurance Companies, and (6) Hedge Funds & Venture Capital are insignificant at all

conventional significance levels. Based on our discussion we can see that only independent institutions

types (Investment Companies and Investment Advisors) have a significant effect on risk-taking for both

foreign and domestic ownership. On the other hand for all typical grey institution types (Banks and

Insurance Companies) both foreign and domestic ownership has no significant effect on risk-taking. This

implies that only institutions with the ability to actively monitor have the capacity to influence firms’

risk-taking characteristics. Even though this result may seem straightforward it contributes to our study in

number of ways. First, this reinforces the reliability of our results since it confirms the results from past

studies. Second, this result indicates that our main findings are only driven by institution types with the

ability or incentive to actively monitor, since they are able to effectively impose their risk-taking

preferences onto the firm they invest in. Lastly, this shows foreign institutional owners have less business

ties with the firms they invest in therefore they are able to motivate risk-taking through active and

effective monitoring of managerial behavior.


5.4. Endogeneity

Our evidence indicates that foreign institutional ownership is positively and significantly related

to corporate risk-taking. However, it is possible that foreign institutional investors are attracted to firms

that take more risks (reverse causality). It is also possible that a third factor affects both foreign

institutional ownership and corporate risk-taking. For example, if a firm has highly effective corporate-

governance mechanisms, it may attract more foreign institutional owners (Leuz, Lins, and Warnock

24

2009). Meanwhile, such effective corporate-governance mechanisms may motivate managers to take

sufficient risks.

To tackle endogeneity, we adopt four different approaches. The first three approaches are

routinely used to account for endogeneity and simply make use of our previously described ownership

dataset. The results are reported for our primary risk-taking measure RISK1 in section 5.3.1.

Alternatively, in section 5.3.2 we employ an event study approach that allows us to both

effectively account for the endogeneity problem while emphasizing the causal relationship between

foreign ownership and risk-taking. This approach incorporates data sources other than our primary

ownership dataset. The results are qualitatively similar when we use alternative risk-taking measures.

5.4.1. Regression Based Approaches

In this section we carefully address endogeneity concerns by using three different approaches.

The first approach is to lag every independent variable to mitigate the concerns about reverse causality.

More specifically, it is difficult to argue that the foreign owners are able to accurately predict the future

risk-taking activities of a firm and as a result invest in it. The results of Model 1 in Table 10 indicate that

even after lagging all the independent variables, our key findings remain the same: FDI and FPI both

display a positive and significant effect on risk-taking, while domestic institutional ownership has a

negative and significant effect.

The second approach is to take the first difference of each variable in the regressions and then run

regressions with these differenced variables. Additionally, we lag every differenced independent variable

to mitigate reverse causality. By taking first difference, we remove any time-invariant firm characteristics

that could be driving the relationship between foreign institutional ownership and corporate risk-taking

(e.g., firm-specific corporate governance mechanisms that are generally persistent over time). Model 2 of

Table 10 reports the results of regressions with first-differenced variables. Our key results still hold that

foreign institutional ownership promotes corporate risk-taking.

The third approach is 2SLS regressions. We use political rights index (POLRIGHTS) from

Freedom House as an instrument for FDI ownership. Boubakri, Cosset, and Saffar (2008) posit that

25

foreign owners are more likely to invest in countries with stronger political institutions. Meanwhile,

political rights do not have direct unambiguous effects on corporate risk-taking. Therefore, political rights

satisfy both the relevance and exogeneity conditions for a valid instrumental variable. We report the

second-stage regression results in Model 3 of Table 10. We find our key results still hold that foreign

institutional ownership is positively and significantly related to corporate risk-taking.


5.4.2. Event Study Approach

For our final approach we must first establish that one of the main channels allowing foreign

investors to access domestic firms is through M&A transactions. By identifying this link it now becomes

apparent to employ an event study framework to illustrate the true causal relationship between foreign

ownership and risk-taking. Consequently, we are required to supplement our ownership dataset with

additional data sources for the announcements of these M&A transactions, namely, the Securities Data

Company (SDC) Platinum Mergers and Acquisitions Database.

We construct our initial sample by specifying that these firms have to be readily available in both

Datastream and SDC. These firms also have to have been the target of large block purchases (>5% of

issued shares reported by SDC) by foreign investor during our sample period since it is closest in nature

to our definition of FDI. Finally, to assure that the event windows are independent of each other we limit

our sample to firms that have only been the target of a foreign large block purchase no more than once.

Using the above criterion we find only 255 event firms in our sample in which foreign ownership

increases due to large block purchases. The announcement dates allows us to confirm whether these

sample firms reporting large foreign shareholders only exhibit risk-taking changes after these

transactions.

We calculate our risk-taking measure (ROA volatility) for both pre- and post-event periods of up

to 5 years. More specifically, for the pre-event period we compute the ROA volatility using ROA values

ranging from year -5, -4, and -3 up until year -1 before the announcement date. Similarly, for the post-

event period we compute the ROA volatility using ROA values from year +1 to year +3, +4, and +5 after

26

the announcement date. These risk-taking measures are computed for each event firm and corresponding

matched control firm. The control firm set contains firms that do not report any foreign block purchases.

For this study, we form a set of 1-to-1 matched firms by matching each event firm to a control firm that

belongs in the same market, industry and closest market capitalization, however if not possible, a firm in

the same market with the closest market capitalization.

The result of this study is reported in Table 11. The results using our aggregate sample presented

in Panel A clearly indicates that there is indeed an increase in risk-taking after large block purchases by

foreign investors. In fact, prior to the transaction the event firm has a similar risk-taking characteristic to

that of the control firm. However, after the announcement of these transactions our risk-taking measure is

significantly higher than that of the control set. This implies that the increase in risk-taking of the event

firm is driven by large foreign block purchases rather than as a result of foreign investors being attracted

to riskier firms. Furthermore, given the event study design, the only difference between the event and

control set is in the presence of a large foreign block purchase. This eliminates the possibility of a third

factor skewing our results. In Panel B, we divide our aggregate sample into different groups based on the

development indicator (developed/developing) of the acquirer and target nation involved in the

transaction. Again, even in these groups we find that the risk-taking measure between the event and

control firms prior to the announcement of the transaction is never significantly different, with significant

differences only occurring after the announcement of these transactions. This provides further evidence

against reverse causality. Furthermore, we find that the results from Panel A are mainly driven by

transactions between the same development groups. However, the fact that there is a significantly positive

difference in the group with transactions from developed to developing countries and none when reversed

again lends support to our previous findings that portray foreign ownership as a substitute for corporate

governance.


In summary, we find that foreign institutional ownership continues to have a positive and

significant effect on corporate risk-taking even after addressing endogeneity. It is therefore unlikely that

27

our main findings are driven by endogeneity. For this reason, foreign ownership does appear to bring

important risk-taking characteristics to domestic firms.

6. The Effects of Foreign Institutional Ownership on Crash Risk

Our results in the previous section show that foreign institutional ownership significantly

increases corporate risk-taking. Does it also lead to greater crash risk for the firm? To answer this

question, we estimate OLS (Ordinary Least Square) regression equation (4). Robust standard errors

clustered at both country and firm levels are estimated, and corresponding p-values are reported in the

tables.

6.1. The Effects of Foreign Strategic Institutional Ownership on Crash Risk

We first examine how foreign strategic institutional ownership affects a firm’s crash risk, where

foreign strategic institutional ownership is defined as the total ownership by foreign owners with more

than 5% ownership in the firm. In Models 1 and 2 of Table 12, we use NCSKEW as the dependent

variable and include lagged foreign strategic institutional ownership as the key independent variable. We

also include lagged control variables that are commonly used in existing crash-risk literature (Chen,

Hong, and Stein 2001; Hutton, Marcus, and Tehranian 2009; Kim, Li, and Zhang 2011a,b; An and Zhang

2013). Meanwhile, we control for industry and year fixed effects. We find that the proportion of foreign

strategic institutional ownership is negatively and significantly related to a firm’s crash risk. In Models 3

and 4, we use DUVOL as an alternative measure of a firm’s crash risk, and the results are qualitatively

similar. In Models 5 and 6, we use COUNT as the third measure of a firm’s crash risk, and we obtain

qualitatively consistent results in Model 6 as in Models 2 and 4. These results strongly support H4.


6.2. The Effects of Foreign Direct Investor (FDI) Ownership, Foreign Portfolio Investor (FPI)

Ownership and Domestic Institutional Ownership on Crash Risk

In this section, we examine the effects of FDI, FPI and domestic institutional ownership on crash

risk, respectively. The proxy we use for FDI is the foreign strategic institutional ownership from

Datastream. As a proxy for FPI ownership, we aggregate all foreign mutual fund holdings below 5% to

28

represent small foreign portfolio owners who are interested only in the firm for the purpose of

international diversification.

In Model 1 of Table 13, we find that while FDI ownership is negatively and significantly related

to a firm’s crash risk, FPI ownership is insignificantly related to it. In Model 2, we include the

interactions between risk-taking and various types of institutional ownership (FDI, FPI and domestic) to

investigate how a firm’s ownership structure affects the effect of corporate risk-taking on crash risk. We

find that while domestic institutional ownership significantly increases a firm’s crash risk when the firm

takes more risks, neither FDI nor FPI ownership significantly affects the relationship between corporate

risk-taking and crash risk. In Models 3 and 4, we use DUVOL as an alternative measure of crash risk, and

get qualitatively similar results as in Models 1 and 2. In Models 5 and 6, we use COUNT as the third

measure of crash risk, and we get qualitatively similar results in Model 6 as in Models 2 and 4. These

results strongly support H5.


In summary, the results of this section show that a firm’s ownership structure plays a significant

role as a determinant of the firm’s crash risk. Specifically, our results show that FDI ownership

significantly decreases crash risk, while domestic institutional ownership significantly increases crash

risk. Moreover, domestic institutional ownership significantly increases a firm’s crash risk when it takes

more risks, but foreign institutional ownership does not have a significant impact on the relationship

between a firm’s risk-taking and its crash risk.

6.3. Endogeneity

The negative relationship between foreign institutional ownership and corporate crash risk may

have an alternative interpretation: foreign institutional investors choose to invest in firms with greater

information transparency, which tend to have lower crash risk. To rule out the possibility of reverse

causality, we regress the first-differenced crash risk on the first-differenced foreign institutional

ownership. We also use the 2SLS method, with foreign sales over total assets as the instrumental variable

29

for FDI ownership. We do not lag all the independent variables as we do in Section 5 because our

baseline specification already lags each independent variable. Table 14 reports the results.


We first run a regression of first-differenced crash risk on first-differenced foreign institutional

ownership. By examining the change instead of the level, we remove any time-invariant firm

characteristics that could be driving the relationship between ownership structure and crash risk. Model 1

of Table 14 shows that the change in FDI ownership is negatively and significantly associated with the

change in a firm’s crash risk. In Model 2, we use foreign sales scaled by total assets (FSALES) as an

instrumental variable because while it is expected to directly influence FDI ownership, it is difficult to

argue that FSALES has direct effects on a firm’s crash risk.10

This instrument is highly correlated with

foreign institutional ownership because foreign institutional investors tend to prefer investing in firms

with more international exposure, which will consequently lead to more visibility to foreign investors

(Covrig, Lau, and Ng 2006; Ferreira and Matos 2008).

In Model 2, we report the second-stage regression results, which indicate that the coefficient of

FDI ownership is still significantly negative, while the coefficient of domestic institutional ownership

remains significantly positive. Interestingly, after we address the endogeneity issue, FPI ownership begins

to show positive and significant effects on crash risk, again highlighting the different effects of FDI and

FPI ownership on crash risk. For brevity, we have reported the results only when we use NCSKEW as the

dependent variable. When we use DUVOL and COUNT as the dependent variables, we obtain

qualitatively similar results, which are available upon request.

7. Conclusion

10We use FSALES as opposed to POLRIGHTS since FSALES appears to be much more relevant as an

instrument. However, we disregard FSALES as an instrument for our previous 2SLS analysis since it is

very likely that FSALES is endogenous there, given that the measure of corporate risk-taking is the

volatility of earnings.

30

We study a large sample of 48,548 firms in 72 countries from 2000 to 2008 to investigate the

effect of ownership structure on corporate risk-taking and crash risk, with a focus on how foreign

institutional ownership affects corporate risk-taking and crash risk. Motivating corporate risk-taking is

pivotal because corporate risk-taking is essential to corporate growth and economic growth (John, Litov,

and Yeung 2008; Baumol, Litan, and Schramm 2009). However, corporate risk-taking can also be

detrimental to a firm, because more risk-taking is associated with potentially more severe losses, and

severe losses create incentives for managers to manage earnings and hide bad news from the market.

When bad news eventually accumulates beyond a threshold, all the negative information will be released

at once, leading to an excessive fall in the firm’s stock price. Therefore, more corporate risk-taking may

lead to greater crash risk. We evaluate the effects of two different types of foreign institutional ownership:

foreign direct investor (FDI) ownership and foreign portfolio investor (FPI) ownership on corporate risk-

taking and crash risk.

We have made several discoveries. First, we find that both foreign direct investor ownership and

foreign portfolio investor ownership in a firm significantly increases the firm’s risk-taking. We believe

that foreign institutional investors monitor the managers more intensely due to their lack of extensive

existing business ties with the managers. As a result, foreign institutional investors are able to motivate

managers more effectively to take risks. On the other hand, compared with domestic institutional

investors, foreign institutional investors generally face more information asymmetry with the managers.

As a result, they may be less effective monitors, and may lead to significantly lower corporate risk-taking.

Our results show that the positive effects of foreign institutional ownership on corporate risk-taking

dominate the negative effects.

Second, we examine whether foreign institutional ownership and country-level corporate

governance institutions are substitutes or complements. We find that the relationship between foreign

institutional ownership and corporate risk-taking is stronger in countries with poorer governance

institutions. This supports the view that foreign institutional owners play stronger roles in motivating

31

managers to take risks in countries with weaker governance institutions as domestic institutional investors

there are unable to incentivize managers effectively to take risks.

Third, we examine the effects of FDI, FPI and domestic institutional ownership on a firms’ crash

risk. We find that among these three types of owners, only FDI ownership significantly decreases a firms’

crash risk. Similar to FPI owners, FDI owners are more independent of management compared with

domestic institutional owners. As a result, FDI owners have stronger incentives to improve corporate

governance. Different from FPI owners, FDI owners have control over corporate operations. Since FDI

owners have both the incentives to improve firm-level corporate governance and the controlling power to

influence corporate operations, they can effectively reduce the earnings management activities of

managers, which consequently decrease the crash risk of a firm.

Finally, we find that when firms take more risks, domestic institutional ownership significantly

increases the crash risk, while foreign institutional ownership does not.

Our findings are robust to the inclusion of industry and year fixed effects, the use of alternative

measures of corporate risk-taking and crash risk, lagging all the independent variables, taking the first

differences, and 2SLS estimation techniques.

Our findings highlight the benefits associated with foreign institutional investors. These findings

have broad implications for academia, practitioners and policy makers. When considering the crash risk

of a firm, researchers should consider the ownership identity of its institutional investors since it has a

strong effect on the firm’s crash risk. When policy makers reduce the barriers to foreign investments in

the hope of developing the local markets, they should consider the costs and benefits associated with

foreign investment. For example, based on our findings, foreign investors are particularly effective at

motivating corporate risk-taking in countries with poor governance environments. This provides a new

channel through which foreign investment can promote economic growth in developing countries.

However, attracting FPI investors as opposed to FDI investors could significantly increase the crash risk

of local firms’ stock prices.

32

Table 1. Sample Distribution

This table reports the distribution of our sample of 48,548 firms by year, industry and region.

Panel A: Distribution by year

Year Number Percentage

2000 5463 2.34

2001 8936 3.83

2002 28665 12.28

2003 29166 12.49

2004 31073 13.31

2005 31012 13.28

2006 31941 13.68

2007 33542 14.36

2008 33701 14.43

Total 233499 100.00

Panel B: Distribution by industry

Industry Number Percentage

Basic Materials 5703 11.75

Consumer Goods 7305 15.05

Consumer Services 6969 14.35

Health Care 4573 9.42

Industrials 12413 25.57

Oil & Gas 3252 6.7

Technology 7279 14.99

Telecommunications 945 1.95

Other 109 0.22

Total 48548 100

Panel C: Distribution by region

Region (Countries) Number Percentage

East Asia and the Pacific (14) 17795 36.65

Europe and Central Asia (28) 10547 21.72

Latin America and the Caribbean (8) 717 1.48

Middle East and North Africa (8) 475 0.98

North America (2) 16934 34.88

South Asia (4) 1469 3.03

Sub-Saharan Africa (8) 611 1.26

Total (72) 48548 100

33

Table 2. Summary Statistics

This table reports the summary statistics for the variables used in this paper. The full sample contains 48,548 firms from 72 countries. Panel A reports the summary statistics for variables in our corporate risk-taking regressions. Panel B reports the summary statistics for variables in our crash-risk regressions. The definitions and data sources for the variables are available in the Appendix. Panel A: Corporate risk-taking regression variables

Mean Median S.D. Minimum Maximum

RISK1 0.076 0.041 0.091 0.004 0.462

RISK2 0.185 0.100 0.221 0.009 1.114

RISK3 0.076 0.041 0.089 0.006 0.459

RISK4 0.075 0.043 0.086 0.004 0.449

R&D 0.059 0.020 0.103 0.000 0.632

FSIO 0.061 0.000 0.160 0.000 1.000

FIO 0.05 0.019 0.077 0.000 0.416

FDI 0.061 0.000 0.160 0.000 1.000

FPI 0.047 0.020 0.067 0.000 0.350

DINST 0.115 0.024 0.176 0.000 0.777

ROA 0.006 0.046 0.197 -1.032 0.355

LEVERAGE 0.217 0.185 0.202 0.000 1.004

SIZE 13.926 13.881 3.372 5.352 21.793

SALESGROWTH 0.252 0.091 0.830 -0.752 6.349

CAPEX 0.055 0.035 0.063 0.000 0.362

GDPGROWTH 3.189 2.660 2.627 -1.04 12.700

ECONFREEDOM 7.879 8.020 0.686 6.060 9.110

GDP 10.08 10.485 1.003 6.681 10.922

MARKETRATES (%) 3.389 3.000 2.325 -3.280 11.710

34

Panel B: Stock price crash-risk regression variables

Mean Median S. D. Minimum Maximum

NCSKEW -0.061 -0.073 0.685 -1.989 2.194

DUVOL -0.04 -0.045 0.339 -0.875 0.884

FSIO 0.046 0.000 0.139 0.000 0.990

FIO 0.047 0.019 0.071 0.000 0.414

FDI 0.046 0.000 0.139 0.000 0.990

FPI 0.044 0.020 0.062 0.000 0.348

DINST 0.134 0.038 0.184 0.000 0.775

ROA -0.002 0.037 0.192 -1.028 0.355

LEVERAGE 0.219 0.184 0.205 0.000 1.000

MV 5.054 4.938 1.973 0.820 10.232

MTB 2.177 1.460 3.217 -8.160 20.510

RISK1 0.071 0.036 0.088 0.003 0.442

OPACITY 0.762 0.245 1.384 0.026 8.510

SIGMA 0.060 0.053 0.030 0.016 0.155

RET -0.218 -0.138 0.224 -1.163 -0.012

DTURN 0.000 0.000 0.007 -0.038 0.024

GDPGROWTH 3.134 2.550 2.783 -4.240 12.700

GDP 10.101 10.491 0.976 6.681 10.922

MARKETRATES (%) 3.217 3.000 2.163 -3.280 11.710

35

Table 3. Foreign Strategic Institutional Ownership and Corporate Risk-Taking

This table reports the OLS estimation results of the following regression model:

where “Risk-taking” refers to five risk-taking measures, FSIO is the percentage of strategic ownership

held by foreign institutional investors and CONTROLS is a vector of control variables (firm and country

characteristics). All regressions includes year, industry and country fixed effects whose coefficients are

suppressed. Beneath each coefficient estimate is p-statistics in parentheses based on robust standard errors

clustered at both the firm- and country-levels. ***, **, and * denote statistical significance at the 1%, 5%,

and 10% levels, respectively. Variable definitions and data sources are available in the Appendix.

Dependent Variable RISK1 RISK2 RISK3 RISK4 SRVOL

Model (1) (2) (3) (4) (5)

FSIO 0.010** 0.025** 0.009** 0.009** 0.029***

(0.020) (0.018) (0.012) (0.034) (0.000)

ROA -0.192*** -0.461*** -0.184*** -0.171*** -0.114***

(0.000) (0.000) (0.000) (0.000) (0.000)

LEVERAGE 0.018*** 0.045*** 0.018*** 0.018*** 0.079***

(0.000) (0.000) (0.000) (0.000) (0.000)

SIZE -0.010*** -0.024*** -0.010*** -0.010*** -0.014***

(0.000) (0.000) (0.000) (0.000) (0.000)

SALESGROWTH 0.005*** 0.012*** 0.005*** 0.005*** 0.004***

(0.000) (0.000) (0.000) (0.000) (0.000)

CAPEX -0.012* -0.030* -0.014** -0.018** -0.032***

(0.092) (0.081) (0.021) (0.010) (0.007)

GDPGROWTH -0.007*** -0.019*** -0.009*** -0.009*** 0.001

(0.009) (0.008) (0.004) (0.003) (0.617)

ECONFREEDOM 0.001** 0.003** 0.001* 0.001 0.013*

(0.022) (0.027) (0.070) (0.136) (0.088)

GDP -0.015** -0.035* -0.020*** -0.017** 0.102**

(0.039) (0.056) (0.005) (0.014) (0.028)

MARKETRATES 0.001*** 0.002*** 0.001** 0.001** -0.001

(0.009) (0.008) (0.013) (0.015) (0.312)

Intercept YES YES YES YES YES

YEAR DUMMIES YES YES YES YES YES

COUNTRY DUMMIES YES YES YES YES YES

INDUSTRY DUMMIES YES YES YES YES YES

Adjusted R-Squared 0.432 0.429 0.435 0.444 0.383

No. of Observations 199693 199693 199693 199693 174714

36

Table 4. Foreign Institutional Ownership and Corporate Risk-Taking

This table reports the OLS estimation results of the following regression model: , where “Risk-taking” refers to five risk-taking measures, FIO is the percentage of

foreign institutional ownership in a firm, DIO is the percentage of total domestic institutional ownership

in a firm, and CONTROLS is a vector of control variables (firm and country characteristics). All

regressions include year, industry and country fixed effects whose coefficients are suppressed. Beneath

each coefficient estimate is p-statistics in parentheses based on robust standard errors clustered at both the

firm- and country-levels. ***, **, and * denote statistical significance at the 1%, 5%, and 10% levels,

respectively. Variable definitions and data sources are available in the Appendix.

Dependent Variable RISK1 RISK2 RISK3 RISK4 SRVOL R&D

Model (1) (2) (3) (4) (5) (6)

FIO 0.055*** 0.134*** 0.047*** 0.047*** 0.066*** 0.117***

(0.000) (0.000) (0.000) (0.000) (0.000) (0.000)

DIO -0.018*** -0.046*** -0.018*** -0.022*** -0.046*** -0.013**

(0.000) (0.000) (0.000) (0.000) (0.000) (0.039)

ROA -0.196*** -0.476*** -0.186*** -0.172*** -0.137*** -0.171***

(0.000) (0.000) (0.000) (0.000) (0.000) (0.000)

LEVERAGE 0.004 0.012 0.005 0.006 0.062*** -0.028***

(0.329) (0.279) (0.280) (0.138) (0.000) (0.000)

SIZE -0.009*** -0.022*** -0.009*** -0.009*** -0.012*** -0.006***

(0.000) (0.000) (0.000) (0.000) (0.000) (0.000)

SALESGROWTH 0.008*** 0.020*** 0.008*** 0.007*** 0.006*** 0.000

(0.000) (0.000) (0.000) (0.000) (0.000) (0.892)

CAPEX -0.001 -0.003 -0.003 -0.008 -0.002 -0.023

(0.944) (0.890) (0.752) (0.359) (0.848) (0.163)

GDPGROWTH 0.001* 0.002* 0.001 0.001* -0.000 -0.000

(0.054) (0.061) (0.136) (0.062) (0.847) (0.665)

ECONFREEDOM -0.000 -0.001 -0.001 -0.000 0.040*** -0.012***

(0.888) (0.793) (0.800) (0.899) (0.000) (0.001)

GDP

-0.008 -0.018 -0.015** -0.014** 0.022 0.017***

(0.232) (0.284) (0.010) (0.023) (0.478) (0.008)

MARKETRATES 0.000 0.001 0.000 0.000 0.001 0.000

(0.111) (0.110) (0.116) (0.172) (0.184) (0.815)

Intercept YES YES YES YES YES YES

YEAR DUMMIES YES YES YES YES YES YES

COUNTRY DUMMIES YES YES YES YES YES YES

INDUSTRY DUMMIES YES YES YES YES YES YES

Adjusted R-Squared 0.417 0.414 0.421 0.431 0.436 0.529

No. of Observations 91211 91211 91211 91211 110187 45263

37

Table 5. FDI, FPI Ownership and Corporate Risk-Taking

This table reports the OLS estimation results of the following regression model: where “Risk-taking” refers to five risk-taking

measures, FDI is the percentage of foreign direct investors’ ownership in a firm, FPI is the percentage of

foreign portfolio investors’ ownership in a firm, DIO is the percentage of total domestic institutional

ownership in a firm, and CONTROLS is a vector of control variables (firm and country characteristics).

All regressions include year, industry and country fixed effects whose coefficients are suppressed.

Beneath each coefficient estimate is p-statistics in parentheses based on robust standard errors clustered at

both the firm- and country-levels. ***, **, and * denote statistical significance at the 1%, 5%, and 10%

levels, respectively. Variable definitions and data sources are available in the Appendix.

Dependent Variable RISK1 RISK2 RISK3 RISK4 SRVOL R&D

Model (1) (2) (3) (4) (5) (6)

FDI 0.007** 0.019** 0.008*** 0.006** 0.011*** 0.003

(0.011) (0.010) (0.005) (0.045) (0.003) (0.480)

FPI 0.057*** 0.138*** 0.049*** 0.044*** 0.049** 0.135***

(0.000) (0.000) (0.000) (0.000) (0.016) (0.000)

DIO -0.017*** -0.045*** -0.017*** -0.021*** -0.051*** -0.011*

(0.000) (0.000) (0.000) (0.000) (0.000) (0.073)

ROA -0.193*** -0.469*** -0.183*** -0.172*** -0.121*** -0.185***

(0.000) (0.000) (0.000) (0.000) (0.000) (0.000)

LEVERAGE 0.010** 0.025** 0.010*** 0.010*** 0.068*** -0.021***

(0.012) (0.011) (0.009) (0.005) (0.000) (0.001)

SIZE -0.009*** -0.022*** -0.009*** -0.009*** -0.011*** -0.007***

(0.000) (0.000) (0.000) (0.000) (0.000) (0.000)

SALESGROWTH 0.006*** 0.016*** 0.006*** 0.006*** 0.005*** 0.006***

(0.000) (0.000) (0.000) (0.000) (0.000) (0.000)

CAPEX 0.000 0.000 -0.003 -0.009 -0.006 -0.022

(0.987) (0.989) (0.684) (0.275) (0.516) (0.264)

GDPGROWTH -0.008** -0.019** -0.008** -0.008** -0.002* -0.015***

(0.016) (0.014) (0.025) (0.027) (0.086) (0.000)

ECONFREEDOM 0.001** 0.003** 0.001* 0.001 0.021*** 0.000

(0.018) (0.021) (0.063) (0.113) (0.003) (0.618)

GDP 0.003 0.011 -0.007 -0.004 0.052 0.032***

(0.756) (0.651) (0.495) (0.676) (0.214) (0.001)

MARKETRATES 0.000** 0.001** 0.000* 0.000* -0.001 0.001*

(0.049) (0.036) (0.072) (0.084) (0.389) (0.065)







38

Table 6.FDI and FPI Ownership, Country-level Information Transparency and Corporate Risk-

Taking


where “Risk-

taking” refers to five risk-taking measures, FDI is the percentage of foreign direct investors’ ownership

in a firm, FPI is the percentage of foreign portfolio investors’ ownership in a firm, DIO is the percentage

of domestic institutional investors’ ownership in a firm, IT is country-level information transparency, and

CONTROLS is a vector of control variables (firm and country characteristics). All regressions include

year, industry and country fixed effects whose coefficients are suppressed. Beneath each coefficient

estimate is p-statistics in parentheses based on robust standard errors clustered at both the firm- and

country-levels. ***, **, and * denote statistical significance at the 1%, 5%, and 10% levels, respectively.

Variable definitions and data sources are available in the Appendix. For brevity, this table only reports the

regression when we measure risk-taking by RISK1. The results when we use other risk-taking measures

are qualitatively similar.

Dependent Variable RISK1 Information

Transparency AUDIT ANALYST ITENF INFSHA03 DISREQ LIASTA

FDI 0.053*** 0.032*** 0.025*** 0.032*** 0.037*** 0.024***

(0.001) (0.000) (0.000) (0.000) (0.009) (0.006)

FPI 0.141*** 0.118*** 0.040 0.090*** 0.083* 0.054**

(0.000) (0.000) (0.113) (0.000) (0.061) (0.028)

DIO -0.061* -0.028 -0.014 -0.046** -0.044 0.002

(0.079) (0.217) (0.593) (0.017) (0.326) (0.929)

FDI*IT -0.014*** -0.001*** -0.026*** -0.032*** -0.038** -0.023***

(0.006) (0.001) (0.001) (0.000) (0.031) (0.002)

FPI*IT -0.032*** -0.005*** -0.017 -0.072*** -0.073 -0.031

(0.003) (0.000) (0.542) (0.000) (0.191) (0.361)

DIO*IT 0.014 0.001 0.008 0.040* 0.038 -0.020

(0.127) (0.336) (0.777) (0.055) (0.428) (0.472)

IT 0.007*** 0.000 0.007 0.002 0.012 0.026***

(0.010) (0.569) (0.331) (0.831) (0.469) (0.002)

ROA -0.205*** -0.205*** -0.204*** -0.204*** -0.203*** -0.204***

(0.000) (0.000) (0.000) (0.000) (0.000) (0.000)

LEVERAGE 0.006 0.005 0.004 0.005 0.008* 0.005

(0.219) (0.325) (0.378) (0.310) (0.297) (0.281)

SIZE -0.007*** -0.007*** -0.007*** -0.007*** -0.007*** -0.007***

(0.000) (0.000) (0.000) (0.000) (0.000) (0.000)

SALESGROWTH 0.007*** 0.007*** 0.007*** 0.007*** 0.007*** 0.007***

(0.000) (0.000) (0.000) (0.000) (0.000) (0.000)

CAPEX 0.009 0.005 0.007 0.008 0.006 0.006

(0.472) (0.622) (0.573) (0.500) (0.606) (0.627)

GDPGROWTH 0.002 0.001 0.001 0.000 0.001 0.001

(0.257) (0.453) (0.431) (0.838) (0.583) (0.495)

ECONFREEDOM 0.006 0.008 0.007 0.011** 0.007 0.002

(0.258) (0.146) (0.173) (0.023) (0.324) (0.644)

GDP -0.005 -0.003 -0.004 -0.007 -0.004 -0.003

(0.167) (0.353) (0.238) (0.145) (0.381) (0.403)

MARKETRATES 0.001* 0.001** 0.001** 0.001** 0.001* 0.001**

(0.070) (0.031) (0.041) (0.027) (0.076) (0.033)

Intercept YES YES YES YES YES YES YEAR DUMMIES YES YES YES YES YES YES COUNTRY DUMMIES YES YES YES YES YES YES INDUSTRY DUMMIES YES YES YES YES YES YES



39

Table 7.FDI and FPI Ownership, Country-level Institutional Environment (Legal Environment and

Investor Protection) and Corporate Risk-Taking

This table reports the OLS estimation results of the following regression model where “Risk-taking”

refers to four risk-taking measures: , FDI is the percentage of foreign direct investors’

ownership, FPI is the percentage of foreign portfolio investors’ ownership, DIO is the percentage of

domestic institutional investors’ ownership, IE is country-level institutional environment (legal

environment and investor protection here), and CONTROLS is a vector of control variables (firm and

country characteristics). All regressions include year, industry and country fixed effects whose

coefficients are suppressed. Beneath each coefficient estimate is p-statistics in parentheses based on

robust standard errors clustered at both the firm- and country-levels. ***, **, and * denote statistical

significance at the 1%, 5%, and 10% levels, respectively. Variable definitions and data sources are in the

Appendix. For brevity, this table only reports the regression when we measure risk-taking by RISK1. The

results when we use other risk-taking measures are qualitatively similar.

Dependent Variable RISK1

Institutional environment (IE) EFFJUD LEGCOM ANTISELF

FDI 0.041*** 0.018*** 0.026**

(0.002) (0.000) (0.024)

FPI 0.069 0.043*** 0.037

(0.147) (0.004) (0.217)

DIO -0.155 -0.020 -0.030

(0.131) (0.414) (0.367)

FDI*IE -0.004** -0.018*** -0.033*

(0.018) (0.003) (0.084)

FPI*IE -0.005 -0.043** -0.026

(0.377) (0.022) (0.508)

DINST*IE 0.015 0.010 0.036

(0.140) (0.672) (0482)

IE -0.002 0.007* -0.005

(0.220) (0.094) (0.669)

ROA -0.205*** -0.206*** -0.204***

(0.000) (0.000) (0.000)

LEVERAGE 0.005 0.005 0.005

(0.281) (0.345) (0.289)

SIZE -0.007*** -0.006*** -0.007***

(0.000) (0.000) (0.000)

SALESGROWTH 0.007*** 0.007*** 0.007***

(0.000) (0.000) (0.000)

CAPEX 0.005 0.006 0.007

(0.647) (0.625) (0.585)

GDPGROWTH 0.000 0.001 0.000

(0.694) (0.643) (0.927)

ECONFREEDOM 0.011** 0.006 0.014***

(0.017) (0.266) (0.003)

GDP

-0.004 -0.003 -0.008**

(0.299) (0.411) (0.042)

MARKETRATES 0.001** 0.001* 0.001**

(0.031) (0.062) (0.012)

Intercept YES YES YES

YEAR DUMMIES YES YES YES COUNTRY DUMMIES YES YES YES INDUSTRY DUMMIES YES YES YES

Adjusted R-Squared 0.393 0.392 0.389

No. of Observations 61398 61398 62867

40

Table 8. Development of Foreign institutional Investor and Corporate Risk-Taking


after restricting the where “Risk-taking” refers to five risk-taking measures, FIO is the percentage of

foreign institutional ownership in a firm, DIO is the percentage of total domestic institutional ownership

in a firm, and CONTROLS is a vector of control variables (firm and country characteristics). All

regressions include year, industry and country fixed effects whose coefficients are suppressed. Beneath

each coefficient estimate is p-statistics in parentheses based on robust standard errors clustered at both the




Holding Type

Developed to

Developed

Developed to

Developing

Developing to

Developing

Developing to

Developed

Model (1) (2) (3) (4)

FIO 0.056*** 0.011* 2.029 -0.068

(0.000) (0.099) (0.346) (0.786)

DIO -0.011** -0.015 -0.029 0.008

(0.014) (0.322) (0.683) (0.215)

ROA -0.201*** -0.044** -0.012 -0.191***

(0.000) (0.036) (0.614) (0.000)

LEVERAGE 0.004 0.025*** -0.010 -0.011

(0.505) (0.001) (0.523) (0.100)

SIZE -0.009*** -0.007*** -0.003** -0.008***

(0.000) (0.000) (0.046) (0.000)

SALESGROWTH 0.008*** 0.004*** 0.012*** 0.011***

(0.000) (0.007) (0.000) (0.004)

CAPEX 0.014 -0.024*** -0.022 0.032**

(0.210) (0.007) (0.280) (0.012)

GDPGROWTH 0.001*** -0.001 -0.001 0.001

(0.000) (0.253) (0.238) (0.148)

ECONFREEDOM -0.003 0.011** -0.006 0.007

(0.114) (0.012) (0.328) (0.117)

GDP 0.003 -0.010 -0.058* -0.000

(0.857) (0.544) (0.064) (0.991)

MARKETRATES 0.001* 0.000 0.000 0.000

(0.081) (0.395) (0.774) (0.974)

Intercept YES YES YES YES

YEAR DUMMIES YES YES YES YES

COUNTRY DUMMIES YES YES YES YES

INDUSTRY DUMMIES YES YES YES YES

Adjusted R-Squared 65938 11007 738 7151

No. of Observations 0.419 0.145 0.261 0.305

41

Table 9. Institution Type and Corporate Risk-Taking


after restricting the institutional holdings by both foreign and domestic to the distinct institution

type.“Risk-taking” refers to five risk-taking measures, FIO is the percentage of foreign institutional

ownership in a firm, DIO is the percentage of total domestic institutional ownership in a firm, and

CONTROLS is a vector of control variables (firm and country characteristics). All regressions include

year, industry and country fixed effects whose coefficients are suppressed. Beneath each coefficient

estimate is p-statistics in parentheses based on robust standard errors clustered at both the firm- and

country-levels. ***, **, and * denote statistical significance at the 1%, 5%, and 10% levels, respectively.

Variable definitions and data sources are available in the Appendix.


Institution Type Banks Insurance

Companies

Investment

Companies

Investment

Advisors

Pension

Funds &

Endowments

Hedge

Funds &

Venture

Capital

Model (1) (2) (3) (4) (5) (6)

FIO 0.429 1.351 0.064*** 0.070*** 0.301*** -0.184

(0.435) (0.258) (0.001) (0.000) (0.003) (0.112)

DIO -0.005 -0.109 -0.017*** -0.030*** -0.096*** -0.068

(0.986) (0.697) (0.003) (0.000) (0.006) (0.554)

ROA -0.191*** -0.191*** -0.189*** -0.197*** -0.179*** -0.187***

(0.000) (0.000) (0.000) (0.000) (0.000) (0.000)

LEVERAGE -0.005 -0.007*** -0.002 0.003 -0.003 -0.009***

(0.248) (0.005) (0.517) (0.508) (0.365) (0.001)

SIZE -0.007*** -0.008*** -0.009*** -0.009*** -0.007*** -0.008***

(0.000) (0.000) (0.000) (0.000) (0.000) (0.000)

SALESGROWTH 0.010*** 0.012*** 0.009*** 0.008*** 0.010*** 0.006***

(0.000) (0.000) (0.000) (0.000) (0.000) (0.000)

CAPEX 0.022 -0.008 0.002 0.004 0.006 0.002

(0.113) (0.271) (0.835) (0.655) (0.580) (0.809)

GDPGROWTH 0.000 0.001** 0.001 0.001** 0.001* 0.001

(0.478) (0.023) (0.137) (0.039) (0.074) (0.104)

ECONFREEDOM 0.005 -0.002 0.000 -0.001 -0.002 -0.000

(0.128) (0.590) (0.996) (0.589) (0.377) (0.933)

GDP -0.004 0.009 -0.013* -0.004 -0.010 0.001

(0.796) (0.812) (0.067) (0.605) (0.344) (0.892)

MARKETRATES -0.000 0.001*** 0.001** 0.000* 0.001** 0.000

(0.650) (0.001) (0.042) (0.055) (0.042) (0.388)







42

Table 10. FDI, FPI Ownership and Corporate Risk-Taking: Correcting for Endogeneity

This table reports the regression results for three regressions estimated to tackle the endogeneity issue in

Table 5. Model 1 is an OLS model with lagged independent variables. Model 2 is an OLS model with the

first difference of the dependent variable and the independent variables. Model 3 is a 2SLS model using

political rights as the instrumental variable for FDI. All regressions include year, industry and country

fixed effects whose coefficients are suppressed. Beneath each coefficient estimate is p-statistics in

parentheses based on robust standard errors clustered at both the firm- and country-levels. ***, **, and *

denote statistical significance at the 1%, 5%, and 10% levels, respectively. Variable definitions and data

sources are available in the Appendix.


Model Specification

Lagged independent

variables First Difference 2SLS

Model (1) (2) (3)

FDI 0.008** 0.002* 0.090**

(0.038) (0.067) (0.045)

FPI 0.058*** 0.010*** 0.047***

(0.000) (0.004) (0.001)

DIO -0.015*** 0.000 -0.011***

(0.000) (0.881) (0.006)

ROA -0.144*** 0.014*** -0.195***

(0.000) (0.000) (0.000)

LEVERAGE 0.013*** 0.004* 0.011***

(0.000) (0.091) (0.002)

SIZE -0.010*** 0.004*** -0.009***

(0.000) (0.009) (0.000)

SALESGROWTH 0.007*** -0.001*** 0.006***

(0.000) (0.008) (0.000)

CAPEX -0.014 -0.003 -0.005

(0.209) (0.286) (0.548)

GDPGROWTH -0.010** -0.004* -0.006

(0.038) (0.098) (0.185)

ECONFREEDOM 0.001 0.000 0.001**

(0.224) (0.425) (0.019)

GDP -0.006 0.012 -0.002

(0.516) (0.602) (0.846)

MARKETRATES 0.000 0.000 0.001**

(0.810) (0.171) (0.035)

Intercept YES YES YES

YEAR DUMMIES YES YES YES

COUNTRY DUMMIES YES YES YES

INDUSTRY DUMMIES YES YES YES

Adjusted R-Squared 0.349 0.0150 0.394

No. of Observations 48255 35060 60813

43

Table 11. Event Study of ROA volatility around Large Block Purchases by Foreign Investors

This table presents our results for an event study around the announcement of acquisitions of large block

of shares (>5%) by foreign investors available in the SDC Platinum M&A Database. Panel A compares

the volatility of ROA between event and control firms that are matched 1-to-1 by finding firms in the

same market and industry with the closest market capitalization, or if not possible, a firm in the same

market with the closest market capitalization. ROA volatility is reported for pre-event periods from year

-5, year -4, and year -3 up to year -1 before the announcement date and post-event periods from year +1

to year +3, year +4, and year +5 after the announcement date. Column (1) reports the average ROA

volatility for our event sample. Column (2) reports the average ROA volatility for our matched control

sample. Panel B provides further analysis by differentiating between acquisitions from developed or

developing acquirers to developed or developing targets. See Appendix I for variable definitions. The

difference and the t-statistic associated with the difference are reported in the last two columns. ***, **, *

denotes statistical significance at the 1%, 5%, and 10% levels, respectively.

Average ROA

Volatility -

Event Sample

Average ROA

Volatility -

Control Sample

Difference t-test of

Difference

(1) (2) (1)-(2) (1)-(2)

Panel A. ROA Volatility Before and After Large Block Purchases by Foreign Investors (N=255)

Year -5 0.084 0.078 0.006 0.703

Year -4 0.073 0.070 0.003 0.383

Year -3 0.066 0.062 0.005 0.630

Year +3 0.084 0.058 0.026 2.817***

Year +4 0.086 0.060 0.026 2.970***

Year +5 0.097 0.063 0.034 3.775***

Panel B. ROA Volatility Before and After Large Block Purchases by Foreign Investors from:

Developing to Developing Countries (N=15)

Year -5 0.046 0.051 -0.006 -0.418

Year -4 0.040 0.038 0.002 0.227

Year -3 0.034 0.037 -0.003 -0.303

Year +3 0.064 0.029 0.035 2.113**

Year +4 0.062 0.032 0.031 2.120**

Year +5 0.071 0.032 0.039 2.214**

Developing to Developed Countries (N=36)

Year -5 0.051 0.060 -0.008 -0.451

Year -4 0.045 0.057 -0.012 -0.716

Year -3 0.044 0.043 0.001 0.046

Year +3 0.037 0.041 -0.004 -0.247

Year +4 0.045 0.041 0.004 0.245

Year +5 0.048 0.040 0.009 0.641

Developed to Developing Countries (N=12)

Year -5 0.111 0.078 0.033 0.811

Year -4 0.088 0.079 0.009 0.249

Year -3 0.093 0.069 0.024 0.736

Year +3 0.129 0.071 0.058 1.117

Year +4 0.145 0.065 0.081 1.608

Year +5 0.165 0.060 0.105 2.044*

Developed to Developed Countries (N=192)

Year -5 0.091 0.084 0.007 0.606

Year -4 0.080 0.074 0.005 0.542

Year -3 0.072 0.067 0.005 0.475

44

Year +3 0.092 0.062 0.030 2.607***

Year +4 0.093 0.066 0.027 2.478**

Year +5 0.104 0.070 0.034 3.075***

45

Table 12. Foreign Strategic Institutional Ownership and Crash Risk


where FSIO is the percentage of foreign strategic institutional ownership in a

firm, RISK1 is a corporate risk-taking measure, and CONTROLS is a vector of control variables (firm

and country characteristics). All regressions include year and industry fixed effects whose coefficients

are suppressed. Beneath each coefficient estimate is p-statistics in parentheses based on robust standard

errors clustered at both the firm- and country-levels. ***, **, and * denote statistical significance at the

1%, 5%, and 10% levels, respectively. Variable definitions and data sources are available in the

Appendix.

Dependent Variable NCSKEW DUVOL COUNT

Model (1) (2)

(3) (4)

(5) (6)

FSIO -0.103* -0.175***

-0.059** -0.097***

-0.054 -0.088***

(0.061) (0.000)

(0.036) (0.000)

(0.105) (0.000)

FSIO*RISK1

0.490

0.275

0.277

(0.152)

(0.144)

(0.226)

RISK1

0.016

-0.025

0.091

(0.861)

(0.581)

(0.162)

Lagged NCSKEW 0.061*** 0.057***

0.030*** 0.029***

0.034*** 0.031***

(0.000) (0.000)

(0.000) (0.000)

(0.000) (0.000)

SIGMA -1.501 -1.279

-1.075 -0.902

-0.632 -0.827

(0.379) (0.490)

(0.198) (0.318)

(0.584) (0.489)

RET -0.408** -0.390**

-0.236*** -0.222**

-0.254** -0.280**

(0.011) (0.030)

(0.002) (0.010)

(0.021) (0.015)

DTURN 1.614*** 1.730***

0.866*** 0.976***

0.919*** 1.153***

(0.001) (0.000)

(0.001) (0.001)

(0.010) (0.006)

OPACITY 0.016*** 0.015***

0.007*** 0.007***

0.012*** 0.012***

(0.000) (0.001)

(0.000) (0.001)

(0.000) (0.001)

LEVERAGE 0.072*** 0.066**

0.041*** 0.037***

0.027 0.024

(0.005) (0.014)

(0.002) (0.009)

(0.187) (0.269)

ROA -0.003 0.029

-0.002 0.010

-0.010 0.020

(0.952) (0.678)

(0.936) (0.790)

(0.765) (0.680)

MCAP 0.023*** 0.026***

0.012*** 0.013***

0.015*** 0.018***

(0.000) (0.000)

(0.001) (0.001)

(0.000) (0.000)

MTB 0.004** 0.004***

0.002** 0.002**

0.002*** 0.002***

(0.013) (0.004)

(0.040) (0.023)

(0.001) (0.004)

GDPGROWTH 0.007 0.006

0.004 0.003

0.004 0.003

(0.278) (0.383)

(0.298) (0.387)

(0.252) (0.365)

GDP 0.018 0.018

0.008 0.009

0.014 0.013

(0.376) (0.380)

(0.436) (0.427)

(0.266) (0.317)

MARKETRATES -0.007 -0.008

-0.004 -0.005

-0.003 -0.003

(0.225) (0.223)

(0.166) (0.164)

(0.368) (0.329)

Intercept YES YES

YES YES

YES YES

YEAR DUMMIES YES YES

YES YES

YES YES

INDUSTRY DUMMIES YES YES

YES YES

YES YES

Adjusted R-Squared 0.034 0.035

0.042 0.043

0.017 0.017


46

Table 13.FDI and FPI Ownership and Crash Risk


where FDI is foreign direct investors’ ownership, FPI is foreign portfolio

investors’ ownership, DIO is the percentage of total domestic institutional ownership in a firm, RISK1 is

a corporate risk-taking measure, and CONTROLS is a vector of control variables (firm and country

characteristics). All regressions include year and industry fixed effects whose coefficients are suppressed.

Beneath each coefficient estimate is p-statistics in parentheses based on robust standard errors clustered at

both the firm- and country-levels. ***, **, and * denote statistical significance at the 1%, 5%, and 10%

levels, respectively. Variable definitions and data sources are available in the Appendix.

Dependent Variable NCSKEW DUVOL COUNT

Model (1) (2)

(3) (4)

(5) (6)

FDI -0.085** -0.107***

-0.045** -0.063***

-0.048 -0.047*

(0.037) (0.000)

(0.025) (0.000)

(0.112) (0.053)

FPI 0.114 0.097

0.038 0.039

0.134* 0.117

(0.270) (0.311)

(0.477) (0.422)

(0.079) (0.123)

DIO 0.247*** 0.215***

0.126*** 0.108***

0.153*** 0.133***

(0.000) (0.000)

(0.000) (0.000)

(0.000) (0.000)

FDI*RISK1 -0.008

0.090

-0.108

(0.981)

(0.552)

(0.679)

FPI*RISK1

-0.019

-0.097

-0.122

(0.968)

(0.724)

(0.854)

DIO*RISK1 0.305**

0.191**

0.233**

(0.045)

(0.021)

(0.042)

RISK1

0.072

0.004

0.117

(0.469)

(0.927)

(0.156)

Lagged NCSKEW 0.031*** 0.030***

0.016*** 0.016***

0.016*** 0.015***

(0.000) (0.000)

(0.000) (0.000)

(0.007) (0.009)

SIGMA -0.653 -0.920

-0.465 -0.486

-0.295 -0.832

(0.735) (0.639)

(0.602) (0.596)

(0.813) (0.479)

RET -0.292 -0.338

-0.144 -0.152

-0.211 -0.285**

(0.157) (0.118)

(0.100) (0.102)

(0.107) (0.025)

DTURN 1.213* 1.417*

0.338** 0.423**

0.468 0.724

(0.097) (0.095)

(0.036) (0.036)

(0.425) (0.283)

OPACITY 0.008* 0.006

0.003 0.002

0.007** 0.006*

(0.066) (0.143)

(0.134) (0.220)

(0.010) (0.051)

LEVERAGE 0.015 0.007

0.009 0.005

-0.005 -0.008

(0.628) (0.826)

(0.587) (0.789)

(0.822) (0.768)

ROA 0.054 0.110***

0.029 0.053***

0.025 0.066**

(0.168) (0.008)

(0.147) (0.009)

(0.309) (0.039)

MCAP 0.023*** 0.025***

0.012*** 0.013***

0.015*** 0.016***

(0.000) (0.000)

(0.000) (0.000)

(0.000) (0.000

MTB 0.003*** 0.003**

0.001* 0.001

0.002*** 0.002**

(0.008) (0.015)

(0.056) (0.112)

(0.002) (0.022)

GDPGROWTH -0.009* -0.011**

-0.006* -0.006**

-0.005 -0.006*

(0.085) (0.043)

(0.061) (0.038)

(0.184) (0.083)

GDP 0.019 0.019

0.008 0.009

0.012 0.010

(0.156) (0.162)

(0.215) (0.204)

(0.161) (0.227)

47

MARKETRATES -0.005 -0.006

-0.002 -0.003

-0.004 -0.005

(0.297) (0.249)

(0.313) (0.271)

(0.225) (0.150)

Intercept YES YEs

YES YES

YES YES

YEAR DUMMIES YES YES

YES YES

YES YES

INDUSTRYDUMMIES YES YES

YES YES

YES YES

Adjusted R-Squared 0.040 0.041

0.046 0.047

0.018 0.019 No. of Observations 56626 52199 56626 52199 56651 52218

48

Table 14. FDI and FPI Ownership and Crash Risk: Correcting for Endogeneity

This table reports the regression results for two regressions estimated to tackle the endogeneity issue in

Table 13. Table 1 is an OLS model with the first difference of the dependent variable and the independent

variables. Model 2 is a 2SLS model where FDI is instrumented on foreign sales scaled by total assets. All

regressions include year and industry fixed effects whose coefficients are suppressed. Beneath each

coefficient estimate is p-statistics in parentheses based on robust standard errors clustered at both the



Dependent Variable NCSKEW

Model specification First-differenced Variables 2SLS

Model (1) (2) FDI -0.117*** -0.752**

(0.009) (0.030)

FPI 0.367** 0.228**

(0.027) (0.000)

DIO 0.221*** 0.199***

(0.000) (0.000)

Lagged NCSKEW -0.489*** 0.026***

(0.000) (0.000)

SIGMA -1.136 0.004

(0.318) (0.998)

RET -0.164 -0.209

(0.156) (0.224)

DTURN 0.706 0.798

(0.101) (0.131)

OPACITY -0.017*** 0.004

(0.000) (0.410)

LEVERAGE 0.141** 0.006

(0.011) (0.854)

ROA 0.102*** 0.083*

(0.004) (0.081)

MCAP 0.215*** 0.023***

(0.000) (0.000)

MTB

0.002* 0.002*

(0.086) (0.069)

GDPGROWTH 0.009 0.011

(0.463) (0.293)

GDP -0.411 0.047**

(0.507) (0.020)

MARKETRATES -0.012 0.005

(0.110) (0.379)

Intercept YES YES YEAR DUMMIES YES YES INDUSTRY DUMMIES YES YES

Adjusted R-Squared 0.264 0.020 No. of Observations 41108 39628

49

Appendix: Variable Definitions and Data Sources

Variable Acronym Definition Data Source

Panel A: Corporate risk-taking variables

Company Earnings Volatility RISK1

is the earnings before interest and taxes of firm i in year t; is the total

assets of firm i in year t; with T over the windows (0 to +4, +1 to +5, +2 to +6, +3

to +7, +4 to +8, +5 to +9, +6 to +10, etc.).

Worldscope

Company Earnings Range RISK2



to +7, +4 to +8, +5 to +9, +6 to +10 etc.).

Worldscope

Company Earnings Volatility

(Adjusted for Country)

RISK3

indexes the number of firms within the country c and year t, is the

earnings before interest and taxes of firm i in year t; is the total assets of

firm i in year t; with T over the windows (0 to +4, +1 to +5, +2 to +6, +3 to +7, +4

to +8, +5 to +9, +6 to +10, etc.).

Worldscope

50

Company Earnings Volatility

(Adjusted for Country and Industry)

RISK4

indexes the number of firms within the country c, industry d and year

t, is the earnings before interest and taxes of firm i in year t; is the

total assets of firm i in year t; with T over the windows (0 to +4, +1 to +5, +2 to

+6, +3 to +7, +4 to +8, +5 to +9, +6 to +10, etc.).

Worldscope

Stock Return Volatility SRVOL Standard deviation of monthly stock returns over the next two years including the

current fiscal year.

Datastream

Research and Development Ratio R&D

is the research and development expenses of firm i in year t; is the total


to +7, +4 to +8, +5 to +9, +6 to +10, etc.).

Worldscope

Panel B: Crash-Risk Measures

Negative Skewness NCSKEW

is de-meaned firm-specific weekly return of firm i in year t. This is the

negative sample skewness of the firm-specific weekly returns over the year.

Datastream

Return Asymmetries DUVOL

and are de-meaned firm-specific weekly return for up and down weeks

respectively; and are the number of up and down weeks respectively.

Datastream

51

Number of crashes minus number of

jumps

COUNT We first detect crash (jump), which occurs when the firm-specific weekly return is

3.09 standard deviations below (above) its mean over the fiscal year, and then we

compute COUNT as the number of crashes minus the number of jumps over the

fiscal year.

Datastream

Panel C: Firm Ownership Measures

Foreign Strategic Institutional

Ownership

FSIO Aggregate strategic holdings by investors domiciled in a foreign country denoted

as a percentage of the total number of shares outstanding at the end of the previous

year, with a disclosed holding above 5 percent of the total number of shares

outstanding.

Datastream

Foreign Institutional Ownership FIO Aggregate equity portfolio holdings by institutions domiciled in a foreign country

denoted as a percentage of the total number of shares outstanding at the end of the

previous year.

FactSet Ownership (LionShares)

Database

Foreign Direct Investment

Ownership

FDI Aggregate strategic holdings by investors domiciled in a foreign country denoted

as a percentage of the total number of shares outstanding at the end of the previous

year, with a disclosed holding above 5 percent of the total number of shares

outstanding.

Datastream

Foreign Portfolio Investment

Ownership

FPI Aggregate equity portfolio holdings by institutions domiciled in a foreign country

denoted as a percentage of the total number of shares outstanding at the end of the

previous year, with individual firm holdings below 5 percent.


Database

Domestic Institutional Ownership DIO Aggregate equity holdings by domestic institutional investors as a percentage of

total number of shares outstanding at the end of the previous year.


Database

Panel D: Firm-level control variables

Return on Assets ROA


assets of firm i in year t;

Worldscope

Leverage LEVERAGE

is net debt of firm i in year t; is the total assets of firm i in year t;

Worldscope

Firm Size SIZE The natural logarithm of total sales denominated in U.S. dollar Worldscope

Annual Sales Growth SALESGROWTH The annual change in total sales. Worldscope

Market Capitalization MCAP The natural logarithm of market capitalization Worldscope

52

Market to Book Ratio MTB Market value of equity divided by the book value of equity. Worldscope

Sigma SIGMA The standard deviation of firm-specific weekly returns Worldscope

Return RET The mean of the firm-specific weekly returns for firm i in given year t. Worldscope

Detrended Turnover DTURN The average monthly share turnover in the current year minus the average monthly

share turnover in the previous year, where monthly share turnover is given by the

monthly trading volume divided by the number of shares outstanding.

Worldscope

Earning Opacity OPACITY The sum of the absolute value of discretionary accruals over the prior three years,

where discretionary accruals are estimated from the modified Jones model.

Worldscope

Panel E: Country-level control variables

Economic Freedom ECONFREEDOM The index published in Economic Freedom of the World measures the degree to

which the policies and institutions of countries are supportive of economic

freedom. The cornerstones of economic freedom are personal choice, voluntary

exchange, freedom to compete, and security of privately owned property. Forty-

two variables are used to construct a summary index and to measure the degree of

economic freedom in five broad areas: (1) Size of Government; (2) Legal System

and Property Rights; (3) Sound Money; (4) Freedom to Trade Internationally; (5)

Regulation.

Economic Freedom of the World

(Gwartney et al., 2008)

Market Interest Rates MARKETRATES The real market interest rate. World Development Indicators.

Annual GDP Growth GDPGROWTH Annual percentage growth rate of GDP, at 2005 U.S. dollar. Where GDP is the

sum of gross value added by all resident producers in the economy plus any

product taxes and minus any subsidies not included in the value of the products. It

is calculated without making deductions for depreciation of fabricated assets or for

depletion and degradation of natural resources.

World Development Indicators

Logarithm of GDP per capita GDP The natural logarithm of GDP per capita, which is the gross domestic product

divided by midyear population of a given country, at constant 2005 U.S. dollar.

World Development Indicators

Panel F: Country-level Information Transparency Variables

Credibility of financial accounting

disclosure

AUDIT Variable indicating the percentage of firms in the country audited by the Big 5

accounting firms. AUDIT equals 1, 2, 3 or 4 if the percentage ranges between

[0,25%], (25%,50%], (50%, 75%] and (75%, 100%], respectively.

Bushman et al. (2004)

Number of financial analysts ANALYST Number of analysts following the largest 30 companies in each country in 1996. Bushman et al. (2004)

53

Extent of insider trading activities ITENF Indicator variable equal to one if the country enforced insider trading laws before

1995, zero otherwise.

Bushman et al. (2004)

Disclosure requirements index DISREQ The index of disclosure equals the arithmetic mean of: (1) Prospect; (2)

Compensation; (3) Shareholders; (4) Inside ownership; (5) Contracts Irregular; (6)

and Transactions.

La Porta et al. (2006)

Liability standard index LIASTA The index of liability standards equals the arithmetic mean of: (1) Liability

standard for the issuer and its directors; (2) Liability standard for the distributor;

and (3) Liability standard for the accountant.


Information sharing INFSHA03 Dummy equals to one if information sharing operates in 2003, and zero otherwise. Djankov et al. (2007)

Panel G: Legal Environment and Shareholder Protection Variables

Efficiency of judicial system EFFJUD

Assessment of the ‘‘efficiency and integrity of the legal environment as it affects

business, particularly foreign firms’’ produced by the country risk rating agency

Business International Corp. It ‘‘may be taken to represent investors’ assessments

of conditions in the country in question.’’ Average between 1980 and 1983. Scale

from zero to 10; with lower scores, lower efficiency levels.


Anti-self-dealing index ANTISELF Average of ex-ante and ex-post private control of self-dealing. Djankov et al. (2008)

English common law LEGCOM A dummy equals 1 if a country adopts the common law system, zero otherwise. La Porta et al. (1998)

Panel H: Instrumental Variables

Political rights index POLRIGHTS

A higher political rights rating indicates a political system that includes free and

fair elections, those who are elected rule, competitive political parties or other

political groupings, the opposition plays an important role and has actual power,

and minority groups have reasonable self-government or can participate in the

government through informal consensus. The political rights index ranges from

one to seven, where a higher rating corresponds to stronger political rights. (Qi,

Roth, and Wald, 2010)

Freedom House (2013)

Foreign sales scaled by total assets FSALES A firm’s sales in foreign countries divided by its total assets. Worldscope

54

References

Acharya VV, Amihud Y, Litov L (2011) Creditor rights and corporate risk-taking.J. Financial

Econom.102: 150-166.

Aggarwal R, Erel I, Ferreira M, Matos P(2011) Does governance travel around the world? Evidence

from institutional investors,J. Financial Econom.100: 154-181.

AghionP, Reene JV, ZingalesL (2013) Innovation and institutional ownership. Am. Econ. Rev.103:

277-304.

Albuquerque R, Durnev A, Koskinen Y (2013)Corporatesocial responsibility and firm risk: Theory

and empirical evidence. Working paper.

Almazan, Andres, Jay Hartzell, and Laura Starks (2005) Active Institutional Shareholders and Cost

of Monitoring: Evidence from Executive Compensation. Financial Management 34, 5-34.

Amihud Y, Lev B (1981) Risk reduction as a managerial motive for conglomerate

mergers.BellJ.Econom. 12: 605-617.

An H, Zhang T (2013) Stock price synchronicity, crash risk, and institutional investors.J. Corp.

Finance 21: 1-15.

Anderson RC, Mansi SA, Reeb DM (2003) Founding family ownership and the agency cost of debt.J.

Financial Econom.68: 263-285.

Ball R (2009)The global financial crisis and the efficient market hypothesis: What have we learned?

J. Applied Corp. Finance 21: 8-16.

Barber, B. (2006), Monitoring the Monitor: Evaluating CalPERS Activism, University of California at

Davis working paper.

Baumol W, Litan RE, Schramm CJ (2009) Good capitalism, bad capitalism, and the economics of

growth and prosperity. New Haven, CT: Yale University Press.

BargeronLL, Lehn KM, Zutter CJ (2010) Sarbanes-Oxley and corporate risk-taking. J. Accounting

and Econ.49: 34-52.

Bekaert G, Harvey CR, LumsdaineRL (2002) Dating the integration of world equity markets.J.

Financial Econom.65: 203-247.

Bleck A, Liu X (2007) Market transparency and the accounting regime.J. Account. Res.45: 229-256.

Boubakri N, Cosset JC, Saffar W (2008) Political connections of newly privatized firms.J. Corp.

Finance 14: 654-673.

Boubakri N, Cosset JC, Saffar W (2013)The role of state and foreign owners in corporate risk-taking:

Evidence from privatization.J. Financial Econom.Forthcoming.

Brennan MJ, Cao HH(1997) International portfolio investment flows. J. Finance52, 1851-1880.

Brickley, J., R. Lease, and C. Smith (1988) Ownership Structure and Voting on Antitakeover

amendments, J. of Financial Econom.20, 267-292.

Bushman RM, Piotroski JD, Smith AJ (2004) What determines corporate transparency? J. Accounting

Res.42: 207-252.

Callen JL, Fang X(2013) Institutional investor stability and crash risk: Monitoring versus short-

termism? J. Bank. Finance 37:3047-3063.

Carleton, W.T., J.M. Nelson, and M. S. Weisbach (1998) The Influence of Institutions on Corporate

Governance through Private Negotiations: Evidence from TIAACREF. J. of Finance, 1135-

1014.

Chan K, Menkveld AJ, Yang Z(2008) Information asymmetry and asset prices: Evidence from the

China foreign share discount.J. Finance63: 159-196.

Chen X, Harford J, Li K (2007) Monitoring: Which institutions matter?J. Financial Econom. 86: 279-

305.

Chen J, Hong H, SteinJC(2001) Forecasting crashes: Trading volume, past returns, and conditional

skewness in stock prices.J. Financial Econom.61: 345-381.

Chen CR, Steiner TL, Whyte AM (2006) Does stock option-based executive compensation induce

risk-taking? An analysis of the banking industry.J.Bank.Finance30: 915-945.

ChoeH, Kho BC, StulzRM(2005) Do domestic investors have an edge? The trading experience of

foreign investors in Korea.Rev. Financial Stud.18: 795-829.

Coles JL, Daniel ND, NaveenL(2006) Managerial incentives and risk-taking.J. Financial Econom.79:

431-468.

55

Cornett MM, Marcus AJ, TehranianH(2008) Corporate governance and pay-for-performance: The

impact of earnings management.J. Financial Econom.87: 357-373.

Covrig V, Lau ST, NgL(2006) Do domestic and foreign fund managers have similar preferences for

stock characteristics? A cross-country analysis.J. Int. Bus. Stud. 37: 407-429.

Csermely Á, VinczeJ(1999)Leverage and foreign ownership in Hungary. Working paper.

Del Guerio, Diane, and Jennifer Hawkins (1998) The Motivation and Impact of Pension Fund

Activism.J. of Financial Econom 52, 293-340.

Dimson E (1979) Risk measurement when shares are subject to infrequent trading.J. Financial

Econom.7: 197-226.

DjankovS,McLiesh C, Shleifer A (2007)Private credit in 129 countries. J. Financial Econom.84: 299-

329.

DjankovS, La Porta R, Lopez-de-Silanes F, ShleiferA (2008)The law and economics of self-

dealing.J.Financial Econom. 88:430-465.

Ertimur, Yonca, Fabrizio Ferri, and Stephen R. Stubben (2006) Board of Directors’Responsiveness to

Shareholders: Evidence from Majority-Vote Shareholder Proposals, Harvard Business

School working paper.

Faccio MA, Marchica MT, Mura R (2011)Large shareholder diversification and corporate risk-

taking.Rev. Financial Stud.24: 3601-3641.

Ferreira MA, Matos P (2008) The colors of investors’ money: The role of institutional investors

around the world.J. Financial Econom.88: 499-533.

Gillan, Stuart L., and Laura T. Starks (2007) The evolution of shareholder activism in the United

States, J. of Applied Corp. Finance 19, 55-73.

Griffin JM, Lemmon ML (2002) Book-to-market equity, distress risk, and stock returns. J. Finance

57: 2317-2336.

Gurunlu M, GursoyG(2010) The influence of foreign ownership on capital structure of non-financial

firms: Evidence from Istanbul stock exchange. IUP J. Corporate Governance,9, 21-29.

Gwartney JD, LawsonR, Norton S (2008) Economic Freedom of the World: 2008 Annual

Report.Source: http://www.freetheworld.com/release.html.

Hagendorff J, Vallascas F (2011) CEO pay incentives and risk-taking: Evidence from bank

acquisitions. J. Corp. Finance17: 1078-1095.

Haugen RA, SenbetLW (1988) Bankruptcy and agency costs: Their significance to the theory of

optimal capital structure. J. Financial Quant. Anal. 23: 27-38.

Hilary G, HuiKW(2009) Does religion matter in corporate decision making in America? J. Financial

Econom.93: 455-473.

Hirshleifer D, Hsu P, Li D (2013)Innovative efficiency and stock returns.J. Financial Econom.107:

632-654.

Hirshleifer D, Thakor AV (1992) Managerial conservatism, project choice, and debt.Rev. Financial

Stud.5: 437-470.

HolmstromB,Ricart I CostaJ(1986) Managerial incentives and capital management. Q. J. Econ.101:

835-860.

Hutton AP, Marcus A, TehranianH (2009) Opaque financial reports, , and crash risk. J. Financial

Econom. 94: 67-86.

Jensen MC, MecklingWH(1976) Theory of the firm: Managerial behavior, agency costs and

ownership structure. J. Financial Econom.3: 305-360.

Jin L,Myers SC(2006)R2 around the world: New theory and new tests.J. Financial Econom.79: 257-

292.

John K, LitovL,Yeung B(2008) Corporate governance and risk-taking.J. Finance63: 1679-1728.

Kang JK, StulzR(1997) Why is there a home bias? An analysis of foreign portfolio equity ownership

in Japan.J. Financial Econom.46: 3-28.

Kim JB, Li Y, Zhang L (2011a)Corporate tax avoidance and stock price crash risk: Firm-level

analysis.J. Financial Econom.100: 639-662.

Kim JB, Li Y, Zhang L (2011b) CFOs versus CEOs: Equity incentives and crashes. J. Financial

Econom.101: 713-730.

KirschenheiterM, MelumadND (2002) Can “big bath” and earnings smoothing co‐exist as equilibrium

financial reporting strategies?J. Accounting Res. 40: 761-796.

http://rady.ucsd.edu/docs/faculty/Innovative-Efficiency-and-Stock-Returns.pdf

56

Kothari SP, Shu S, WysockiPD(2009) Do managers withhold bad news? J. Accounting Res.47: 241-

276.

La Porta R, Lopez-de-Silanes F, Shleifer A, Vishny R (1998) Law and finance. J. Political Economy

106: 1113-1155.

La Porta R, Lopez-de-Silanes F, ShleiferA(2006) What works in securities laws? J. Finance61, 1-32.

Leuz C (2006) Cross listing, bonding and firms’ reporting incentives: A discussion of Lang, Raedy

and Wilson (2006). J. Accounting and Econ.42: 285-299.

LeuzC, Lins K, Warnock F (2009) Do foreigners invest less in poorly governed firms? Rev. Financial

Stud.22: 3245-3285.

Low A (2009) Managerial risk-taking behavior and equity-based compensation.J. Financial Econom.

92: 470-490.

Mankiw, NG, Romer D, Weil DN (1992) A contribution to the empirics of economic growth.Q. J.

Econ.107: 407-437.

Mieno F (2009)Foreign ownership, listed status, and the financial system in EastAsia: Evidence from

Thailand and Malaysia. Working paper.

Morck R, Nakamura M (1999)Banks and corporate control in Japan. J. Finance 54: 319-339.

Myers SC (1977)Determinants of corporate borrowing.J. Financial Econom.5: 147-75.

Myers SC, MajlufNS (1984) Corporate financing and investment decisions when firms have

information that investors do not have.J. Financial Econom.13: 187-221.

Portes R, Rey H (2005)The determinants of cross-border equity flows.J. Int. Econ. 65: 269-296.

Qi Y, Roth L, Wald JK(2010) Political rights and the cost of debt.J. Financial Econom.95: 202-226.

Rajgopal S, Shevlin T (2002) Empirical evidence on the relation between stock option compensation

and risk-taking.J. Accounting and Econ. 33, 145-171.

Rossi S, Volpin PF (2004) Cross-country determinants of mergers and acquisitions.J. Financial

Econom.74: 277-304.

Smith CW, StulzRM (1985)The determinants of firms’ hedging policies.J. Financial Quant. Anal. 20:

391-405.

Stiglitz JE (2000) Capital market liberalization, economic growth, and instability. World

Development28, 1075-1086.

UNCTAD (United Nations Conference on Trade and Development) (2002, 2010, 2013) World

Investment Report.

World Bank.World Development Indicators.

Source: http://data.worldbank.org/data-catalog/world-development-indicators

Wright P, Perris S, Sarin AY, Awasthi V (1996) Impact of corporate insider, blockholder and

institutional equity ownership on firm risk-taking.Acad. Manage. J. 39: 441-463.

http://ideas.repec.org/a/bla/jfinan/v54y1999i1p319-339.html

http://ideas.repec.org/s/bla/jfinan.html

Foreign Institutional Ownership, Risk-Taking, and Crash ... · Crash Risk around the World Abstract...

Documents

Transcript of Foreign Institutional Ownership, Risk-Taking, and Crash ... · Crash Risk around the World Abstract...