Law and Executive Compensation How the Legal System Affects

Document Sample
Law and Executive Compensation How the Legal System Affects Powered By Docstoc
					   Law and Executive Compensation: How the Legal System Affects the Equity Mix in
                            Executive Compensation


We examine variation in relative use of equity-based compensation (equity mix) across firms
from different legal environments by studying 381 non-U.S. firms from 43 countries during the
1996-2000 period. These firms are from countries that provide varying degrees of legal
protection for shareholders. The data indicate association between equity mix and the degree of
legal protection of shareholder rights. Specifically, firms use relatively more equity-based
compensation if in a legal environment where shareholder rights are more strongly protected and
where laws are more effectively enforced. These findings add to the literature demonstrating a
relationship between institutional factors and financial decisions.




Stephen Bryan is a Professor of Accounting in the Graduate School of Business and the College
of Business Administration at Fordham University, New York, NY 10023.

Robert Nash is an Associate Professor of Finance in the Babcock Graduate School of
Management at Wake Forest University, Winston-Salem, NC 27109-7424.

Ajay Patel is a Professor and GMAC Chair in Finance in the Babcock Graduate School of
Management at Wake Forest University, Winston-Salem, NC 27109-7424.
       A growing stream of research has linked financial decisions to institutional settings.

Utilizing primarily the models of LaPorta, Lopez-de-Silanes, Shleifer, and Vishny (1997), this

body of work has identified that legal and other institutional characteristics are related to most of

the firm’s major financial policies. For example, the extant literature has established that legal

characteristics are related to decisions regarding dividend payout (LaPorta, Lopez-de-Silanes,

Shleifer, and Vishny, 2000), capital structure (Bancel and Mittoo, 2004; Giannetti, 2003; Booth,

Aivazian, Demirguc-Kunt, and Maksimovic, 2001; Demirguc-Kunt and Maksimovic, 1999;

Rajan and Zingales, 1995), financing (Denis and McConnell, 2003; Lang, Lins, and Miller,

2003; Licht, 2003; Johnson, McMillan, and Woodruff, 2002; Reese and Weisbach, 2002;

Modigliani and Perotti, 2000; Rajan and Zingales, 1998; Demirguc-Kunt and Maksimovic,

1998), derivatives usage (Bartram, Brown, and Fehle, 2007), restructuring (Claessens and

Klapper, 2005), type of management (Denis and McConnell, 2003; Anderson, Lee, and Murrell,

2000), investment (Carlin and Mayer, 2003; Dittmar, Mahrt-Smith, and Servaes, 2003), and

ownership structure (LaPorta, Lopez-de-Silanes, Shleifer, and Vishny, 2002; LaPorta, Lopez-de-

Silanes, Shleifer, and Vishny, 1998, LaPorta et al., 1997; Zingales, 1995). However, we know

relatively little about how the institutional environment, particularly the legal system, affects the

design of compensation contracts. Accordingly, we extend the analysis of the impact of

institutional differences by considering whether these factors also affect the structure of

management compensation.

       Surveys by major human resource consulting firms document distinct variation in

compensation structure across countries. Aware of enduring differences in compensation

contracting, Matthews (2006) points out that a “global philosophy [of uniform executive

compensation] does not mean global homogeneity; it is an overarching framework that typically




                                                                                                     2
allows for different applications in different locations or for different executive and/or employee

populations.” Towers Perrin (2006) contends that these differences in compensation design will

also likely persist to the extent that the underlying factors endure as well (such as a country’s

local institutions). Therefore, we examine how institutional environment affects compensation

structure.

       In this study, we focus on the equity mix (i.e., relative amount of equity-based

compensation) used by firms from all major legal systems. We document considerable variation

in compensation structure and identify factors significantly associated with differences in the use

of equity-based pay. Specifically, we find strong evidence that institutional factors are significant

determinants of compensation structure. Firms from countries that provide greater protection of

shareholder rights use a larger relative amount of equity-based compensation. Also, the equity

mix is higher if the country’s legal system ensures strict enforcement of laws. This is consistent

with prior studies that demonstrate the important effect of institutional factors on corporate

finance decisions.

       While we find that institutional factors are the most significant determinants of

compensation structure, we also document that firm-level factors (such as the prevalence of

growth options and the use of leverage) affect the compensation mix. These enterprise-specific

results are consistent with the predictions of contracting theory.

       We organize the remainder of this paper as follows. Section I specifies our hypotheses

regarding potential institutional determinants of compensation structure. Section II defines

control factors. Our control factors are primarily firm-level variables frequently identified as

agency-based determinants of compensation structure. Section III describes our data sources and

our sample, and provides descriptive statistics for the variables in our models. Empirical findings




                                                                                                     3
are located in Section IV, while Section V documents the results of our robustness tests. Our

summary and conclusions are in Section VI.



I. Potential Institutional Determinants of Compensation Structure

         Our primary hypothesis is that institutional characteristics (especially legal factors)

should significantly affect differences in compensation structure. The following section describes

why the use of equity-based compensation is expected to differ across legal environments. We

base our hypothesis on extensive literature that has shown that institutional factors affect

corporate contracting decisions. These studies (LaPorta et al., 1997, 1998) focus primarily on the

impact of differences in legal systems.

         As described by David and Brierley (1985), most commercial law is derived from the two

broad traditions of common law or civil law. Common law is primarily determined by judges

where laws are formulated through precedent and subsequently incorporated into the legislature.

Common law is based on English tradition and, like the other systems, was spread across the

world mostly through occupation and colonization. Civil law relies on statutes and

comprehensive codes (primarily articulated by legal scholars and governmental authorities).

Civil systems, drawing from the principles of Roman law, can be primarily arrayed into three

main families: 1) French, 2) German, and 3) Scandinavian. Additionally, the socialist legal

tradition, based on the law of the former Soviet Union, is closely related to the civil law

systems.1




1
  LaPorta, Lopez-de-Silanes, Shleifer, and Vishny (1999) designate socialist legal systems as another form of civil
law. Countries that LaPorta et al. (1999) classify as having a socialist legal tradition are those previously under the
influence of the Soviet Union or those with communist political orientations. Of our sample firms, only China and
Russia fall under the “socialist” legal origin.


                                                                                                                          4
       An important area of difference between legal systems is in the protection of shareholder

rights. Since stockholders can exert influence through the voting mechanisms, LaPorta et al.

(1997, 1998) primarily evaluate shareholder rights by gauging the legal protection of voting

procedures. LaPorta et al. (1997, 1998) explicitly measure shareholder protection by forming an

index based on whether or not certain legal mechanisms are in place to ensure shareholder rights.

Specifically, the Antidirector Rights Index (ARI) of LaPorta et al. (1997, 1998) measures the

amount of voting powers possessed by stockholders and the strength of the legal support of

shareholder rights. The ARI takes on a higher value as the legal protection of stockholder rights

increases.

       LaPorta et al. (1997, 1998) reveal that shareholder protection varies significantly among

the legal families. In general, the ARI indicates that common law systems provide significantly

stronger protection of shareholder rights than do any of the civil law families. Additionally,

although the English common law system provides significantly stronger protection overall,

LaPorta et al. (1997) note that stockholder rights may vary within each legal system. In our

empirical analysis, we focus on how variation in shareholder legal protection affects the structure

of managerial compensation across countries.

       We expect a greater use of equity-based compensation in countries providing stronger

protection of shareholder rights. This is primarily because equity-based compensation will be

most effective when stock prices are informationally efficient. Specifically, stock prices must be

informationally efficient to provide an accurate retrospective of firm performance and a

reasonable means of motivating and compensating managers. As described by Baker, Jensen,

and Murphy (1988) and Holmstrom (1979), compensation should not be based on factors beyond

the control of managers. Accordingly, stock prices that provide a more accurate retrospective of




                                                                                                    5
managerial performance would be more effectively utilized in managerial compensation. If the

stock market is not informationally efficient, share price has diminished value for monitoring and

for linking pay to performance.2

        Nations providing stronger protection of shareholder rights have stock markets that are

more informationally efficient. Specifically, studies by Morck, Yeung, and Yu (2000), Dow and

Gorton (1997), and Vickers and Yarrow (1991) contend that greater protection of shareholder

rights increases the informational efficiency of stock prices. Furthermore, Ball, Kothari, and

Robin (2000) and Ali and Hwang (2000) identify that legal institutions significantly affect the

value relevance of accounting information. These studies find that, in countries with common

law legal systems (stronger legal protection), accounting information is more timely, is followed

more closely by analysts, and is more powerful in explaining security returns. This stronger

connection between accounting information and firm value in common law countries improves

informational efficiency, facilitates monitoring, and provides a more direct link between stock

price and firm performance. Accordingly, equity-based compensation should be more effective

and should be more widely used in strong protection countries.3

        Additionally, Holmstrom and Tirole (1993) and Bhushan (1989) contend that market

liquidity affects the informational efficiency of stock prices. As LaPorta et al. (1997, 1998)

demonstrate, nations providing weaker protection of shareholder rights have stock markets with

highly concentrated ownership and, therefore, limited liquidity. This provides additional


2
  Baker, Gibbons, and Murphy (1994) further contend that variances in contract structure (i.e., differences in the
equity mix) can be linked to the level of distortion in performance measurement. As a result, in economies where
stock price is a less accurate gauge of managerial performance, we should expect a lower relative use of equity-
based compensation.
3
  In legal environments with weaker protection for shareholders, Gorton and Schmid (2000) argue that bank
relationships substitute for the stock market in providing a retrospective on firm performance. That is, in weak
protection countries, bank monitoring may be relied upon to mitigate agency problems. Thus, we expect lower
equity-based compensation in countries with bank dominated financial systems. Our findings are consistent with this
expectation.


                                                                                                                 6
theoretical and empirical justification for our expectation that the equity mix should be positively

related to the strength of legal protection of shareholder rights.

        Therefore, we expect stronger investor protection to lead to higher use of equity

compensation, either because greater investor protection makes stock prices more

informationally efficient or because it leads to lower ownership concentration. However, in the

following empirical analysis, we are unable to explicitly identify which of the two mechanisms

causes the positive relation between investor protection and the use of equity-based

compensation.

        Our primary indicator of legal protection is the LaPorta et al. (1997) Antidirector Rights

Index (ARI). In our robustness analysis (Section V), we use alternative measures of shareholder

legal protection. Additionally, in all models, we include the Rule of Law Index (LaPorta et al.,

1997). An indicator of law enforcement quality, the Rule of Law Index measures the

effectiveness of law enforcement in each country.4 This index takes on a higher value for nations

with a stronger tradition of law and order.



II. Control Factors: Firm-Specific Determinants of Compensation Structure

        So that we can focus on the impact of legal environment, we control for enterprise-level

factors frequently cited as determinants of compensation structure. Specifically, previous studies

of management compensation (primarily in the U.S.) emphasize the role of firm-specific agency

costs. As summarized below and by Bryan, Hwang, and Lilien (2000) and Yermack (1995), the

extant literature generally concludes that U.S. firms develop compensation contracts to motivate



4
 In empirical studies regarding how the institutional environment affects corporate finance decisions, Durnev and
Kim (2005), Giannetti (2003), and Pistor, Raiser, and Gelfer (2000) find that the quality of law enforcement is a
highly influential factor.


                                                                                                                    7
managers to maximize shareholder wealth and minimize conflicts of interest within the firm.5

Therefore, while we emphasize the effect of institutional factors, our models include variables to

test whether the same firm-level agency problems that influence the structure of U.S.

compensation contracts also impact the compensation design of non-U.S. firms.



A. Agency Costs of Equity and Compensation Structure

         Most studies of the determinants of U.S. compensation structure have focused on the role

of agency costs of equity. When assessing the potential severity of stockholder/manager conflicts

within a firm, the contracting literature primarily examines enterprise-specific characteristics

(i.e., the firm’s growth options, size, and free cash flow).

         Authors such as Bryan et al. (2000), Kole (1997), Bizjak, Brickley, and Coles (1993), and

Gaver and Gaver (1993) contend that firms with greater amounts of growth options have broader

informational asymmetries that create a larger potential for opportunistic behavior by managers.

As a result, these firms should use more equity-based compensation. Our proxy for the

prevalence of growth options is the ratio of the market value to the book value of the firm’s

assets (Market-to-Book). This ratio is equal to the book value of assets minus the book value of

equity plus the market value of equity divided by the book value of assets.

         Also, Jensen and Meckling (1976) contend that agency costs increase with firm size since

a larger span of operations allows for greater managerial opportunism and contributes to less

effective external monitoring. Ryan and Wiggins (2002), Bryan et al. (2000), Kole (1997),



5
  This hypothesis holds that larger amounts of equity-based compensation will cause managers to act more like
stockholders. This would lower the agency costs of equity, but would potentially increase the agency costs of debt.
Alternatively, greater amounts of cash compensation will encourage managers to act more like bondholders
(reducing the agency costs of debt, but contributing to higher agency costs of equity). These countervailing effects
of cash and stock-based compensation lead John and John (1993) to conclude that the optimal equity mix should
balance these agency costs and, thus, minimize the firm’s total agency costs.


                                                                                                                       8
Mehran (1995), Yermack (1995), and Gaver and Gaver (1993) find that bigger firms pay

managers with significantly larger relative amounts of equity-linked compensation. These

authors also attribute this relation to the greater degree of difficulty in monitoring managers of

larger companies. As used by Gabaix and Landier (2008) and Baker and Hall (2004), our proxy

for size is the natural logarithm of the firm’s market capitalization. We predict a positive relation

between firm size and the relative use of equity-based compensation.

        Finally, Jensen (1986) argues that larger amounts of excess cash leads to more severe

agency problems as discretionary cash is more likely to be misinvested or lost through

organizational inefficiencies. Zhang (2008) and Broussard, Buchenroth, and Pilotte (2004)

contend that a stronger link between managerial and shareholder wealth may help mitigate this

“free cash flow problem”.

        A countervailing argument is that excess cash may be negatively associated with equity-

based pay, particularly stock options. This is because stock options’ payoff function may

encourage excessive risk-taking and, therefore, encourage managers to gamble with excess cash.

Additionally, Bryan et al. (2000) identify a negative correlation between cash flow and equity

compensation and argue that firms with low cash flow may rely more heavily on stock option

awards to conserve cash.

        Since the predictions regarding free cash flow are unclear, we test this relation with the

empirical analysis in the following sections. To measure the firm’s cash flow, we use the Lehn

and Poulsen (1989) cash flow statistic. This proxy for the firm’s free cash flow is: operating

income before depreciation less income tax less interest minus dividends paid. The measure of

free cash flow divided by the firm’s market value of equity provides an indication of cash

availability.




                                                                                                     9
B. Agency Costs of Debt and Compensation Structure

         Other papers have recognized the potential impact of the agency problems of debt on the

firm’s choice of managerial compensation (Bryan, Nash, and Patel, 2006; Ryan and Wiggins,

2002; John and John, 1993). The most prominent agency problems of debt are underinvestment

and asset substitution.6 Begley and Feltham (1999), Yermack (1995), Bizjak et al. (1993), and

John and John (1993) contend that greater amounts of equity-based compensation will

exacerbate these agency problems of debt. Furthermore, both underinvestment and asset

substitution become more severe as the firm increases leverage. Accordingly, we expect a

negative relation between a firm’s leverage and its use of equity-based compensation. Our proxy

for leverage is the ratio of the book values of the firm’s total debt to total assets.



III. Data and Descriptive Statistics

         The following section describes the data that we use to measure and explain patterns in

the design of compensation contracts. To obtain detailed compensation information for a wide,

cross-section of countries, we examine the financial disclosures of 381 firms (from 43 countries)

that issued ADRs during 1996-2000. As we will describe below, these ADR issuers must provide

extensive documentation of compensation structure. Therefore, examining ADR issuers is an

effective way to acquire executive compensation data covering a wide sample of non-U.S. firms.7


6
  Myers (1977) identifies a potential underinvestment problem for levered firms. Managers may forgo wealth
increasing projects (i.e., underinvest) if a disproportionate share of the project’s incremental value accrues to the
bondholder. Additionally, in a levered firm, stockholders may expropriate wealth from debtholders by switching
from safer to riskier investments. This is the asset substitution problem.
7
  For a broad cross-section of countries, detailed and consistently presented compensation data are simply
unavailable in any other way. As in most empirical work, there are trade offs involved in the selection of data
sources. For example, Desai, Foley, and Hines (2004), Booth et al. (2001), Rajan and Zingales (1995), and
Megginson, Nash, and Van Randenborgh (1994) note that using cross-country data may lead to measurement error
when comparing countries (due to inconsistencies in accounting practices). However, accounting system differences
are not a problem in our data because all ADR issuers must report standardized financial statements (Form 20-F).



                                                                                                                  10
A. Data Sources

        A foreign company may sell equity, or have its equity traded, in the U.S. via American

Depository Receipts (ADRs). Sponsored ADRs (those managed by a depository bank) are

classified as Level 1, Level 2, Level 3, or Rule 144A. Level 1 ADRs are traded over-the-counter

and must be registered with the SEC, but they are exempt from U.S. reporting requirements

under Rule 12g3-2(b). Rule 144A ADRs are private placements and are also exempt from SEC

reporting requirements. Level 2 ADRs are listed on an exchange or quoted on NASDAQ. This

type of ADR must be registered and foreign companies must file Form 20-F, which contains

detailed financial statements and a reconciliation of foreign GAAP to U.S. GAAP.8 Level 3

ADRs are for new equity offerings. Level 3 ADRs must be registered and foreign companies

must also file Form 20-F.

        We searched Compustat for a listing of all ADRs to identify non-U.S. companies with

ADR programs (Level 2 and Level 3). We then hand-collected the firm’s Form 20-F filings for

years 1996-2000 from the Securities and Exchange Commission’s EDGAR database.



B. Primary Components of Compensation Structure

        We use these data to analyze differences in compensation structure across countries.

Compensation structure refers to the firm’s relative use of stock-option, restricted stock, or cash-

based compensation.

        We measure the value of options granted using the Black-Scholes (1973) model, which

requires an estimate of risk-free interest rates, expected stock return volatility, expected dividend


8
  In November 2007, the U.S. Securities and Exchange Commission did adopt a provision to allow non-U.S.
registrants to file their reports without reconciliation if they used International Financial Reporting Standards
promulgated by the International Accounting Standards Board.



                                                                                                              11
yield, and expected time to maturity. We obtain country-specific risk-free rates from the IMF’s

International Financial Statistics Yearbook. We calculate dividend yields from Compustat data

and stock return volatility from the Center for Research in Securities Prices (CRSP) database. As

is standard in the compensation literature (Yermack, 1995; Bryan et al., 2000), the option

expected time to maturity is 10 years for all firms, unless stated otherwise.

       The 20-F filings also provide the value of any grants of restricted stock. As suggested by

its name, restricted stock has restrictions on resale or transfer until vesting. Accordingly,

restricted stock can be viewed as a stock option with a zero strike price (Berger, Ofek, and

Yermack, 1997). However, unlike stock options whose payoff function is convex in stock price,

restricted stock has a linear payoff schedule (because of the zero exercise price). Thus, risk-

averse managers who are paid in restricted stock bear the potential wealth loss from risky

investments.

       Finally, the 20-F specifies the amount of cash-based compensation received by the

executive. It also discloses any other form of managerial compensation (such as payout from

long-term incentive plans or pay in the form of benefits and perquisites). We include all of these

components of managerial remuneration in our various models of compensation structure.

       Our primary measure of compensation structure (PCTEQ) focuses on the relative use of

equity-based awards. Equity-based compensation can consist of both stock options and restricted

stock. Accordingly, PCTEQ is the sum of the Black-Scholes (1973) option value compensation

and restricted stock compensation divided by total compensation. Total compensation is the sum

of option-based compensation, restricted stock, long-term incentive plans, other compensation,

and cash compensation. Therefore, our primary proxy for compensation structure is:




                                                                                                  12
       PCTEQ = Value of Equity-Based Compensation / Total Compensation



       We use another indicator of compensation design that focuses on option-based

compensation. Stock option compensation is the dominant form of equity-based compensation in

the sample. Our proxy for option-based compensation is:



       PCTOPT = Value of Option Compensation / Total Compensation



       Finally, to complement our measure of option-based compensation, we also calculate the

relative use of restricted stock. This proxy is:



       PCTREST = Value of Restricted Stock Compensation / Total Compensation



       We use Compustat data to form the firm-specific explanatory variables in our empirical

models. Our final sample covers 1996-2000 and consists of 1,022 firm-year observations from

43 non-U.S. countries. These nations are from all of the major legal systems and provide widely

varying degrees of legal protection for shareholders.



C. Descriptive Statistics

       Table I reveals that while 43 countries are represented, the data are clustered around

certain countries. These nations either: 1) have capital-market orientations that are similar to that

of the U.S. (e.g., England), or 2) are geographically close to the U.S. (e.g., Mexico). This finding

is consistent with the evidence in Pagano, Roell, and Zechner (2002) and Tesar and Werner




                                                                                                   13
(1995). Additionally, from Table I, we report that the number of firm years (for which we have

data) steadily increased throughout our sample period.

        Table II presents the averages of our measures of compensation structure by country from

1996-2000. These data provide an early indication that institutional factors may help explain the

cross-country variation in the use of equity-based compensation. We draw on the findings of

LaPorta et al. (1997, 1998) who note that shareholders in countries with an English common law

tradition benefit from much stronger legal protection than those living in nations with other legal

systems. We identify that the average equity mix (PCTEQ) is high for firms in countries with

English legal origin [e.g., England (19.7%), Australia (31.2%), New Zealand (42.3%) and South

Africa (30.7%)], while it is very low for firms in countries with German legal origin [e.g.,

Germany (4.6%) and Switzerland (4%)]. Firms from countries with a Socialist legal origin

(China and Russia) use virtually no equity in the compensation mix. Moreover, in 18 of our

sample countries (42%), no equity-based compensation was paid. All 18 of these countries have

a non-English legal system.9 Therefore, this early evidence suggests that institutional factors

relating to the country’s legal system may contribute to cross-country differences in the use of

equity-based compensation.10



                                          Insert Table II about here.
9
  Of the eight nations with the largest relative use of equity-based compensation, seven have an English common
law system. The exception (Netherlands), while operating under a French civil law system, is noted by Ali and
Hwang (2000) and Ball et al. (2000) as having an institutional environment very similar to that of the common law
countries. Additionally, in a multicountry survey of CEOs, Brounen et al. (2004) find that Dutch firms (unlike
counterparts in Germany and France) exhibit a strong shareholder orientation. Anecdotally, it is also not surprising
that the country that was home to the first modern shareholders (Dutch East India Company) would have a relatively
strong preference for the use of equity in compensation structures.
10
   We acknowledge that the pay practices of our sample firms may not be entirely representative of the average firm
in the respective home countries. However, our intent is not to draw conclusions about whether our sample is
representative of all firms “back home”. Rather, our purpose is to identify whether compensation practices differ
because of regional cultures and norms that should be reflected in local institutional factors. Furthermore, ADR
issuers may share common traits and may be more similar to U.S. firms (than to other firms in the home country).
These factors should bias against findings of cross-country differences in compensation structure.


                                                                                                                 14
       Table III allows us to better understand the behavior of compensation structure across

each of the five years of our sample period. We specifically measure the relative use of all

equity-based compensation (PCTEQ) by non-U.S. firms from 1996-2000. We also separate the

PCTEQ into its two components: 1) PCTOPT (the relative amount of option-based

compensation) and 2) PCTREST (the relative amount of restricted stock compensation).

Throughout the late 1990s, non-U.S. firms use a generally increasing amount of equity-based

compensation. This upward trend in the use of equity-based compensation is broadly consistent

with the time series pattern of compensation structure for U.S. firms (as documented by Bryan et

al., 2006). However, the use of equity-based compensation is considerably lower in the non-U.S.

firms. Additionally, for each of the years, Table III indicates that a vast majority of the equity-

based compensation of non-U.S. firms is in the form of options. Across the sample period, non-

U.S. compensation structures have consistently involved a very small relative amount of

restricted stock. Therefore, while equity-based compensation may be in the form of both stock

options and restricted stock, Table III reveals that non-U.S. firms primarily provide option-based

compensation. Specifically, of all equity-based compensation paid by non-U.S. firms,

approximately 74% consists of stock options.



                                     Insert Table III about here.



       Panel A, Table IV, presents descriptive statistics for the variables sorted by the

Antidirector Rights Index (ARI). The data in Panel A (Table IV) provide support for our general

hypothesis. When ARI is low (0 or 1), the use of equity-based compensation is low (1.7% and




                                                                                                      15
3.7%). However, when shareholder rights are protected strongly (ARI of 4), the use of equity-

based compensation increases to 15.9%.

       Panel B, Table IV, provides descriptive statistics when we sort by origin of legal system.

The five legal origins are 1) English, 2) German, 3) French, 4) Scandinavian, and 5) Socialist.

Panel B indicates that approximately 70% of our ADR issuing firms are from countries with non-

English legal systems. Since a majority of our sample firms are from countries generally

providing weaker shareholder protection (i.e., non-English legal systems), it is not surprising that

a majority of our firms provide little or no equity-based compensation (as indicated by the

quartile values for PCTEQ).

       In Panel B, the data also indicate that firms from countries with an English legal origin

(stronger protection for shareholders) use a greater amount of equity-based compensation. In

contrast, firms from countries with civil law origins (weaker protection of shareholders) use less

equity-based compensation. Not surprisingly, the firms from our countries of Socialist legal

origin use virtually no equity-based compensation. All of these findings are consistent with our

general hypothesis.

       As further indication of how institutional factors affect compensation structure, Panel C

presents descriptive statistics based on effectiveness of law enforcement. We array the data

according to values of the nation’s Rule of Law index (e.g., Quartile 1 contains nations

considered to have the least effective law enforcement). As expected, firms in nations providing

higher quality law enforcement use greater relative amounts of equity-based compensation.

       This preliminary analysis indicates a strong relationship between institutional factors and

the compensation structure of non-U.S. firms. We next use a multivariate framework to further

assess the impact of institutional characteristics on the design of compensation contracts (while




                                                                                                   16
simultaneously controlling for firm-specific factors).



                                           Insert Table IV about here.



IV. Determinants of Compensation Structure

         The following sections describe our empirical findings regarding determinants of

differences in the use of equity-based compensation. Our regressions include industry and year

dummy variables. We use a Tobit model as the data are left censored. Dependent variables are

measures of compensation structure; independent variables are measures of institutional

environment. We also include firm-specific, agency-based control variables. As is standard

practice with panel data, the standard errors are corrected for heteroscedasticity and serial

correlation. The results based on these robust standard errors are presented in Table V.



                                            Insert Table V about here.



A. Institutional Environment and Compensation Structure

         We focus on the characteristics of the issuer’s home legal system. While listing on a U.S.

exchange may contribute to improved disclosure (Lang et al., 2003; Reese and Weisbach, 2002;

LaPorta et al., 2000), foreign issuers are primarily governed according to the corporate laws of

their country of incorporation (Licht, 2003).11 Accordingly, we target the legal environment of



11
   Specifically, Licht (2003) identifies that the corporate laws of the destination country have less significant impact
on the foreign issuer. He notes that critical components of the U.S. institutional framework do not apply to foreign
issuers. For example, ADR issuers are exempt from many stock market listing requirements, are excluded from
Regulation FD, and are not subject to certain measures of the Sarbanes-Oxley Act. This leads Durnev and Kim
(2005) to conclude that ADR listing does not substantially affect investor protection. Therefore, as noted by Siegel
(2005) and Licht (2003), the act of cross-listing is not a remedy for weak legal protection.


                                                                                                                      17
the issuer’s home country as a potential determinant of its compensation structure.

         To assess the impact of the legal system on compensation structure, we focus on the

protection of shareholder rights. We gauge the degree of protection of shareholder rights with the

Antidirector Rights Index (ARI). The findings in Table V (Panel A) provide strong evidence of a

positive association between the strength of investor rights and the relative use of equity-based

compensation. We first focus on Models 1-3, where the dependent variables are our primary

measures of compensation structure (PCTEQ, PCTOPT, and PCTREST). In each model, the

coefficients for the legal environment variables are significantly positive at the 1% level.

         We also expect that the use of equity-based compensation should be positively related to

the quality of law enforcement in a country. The findings in Table V are consistent with our

hypothesis. The coefficient of the Rule of Law variable is significantly positive in all regressions

involving equity-based and option-based compensation.

         Overall, the findings in Models 1-3 (Table V, Panel A,) provide strong support for our

general hypothesis that institutional factors substantially affect compensation structure. That is,

the use of equity-based compensation is significantly related to the strength of shareholder rights

and the quality of law enforcement in each country. The results for institutional factors are robust

to model specification, as all institutional variables are highly significant in our models. 12



B. Agency Costs and Compensation Structure

         While focusing on the institutional environment, we control for the effect of firm-

12
  We note that certain institutional characteristics are difficult to measure directly. For instance, Japan and South
Korea prohibited the granting of stock options until 1997 (Wahlgren, 2001). Although we are unable to measure
these attributes directly, we argue that aspects of each are captured in, or at least correlated with, the institutional
variables that we observe (such as the ARI). To the extent such attributes affect executive compensation and are
uncorrelated with our institutional variables, they represent potential omitted variables. The need for such a caveat is
an unfortunate tradeoff that we make in order to use a large sample of countries. Country-specific studies, such as
Kato, Lemmon, Lou, and Schallheim (2005), Elston and Goldberg (2003), and Kraft and Niederprum (1999) can
provide a more detailed analysis of the institutional factors unique to an individual country.


                                                                                                                     18
specific, agency-based factors. The U.S. compensation literature (e.g., Yermack, 1995) has

shown these variables to be strongly related to compensation design. However, there is little

cross-sectional evidence as to whether these same agency-based factors similarly affect non-U.S.

compensation structures. Table V presents our findings. Our proxy for growth opportunities

(Market-to-Book) is significantly positively as related to the use of equity in the compensation

mix. Additionally, leverage (measuring the prevalence of the underinvestment and asset

substitution problems) has a significantly negative effect on the amount of equity-based

compensation used by non-U.S. firms. These findings are consistent with prior studies of U.S.

firms and provide evidence that non-U.S. firms also consider agency problems when designing

compensation contracts.

       In general, the results suggest that agency problems and the contracting solutions to

mitigate them transcend national boundaries (although the explanatory power is not as significant

when applied to non-U.S. firms). Nevertheless, our data indicate that institutional factors are the

most significant determinants of the compensation structures of non-U.S. firms.



V. Robustness Tests

       The following sections describe additional tests to confirm the robustness of our primary

findings.



A. Alternative Measures of Compensation Structure

       We repeat our regressions using alternative measures of compensation structure. These

measures of compensation mix are, respectively, the ratios of the values of equity compensation,

option compensation, and restricted stock compensation to cash compensation (EQMIX,




                                                                                                   19
OPTMIX, and RESTMIX). The regression results using these alternative measures of

compensation structure are in Models 4-6 of Table V. The Tobit regressions continue to include

industry and year dummy variables. The standard errors and significance values are also robust

to serial correlation and heteroscedasticity. The findings are essentially unchanged from those

using our primary measures of compensation structure. All institutional variables (strength of

shareholder legal protection and strength of the nation’s Rule of Law) are significantly positively

related to the use of equity-based compensation. Also, similar to Models 1-3, several of the firm-

specific factors are significantly related to the compensation structure of our sample firms.



B. Alternative Measures of Legal Environment

       Consistent with the preponderance of the law and finance literature, we primarily

measure legal protection of shareholder rights with the Antidirector Rights Index (ARI) of

LaPorta et al. (1997, 1998). As noted in Section IV.A and in Panel A of Table V, the ARI is

positively and significantly related to the use of equity-based compensation. To help confirm the

robustness of this result, we prepare additional models including the following alternative

measures of legal environment. Panels B-D of Table V present our findings.



1. Revised Antidirector Rights Index

       While the LaPorta et al. (1997, 1998) Antidirector Rights Index has been widely used as

a measure of legal environment, criticisms have been leveled by Pagano and Volpin (2005)

regarding conceptual ambiguities and coding mistakes. In response to these concerns and in an

attempt to define corporate law with better precision, Djankov, LaPorta, Lopez-de-Silanes, and

Shleifer (2008) develop a revised Antidirector Rights Index. The revised Antidirector Rights




                                                                                                  20
Index is highly correlated with the original index, but is designed to overcome the weaknesses

identified by Pagano and Volpin (2005).

          The models in Panel B include the revised Antidirector Rights Index as the indicator of

shareholder legal protection. As with the original index, we find that the revised Antidirector

Rights Index is a highly significant determinant of compensation structure. All other variables

also retain their significance levels.



2. Anti Self-Dealing Index

          The Anti Self-Dealing Index is another measure of legal protection developed by

Djankov et al. (2008). This index focuses on shareholder protection against expropriation by

insiders. The index explicitly considers such control mechanisms as legally mandated disclosure

and approval processes, as well as procedural remedies and penalties if minority shareholders are

wronged. As described by Djankov et al. (2008) and Johnson, LaPorta, Lopez-de-Silanes, and

Shleifer (2000), insider self-dealing (i.e., expropriation by majority owners or “tunneling”) is a

major concern to minority shareholders. By directly targeting the level of legal protection against

this risk, the Anti Self-Dealing Index appears to be an especially relevant measure of shareholder

rights.

          Panel C presents our results after including the Anti Self-Dealing Index as a measure of

shareholder legal protection. As with all of our legal variables, the Anti Self-Dealing Index is

highly significant in all models.



3. Legal Origin: Common Law vs. Civil Law

          For our final measure of the legal environment, we use an indicator variable to




                                                                                                     21
distinguish nations with a common law legal tradition from those with a civil law system.

LaPorta et al. (1997, 1998) document that common law nations generally provide for stronger

protection of investor rights. We include this indicator variable in the models of Panel D. The

indicator variable (i.e., one if the country has a common law legal system) is positive and

significant in all models. Consistent again with our expectations, this confirms that firms use a

greater amount of equity-based compensation in nations with stronger protection of shareholder

rights.

          Overall, the findings in Panels A-D of Table V provide strong and consistent evidence in

support of our general hypothesis. That is, regardless of how it is measured, the legal

environment is a highly significant factor in the structure of managerial compensation.



C. Alternative Measure of Institutional Environment: Political Factors

          In addition to the legal environment, other types of institutional variables may also

impact compensation structure. Studies by Dyck and Zingales (2004) and Stulz and Williamson

(2003) identify that financial decisions may be affected by “extra-legal” characteristics. As

defined by Dyck and Zingales (2004), these extra-legal factors refer to social or cultural norms as

opposed to written or regulatory norms (such as those that we measure with the legal variables).

LaPorta, Lopez-de-Silanes, and Shleifer (2006), Baker and Hall (2004), Conyon and Murphy

(2000), Cheffins (1997), and Jensen and Murphy (1990) specifically consider how extra-legal

characteristics may affect managerial compensation.

          LaPorta et al. (2006) and Jensen and Murphy (1990) identify that a nation’s political

environment is an important extra-legal factor that may affect major financial decisions (such as

compensation policy). To further test the association between political system and compensation,




                                                                                                    22
we include a measure of political environment (the “polity” variable) in Model 1 of Table VI.

We obtain this variable from the Polity Database.13 The polity variable ranges from +10 to -10

with higher values indicating a more democratic government and lower values indicating a more

autocratic government. We argue that, relative to an autocracy, a society that is more democratic

will attach more value to individual effort, and this may facilitate a greater use of equity-based

pay. That is, a society with a democratic political system (where the individual is more

respected) may be more tolerant of compensation systems that reward individual achievement.

This conjecture is supported by Siegel et al. (2008), which contends that highly democratic

societies (such as the U.S.) are more accepting of compensation systems where individual talents

and accomplishments are rewarded.



                                          Insert Table VI about here.



        The results from Model 1 confirm this expectation. The data indicate a significantly

positive correlation between the polity variable and the use of equity-based compensation. We

note that the polity variable is correlated with legal environment. Nevertheless, this finding

provides further evidence of how the extra-legal institutional environment affects contracting

decisions. A deeper analysis of the influence of such factors is an area for subsequent research.

The more important message from Model 1 is that the legal variables (especially the ARI)

remain highly significant after controlling for political environment.



D. Alternative Control Variables


13
   This database is widely used in the political science and development economics literatures. The Polity Database
is prepared and maintained by the CIDCM at the University of Maryland.


                                                                                                                 23
        In the following robustness tests, we consider alternative non-institutional factors which

may affect the structure of compensation contracts.



1. Managerial Labor Market

        Giannetti (2007) demonstrates that the size of the country’s managerial labor market may

affect compensation contracts. Giannetti contends that “CEO job-hopping” is more likely in a

country with a larger managerial labor market.14 To combat short-termism by managers in these

markets, she argues that compensation contracts may be structured to encourage managerial

retention (e.g., by providing larger relative amounts of non-vested options and restricted stock).

Therefore, differences in the dynamics of the managerial labor markets may help explain cross

country differences in compensation. Specifically, larger amounts of equity-based compensation

may be used in countries with a deeper managerial labor market.

        In Models 2 and 3 of Table VI, we add proxies for the depth of the labor market within

each country. Model 2 includes the number of firms that are listed on each national stock

exchange. We argue that a larger number of publicly traded firms is indicative of a deeper labor

market (offering more lucrative opportunities for “job hopping” by managers in that country).

Our results in Model 2 are consistent with the Giannetti (2007) prediction. The data indicate that

a larger labor market (as evidenced by more publicly traded companies) is associated with a

greater relative amount of equity-based compensation.

        Model 3 contains an alternative measure of labor market structure: the proportion of

family-owned enterprises within the economy. This variable is from LaPorta, Lopez-de-Silanes,

and Shleifer (1999). An economy with more family control suggests fewer opportunities for

14
   Gabaix and Landier (2008) provide the related proposition that an increase in firm size expands the CEO’s outside
employment opportunities and, thus, affects the CEO’s compensation. We further evaluate the relationship between
firm size and compensation structure in Section V.D.2.


                                                                                                                 24
unrelated managers and, therefore, a thinner managerial labor market for professional (non-

family) managers. Accordingly, if a greater amount of family control leads to a smaller labor

market for non-family members, the Giannetti (2007) model predicts a lower use of equity-based

compensation in that economy. In Model 3, the data indicate the expected negative association.

This is also consistent with the empirical findings of Conyon and Schwalbach (2000) and Kole

(1997). Perhaps more importantly, after controlling for labor market size, the legal variables

remain highly significant in Models 2 and 3.



2. Firm Size

       Gabaix and Landier (2008), Edmans, Gabaix, and Landier (2007), Baranchuk,

MacDonald, and Yang (2006), and Baker and Hall (2004) describe how firm size may affect

managerial compensation. To that end, we conduct further robustness tests to determine if

alternative measures of firm size impact our models of compensation structure. In our base

models (Table V), we follow Gabaix and Landier (2008) and Baker and Hall (2004) and use

market capitalization as our indicator of firm size. In our robustness tests (Table VI), we

alternatively measure firm size with total sales and with total assets (Models 4 and 5,

respectively). Compensation studies by Baker and Hall (2004), Conyon and Murphy (2000), and

Frost and Pownall (1994) use total sales as a gauge of firm size. Baker and Hall (2004) favor this

measure of size as they contend that annual sales are a less noisy metric than is market value. We

present results (using total sales as our size metric) in Model 4 of Table VI. We also follow Ryan

and Wiggins (2002), Bryan et al. (2000), and Yermack (1995) and measure firm size with total

assets. We present these results in Model 5. As in our primary results (Table V), the alternative

size variables are never significant in any of our models. Furthermore, regardless of our measure




                                                                                                 25
of size, the institutional variables (Antidirector Rights Index and Rule of Law) remain highly

significant in all models.



E. Alternative Samples

       To further verify the robustness of our findings, we next test for consistency across

various subsamples of our data. The following sections describe the results of this analysis.



1. Influence of U.K. Firms

       As documented by Lang et al. (2003) and confirmed in Table I, a large number of ADRs

are from British firms. According to our general hypothesis, the U.K. firms (obviously from an

English Common Law country offering very effective protection of shareholder rights) should

provide large relative amounts of equity-based compensation. This contention is supported by

the findings of Conyon and Schwalbach (2000) and by the data presented in Table II.

       To verify that our conclusions are not driven by the U.K. observations, we exclude

British firms from the regressions. Results in Model 1 of Table VII reveal that our findings

continue to hold for the non-U.K. sample. Specifically, the institutional variables indicate that

the strength and quality of a nation’s legal system continue to matter in non-U.K. countries.

Therefore, our results are not driven by the U.K. firms.



2. Influence of Largest Market Firms

       Carlin and Mayer (2003), Dittmar et al. (2003), and Rajan and Zingales (1998) argue that

firms in wealthier economies, due to similar levels of capital market development, may adopt

similar financial policies. To verify that our results are not driven by the wealthiest and most




                                                                                                    26
developed nations, we remove observations from G7 countries. Examining only the non-G7

countries is important since Demirguc-Kunt and Maksimovic (1999) find that contracting

decisions vary between developed and developing economies. However, our data indicate that

the factors affecting the compensation schemes of non-G7 firms are identical to those affecting

all firms (see Model 2 of Table VII). As with the entire sample, the institutional environment

(strength and quality of the legal system) is the most significant determinant of compensation

structure in non-G7 countries. Therefore, the data indicate that determinants of compensation

structure appear to be consistent across countries of various levels of economic development.



                                    Insert Table VII about here.



3. Influence of U.S. Firms

       Our primary sample consists of non-U.S. firms. As a further robustness test and to more

thoroughly examine the correlation between the compensation structures of U.S. and non-U.S.

firms, we conduct additional regressions in which we complement our ADR sample with data for

U.S. firms. Data for the U.S. are from Execucomp and Compustat. Including all U.S.

observations increases the sample size to 3,805 firm years. When we expand the model to

include the U.S. data (Model 3), the results for our institutional variables are unchanged. All

legal environment factors are significant. This positive relation is expected since, according to

LaPorta et al. (1997, 1998), the U.S. legal system ensures very strong protection of shareholder

rights. As per our hypothesis, U.S. firms provide large amounts of equity-based compensation.

       We also find that the control variables are highly significant. After including the U.S.

observations, the levels of significance for all control variables are generally higher (than those




                                                                                                    27
obtained when using only non-U.S. data).15 Nevertheless, we primarily treat the firm-level factors

as control variables. Our focus is on the relationship between compensation structure and

institutional environment. Further analysis of how firm-level determinants affect non-U.S.

compensation structure is a subject for future research.

         In addition to this model using all of the U.S. observations, we perform regressions in

which each ADR issuer is matched to a U.S. firm (based on industry and size). We again obtain

identical results. Model 4 of Table VII presents these findings.



VI. Conclusions

         Using data from 381 firms from countries providing varying degrees of shareholder legal

protection, we investigate how legal environment affects the design of managerial compensation

contracts. Our analysis focuses on equity mix (i.e., the relative use of equity-based executive

compensation) by non-U.S. firms. We find that the primary determinants of variation in equity

mix are institutional factors related to the strength of shareholder rights and the quality of law

enforcement in each country. Moreover, proxies for the firm-specific agency costs of debt and

equity are also significant in our regression models. Our results are essentially unchanged when

we use alternative measures of compensation structure, consider various measures of legal

environment, include additional control variables, and examine multiple subsamples.

         This study contributes to the literature in several respects. Our unique database allows us

to investigate compensation design for firms from a large sample of non-U.S. countries. We

document significant differences in compensation structure across countries. Second, we identify

15
   For example, leverage is one factor that is consistently more significant when U.S. data are added. The leverage
variable was only marginally significant in the non-U.S. models. This reduced significance of impact for non-U.S.
firms suggests that agency problems of debt are less important determinants of compensation design for non-U.S.
firms. Relating back to our institutional argument, it may be that the more creditor-centric orientation of most non-
U.S. financial markets inoculates non-U.S. firms against some of the threats posed by the agency problems of debt.


                                                                                                                   28
that institutional factors are significant determinants of this variation in equity mix. Our results

are consistent with numerous studies demonstrating an association between legal environment

and financial decisions. However, this is the first evidence indicating that differences in legal

systems specifically affect the design of compensation contracts. Third, this paper finds that

agency theory tested with U.S. compensation data is also applicable to the compensation

decisions of non-U.S. firms. Confirming that agency-based effects are reasonably consistent

across countries supports our contention that institutional environment is a significant

determinant of cross-country differences in compensation structure.




                                                                                                       29
References



Ali, A. and L. Hwang, 2000, “Country-Specific Factors Related to Financial Reporting and the

Value Relevance of Accounting Data,” Journal of Accounting Research 38, 1-21.



Anderson, J., Y. Lee, and P. Murrell, 2000, “Competition and Privatization Amidst Weak

Institutions: Evidence from Mongolia,” Economic Inquiry 38, 527-549.



Baker, G. and B. Hall, 2004, “CEO Incentives and Firm Size,” Journal of Labor Economics 22,

767-798.



Baker, G., R. Gibbons, and K. Murphy, 1994, “Subjective Performance Measures in Optimal

Incentive Contracts,” Quarterly Journal of Economics 109, 1,125-1,156.



Baker, G., M. Jensen, and K. Murphy, 1988, “Compensation and Incentives: Practice vs.

Theory,” Journal of Finance 43, 593-616.



Ball, R., S.P. Kothari, and A. Robin, 2000, “The Effect of International Institutional Factors on

Properties of Accounting Earnings,” Journal of Accounting and Economics 29, 1-51.



Bancel, F. and U. Mittoo, 2004, “Cross-Country Determinants of Capital Structure Choice: A

Survey of European Firms,” Financial Management 33, 103-132.




                                                                                                30
Baranchuk, N., G. MacDonald, and J. Yang, 2006, “The Economics of Super Managers,” Indiana

University Working Paper.



Bartram, S., G. Brown, and F. Fehle, 2007, “International Evidence on Financial Derivatives

Usage,” forthcoming Financial Management.



Begley, J. and G. Feltham, 1999, “An Empirical Examination of the Relation between Debt

Contracts and Management Incentives,” Journal of Accounting and Economics 27, 229-259.



Berger, P., E. Ofek, and D. Yermack, 1997, “Managerial Entrenchment and Capital Structure

Decisions,” Journal of Finance 52, 1,411-1,438.



Bizjak, J., J. Brickley, and J. Coles, 1993, “Stock-Based Incentive Compensation and Investment

Behavior,” Journal of Accounting and Economics 16, 349-372.



Bhushan, R., 1989, “Firm Characteristics and Analyst Following,” Journal of Accounting and

Economics 11, 183-206.



Black, F. and M. Scholes, 1973, “The Pricing of Options and Corporate Liabilities,” Journal of

Political Economy 81, 637-654.



Booth, L., V. Aivazian, A. Demirguc-Kunt, and V. Maksimovic, 2001, “Capital Structure in

Developing Countries,” Journal of Finance 56, 87-130.




                                                                                              31
Brounen, D., A. de Jong, and K. Koedijk, 2004, “Corporate Finance in Europe: Confronting

Theory with Practice,” Financial Management 33, 71-101.



Broussard, J., S. Buchenroth, and E. Pilotte, 2004, “CEO Incentives, Cash Flow, and

Investment,” Financial Management 33, 51-70.



Bryan, S., R. Nash, and A. Patel, 2006, “Can the Agency Costs of Debt and Equity Explain the

Changes in Executive Compensation During the 1990s?” Journal of Corporate Finance 12, 516-

535.



Bryan, S., L. Hwang, and S. Lilien, 2000, “CEO Stock-Based Compensation: An Empirical

Analysis of Incentive-Intensity, Relative Mix, and Economic Determinants,” Journal of Business

73, 661-693.



Carlin, W. and C. Mayer, 2003, “Finance, Investment, and Growth,” Journal of Financial

Economics 69, 191-226.



Cheffins, B., 1997, Company Law: Theory, Structure, and Operations, Oxford, UK, Clarendon

Press.



Claessens, S. and L. Klapper, 2005, “Bankruptcy Around the World: Explanation of Its Relative

Use,” American Law and Economics Review 7, 253-283.




                                                                                               32
Conyon, M. and K. Murphy, 2000, “The Prince and the Pauper? CEO Pay in the United States

and United Kingdom,” The Economic Journal 110, 640-671.



Conyon, M. and J. Schwalbach, 2000, “Executive Compensation: Evidence from the UK and

Germany,” Long Range Planning 33, 504-526.



David, R. and J. Brierley, 1985, Major Legal Systems in the World Today, London, UK, Stevens

and Sons.



Demirguc-Kunt, A. and V. Maksimovic, 1999, “Institutions, Financial Markets, and Firm Debt

Maturity,” Journal of Financial Economics 54, 295-336.



Demirguc-Kunt, A. and V. Maksimovic, 1998, “Law, Finance, and Firm Growth,” Journal of

Finance 53, 2,107-2,137.



Denis, D. and J. McConnell, 2003, “International Corporate Governance,” Journal of Financial

and Quantitative Analysis 38, 1-36.



Desai, M., F. Foley, and J. Hines, 2004, “A Multinational Perspective on Capital Structure

Choice and Internal Capital Markets,” Journal of Finance 59, 2,451-2,487.



Djankov, S., R. LaPorta, F. Lopez-de-Silanes, and A. Shleifer, 2008, “The Law and Economics




                                                                                             33
of Self-Dealing,” Journal of Financial Economics 88, 430-465.



Dittmar, A., J. Mahrt-Smith, and H. Servaes, 2003, “International Corporate Governance and

Corporate Cash Holdings,” Journal of Financial and Quantitative Analysis 38, 111-133.



Dow, J. and G. Gorton, 1997, “Stock Market Efficiency and Economic Efficiency: Is There a

Connection?” Journal of Finance 52, 1,087-1,129.



Durnev, A. and E. Han Kim, 2005, “To Steal Or Not To Steal: Firm Attributes, Legal

Environment, and Valuation,” Journal of Finance 60, 1,461-1,493.



Dyck, A. and L. Zingales, 2004, “Private Benefits of Control: An International Comparison,”

Journal of Finance 59, 537-600.



Edmans, A., X. Gabaix, and A. Landier, 2007, “A Calibratable Model of Optimal CEO

Incentives in Market Equilibrium,” NBER Working Paper.



Elston, J. and L. Goldberg, 2003, “Executive Compensation and Agency Costs in Germany,”

Journal of Banking and Finance 27, 1,391-1,410.



Frost, C. and G. Pownall, 1994, “Accounting Disclosure Practices in the United States and the

United Kingdom,” Journal of Accounting Research 32, 75-102.




                                                                                                34
Gabaix, X. and A. Landier, 2008, “Why Has CEO Pay Increased So Much?” Quarterly Journal

of Economics 123, 49-100.



Gaver, J. and K. Gaver, 1993, “Additional Evidence on the Association Between the Investment

Opportunity Set and Corporate Financing, Dividend, and Compensation Policies,” Journal of

Accounting and Economics 16, 125-160.



Giannetti, M., 2007, “Serial CEO Incentives and the Structure of Managerial Contracts,”

Stockholm School of Economics Working Paper.



Giannetti, M., 2003, “Do Better Institutions Mitigate Agency Problems? Evidence from

Corporate Finance Choices,” Journal of Financial and Quantitative Analysis 38, 185-212.



Gorton, G. and F. Schmid, 2000, “Universal Banking and the Performance of German Firms,”

Journal of Financial Economics 58, 29-80.



Holmstrom, B., 1979, “Moral Hazard and Observability,” Bell Journal of Economics 10, 74-91.



Holmstrom, B. and J. Tirole, 1993, “Market Liquidity and Performance Monitoring,” Journal of

Political Economy 101, 678-709.



Jensen, M., 1986, “Agency Costs of Free Cash Flow, Corporate Finance and Takeovers,”

American Economic Review 76, 323-339.




                                                                                            35
Jensen, M. and W. Meckling, 1976, “Theory of the Firm: Managerial Behavior, Agency Costs

and Ownership Structure,” Journal of Financial Economics 3, 305-360.



Jensen, M. and K. Murphy, 1990, “Performance Pay and Top-Manager Incentives,” Journal of

Political Economy 98, 225-264.



John, T. and K. John, 1993, “Top Management Compensation and Capital Structure,” Journal of

Finance 48, 949-974.



Johnson, S., J. McMillan, and C. Woodruff, 2002, “Property Rights and Finance,” American

Economic Review 92, 1,335-1,356.



Johnson, S., R. LaPorta, F. Lopez-de-Silanes, and A. Shleifer, 2000, “Tunneling,” American

Economic Review 90, 22-27.



Kato, H., M. Lemmon, M. Lou, and J. Schallheim, 2005, “An Empirical Examination of the

Costs and Benefits of Executive Stock Options: Evidence from Japan,” Journal of Financial

Economics 78, 435-461.



Kole, S., 1997, “The Complexity of Compensation Contracts,” Journal of Financial Economics

43, 79-104.




                                                                                             36
Kraft, K. and A. Niederprum, 1999, “Determinants of Management Compensation with Risk-

Averse Agents and Dispersed Ownership of the Firm,” Journal of Economic Behavior and

Organization 40, 17-27.



Lang, M., K. Lins, and D. Miller, 2003, “ADRs, Analysts, and Accuracy: Does Cross-Listing in

the United States Improve a Firm’s Information Environment and Increase Market Value?”

Journal of Accounting Research 41, 317-345.



La Porta, R., F. López-de-Silanes, and A. Shleifer, 2006, “What Works in Securities Laws,”

Journal of Finance 61, 1-32.



La Porta, R., F. López-de-Silanes, and A. Shleifer, 1999, “Corporate Ownership Around the

World,” Journal of Finance 54, 471-517.



La Porta, R., F. López-de-Silanes, A. Shleifer, and R. Vishny, 2002, “Government Ownership of

Banks,” Journal of Finance 57, 265-301.



La Porta, R., F. López-de-Silanes, A. Shleifer, and R.Vishny, 2000, “Agency Problems and

Dividend Policies Around the World,” Journal of Finance 55, 1-33.



La Porta, R., F. López-de-Silanes, A. Shleifer, and R.Vishny, 1999, “The Quality of

Government,” Journal of Law, Economics, and Organization 15, 222-279.




                                                                                             37
La Porta, R., F. López-de-Silanes, A. Shleifer, and R.Vishny, 1998, “Law and Finance,” Journal

of Political Economy 106, 1,113-1,150.



La Porta, R., F. López-de-Silanes, A. Shleifer, and R.Vishny, 1997, “Legal Determinants of

External Finance,” Journal of Finance 52, 1,131-1,150.



Lehn, K. and A. Poulsen, 1989, “Free Cash Flow and Stockholder Gains in Going Private

Transactions,” Journal of Finance 44, 771-787.



Licht, A., 2003, “Cross-Listing and Corporate Governance: Bonding or Avoiding?” Chicago

Journal of International Law 4, 141-163.



Matthews, J., 2006, “Global Stock Plans: Greater Sensitivity to Local Conditions,” Benefits and

Compensation International 35, 18-30.



Megginson, W., R. Nash, and M. Van Randenborgh, 1994, “The Financial and Operating

Performance of Newly-Privatized Firms: An International Empirical Analysis,” Journal of

Finance 49, 403-452.



Mehran, H., 1995, “Executive Compensation Structure, Ownership, and Firm Performance,”

Journal of Financial Economics 38, 163-184.



Modigliani, F. and E. Perotti, 2000, “Security Markets Versus Bank Finance: Legal Enforcement




                                                                                              38
and Investors’ Protection,” International Review of Finance 1, 81-96.



Morck, R., B. Yeung, and W. Yu, 2000, “The Information Content of Stock Markets: Why Do

Emerging Markets Have Synchronous Stock Price Movements?” Journal of Financial

Economics 58, 215-260.



Myers, S., 1977, “Determinants of Corporate Borrowing,” Journal of Financial Economics 5,

147-175.



Pagano, M. and P. Volpin, 2005, “The Political Economy of Corporate Governance,” American

Economic Review 95, 1,005-1,030.



Pagano, M., A. Roell, and J. Zechner, 2002, “The Geography of Equity Listing: Why Do

Countries List Abroad?” Journal of Finance 57, 2,651-2,694.



Pistor, K., M. Raiser, and S. Gelfer, 2000, “Law and Finance in Transition Economies,”

Economics of Transition 8, 325-368.



Rajan, R. and L. Zingales, 1998, “Financial Dependence and Growth,” American Economic

Review 88, 559-586.



Rajan, R. and L. Zingales, 1995, “What Do We Know About Capital Structure? Some Evidence

from International Data,” Journal of Finance 50, 1,421-1,460.




                                                                                            39
Reese, W., and M. Weisbach, 2002, “Protection of Minority Shareholder Interests, Cross-

Listings in the United States, and Subsequent Equity Offerings,” Journal of Financial Economics

66, 65-104.



Ryan, H. and R. Wiggins, 2002, “The Interactions Between R&D Investment Decisions and

Compensation Policy,” Financial Management 31, 5-29.



Shleifer, A. and R. Vishny, 1997, “A Survey of Corporate Governance,” Journal of Finance 52,

737-783.



Siegel, J., 2005, “Can Foreign Firms Bond Themselves Effectively by Renting U.S. Securities

Laws?” Journal of Financial Economics 75, 319-359.



Siegel, J., A. Licht, and S. Schwartz, 2008, “Egalitarianism, Cultural Distance, and FDI: A New

Approach,” Harvard University Working Paper.



Stulz, R. and R. Williamson, 2003, “Culture, Openness and Finance,” Journal of Financial

Economics 70, 313-349.



Tesar, L. and I. Werner, 1995, “Home Bias and High Turnover,” Journal of International Money

and Finance 15, 467-492.




                                                                                              40
Towers P., 2006, Managing Global Pay and Benefits.



Towers P., 2002, Worldwide Total Remuneration (2001-2002).



Vickers, J. and G. Yarrow, 1991, “Economic Perspectives on Privatization,” Journal of

Economic Perspectives 5, 111-132.



Wahlgren, E., 2001, “Spreading the Yankee Way of Pay,” Business Week Online, April 18.



Yermack, D., 1995, “Do Corporations Award CEO Stock Options Effectively?” Journal of

Financial Economics 39, 237-269.



Zhang, Y., 2008, “Are Debt and Incentive Compensation Substitutes in Controlling the Free

Cash Flow Agency Problem?,” forthcoming Financial Management.



Zingales, L., 1995, “Insider Ownership and The Decision to Go Public,” Review of Economic

Studies 62, 425-448.




                                                                                            41
                                   Table I. Observations for Each Country by Year

This table presents the number of observations (Level 2 and Level 3 ADR issues) by country for which we have executive
compensation data from 1996–2000. We identified the ADR issuers from Compustat and obtained the specific executive
compensation data from the Securities and Exchange Commission’s EDGAR database. We express the observations in firm
years.

                                                                     Year
                      Country               1996     1997       1998    1999          2000        Totals
          Argentina                           4        7          8        8            9            36
          Australia                           2        7          9       10            7            35
          Austria                                                                       1             1
          Belgium                                                                       1             1
          Bermuda                            1        2          4           4          2            13
          Brazil                             1        3          10         16         16           46
          British Virgin Islands                                             1                        1
          Chile                              8        10         11         11         12           52
          China                              5         7         7           8         10            37
          Denmark                            1         1                     1         1              4
          Dominican Republic                           1         1           1         1              4
          England                            14       23         33         40         53           163
          Finland                            1         1         1           3         6             12
          France                             9        14         18         17         19           77
          Germany                            5         5         5           6         11            32
          Greece                                                 1           2         2              5
          Hong Kong                          1        1          2           4         6             14
          Hungary                                     1          1           1         1              4
          India                                                  1           2         5              8
          Indonesia                                    1         2           2         1              6
          Ireland                            1         5         8           9         12            35
          Israel                             5         7         7           7         10            36
          Italy                              4         5         8          10         8             35
          Japan                              4         9         10         13         14           50
          Luxembourg                         1         2         2           2         3             10
          Mexico                             11       13         16         21         17           78
          Netherlands                        9        11         16         16         17           69
          New Zealand                                  1         2           2         2              7
          Norway                             2         2         1           1         2              8
          Pakistan                           1         1         1           1         1              5
          Peru                                         1         1           1         1              4
          Philippines                                  1         1           1         2              5
          Portugal                                               1           1         2              4
          Russia                             1        1          2           3         2              9
          Singapore                                                          1         2              3
          South Africa                                2          3           3         6             14
          South Korea                        3        3          3           5         3             17
          Spain                              1        2          3           4         5             15
          Sweden                             5        8          8           9         9             39
          Switzerland                        1        1          2           1         7             12
          Taiwan                                      1          1           1         4              7
          Venezuela                          2        2          2           2                        8
                                            103      162        212         251        294         1,022



                                                                                                                  42
    Table II. Mean Values of Compensation Structure, Firm Size, and Firm Leverage by Country

This table presents the mean values of various measures of compensation structure for firms in each country of our sample.
PCTEQ is the ratio of total equity-based compensation (stock option compensation plus restricted stock compensation) to total
compensation. PCTOPT and PCTREST are, respectively, ratios of stock option compensation and restricted stock compensation
to total compensation. Total compensation is the sum of cash compensation, stock option compensation, restricted stock, other
compensation, and long-term incentive compensation. Stock option compensation is measured using the Black Scholes (1973)
Option pricing model. Size is the natural logarithm of the firm’s market capitalization. Leverage is the ratio of the firm’s total
debt to total assets.

              Country         No. of Obs.       PCTEQ          PCTOPT          PCTREST           Size         Leverage
     Argentina                      36           0.000           0.000            0.000          7.202           0.282
     Australia                      35           0.312           0.279            0.033          7.455           0.372
     Austria                        1            0.000           0.000            0.000          7.860           0.566
     Belgium                        1            0.000           0.000            0.000          8.453           0.681
     Bermuda                        13           0.074           0.074            0.000          4.424           0.376
     Brazil                         46           0.023           0.012            0.011          7.182           0.399
     British Virgin Islands         1            0.337           0.075            0.262          8.195           0.000
     Chile                          52           0.000           0.000            0.000          6.712           0.386
     China                          37           0.014           0.000            0.014          6.819           0.369
     Denmark                        4            0.109           0.109            0.000          9.026           0.211
     Dominican Republic             4            0.000           0.000            0.000          5.635           0.651
     England                       163           0.197           0.108            0.089          8.032           0.360
     Finland                        12           0.029           0.029            0.000          8.232           0.247
     France                         77           0.138           0.110            0.028          8.104           0.268
     Germany                        32           0.046           0.046            0.000          6.769           0.248
     Greece                         5            0.000           0.000            0.000          8.072           0.416
     Hong Kong                      14           0.014           0.014            0.000          6.048           0.334
     Hungary                        4            0.000           0.000            0.000          8.648           0.532
     India                          8            0.142           0.142            0.000          7.419           0.081
     Indonesia                      6            0.000           0.000            0.000          7.949           0.303
     Ireland                        35           0.109           0.032            0.077          6.278           0.278
     Israel                         36           0.160           0.102            0.058          6.173           0.308
     Italy                          35           0.052           0.052            0.000          7.844           0.315
     Japan                          50           0.024           0.014            0.010          9.700           0.322
     Luxembourg                     10           0.000           0.000            0.000          6.621           0.459
     Mexico                         78           0.001           0.001            0.000          6.661           0.397
     Netherlands                    69           0.250           0.250            0.000          9.321           0.331
     New Zealand                    7            0.423           0.423            0.000          8.738           0.458
     Norway                         8            0.000           0.000            0.000          5.793           0.218
     Pakistan                       5            0.000           0.000            0.000          6.786           0.219




                                                                                                                              43
Table II. Mean Values of Compensation Structure, Firm Size and Firm Leverage by Country
                                      (Continued)

Country            No. of Obs.   PCTEQ    PCTOPT     PCTREST      Size      Leverage
Peru                    4        0.000     0.000       0.000      7.913       0.422
Philippines             5        0.000     0.000       0.000      7.253       0.523
Portugal                4        0.002     0.000       0.002      9.327       0.509
Russia                  9        0.000     0.000       0.000      6.640       0.390
Singapore               3        0.132     0.023       0.109      7.904       0.526
South Africa           14        0.307     0.259       0.048      4.675       0.416
South Korea            17        0.000     0.000       0.000      9.370       0.460
Spain                  15        0.023     0.023       0.000      10.070      0.496
Sweden                 39        0.091     0.091       0.000      7.572       0.312
Switzerland            12        0.040     0.029       0.011      8.084       0.321
Taiwan                  7        0.000     0.000       0.000      9.396       0.260
Turkey                  1        0.000     0.000       0.000      8.812       0.539
Venezuela               8        0.000     0.000       0.000      8.248       0.158
Total/Average         1,022      0.094     0.069       0.024      7.660       0.346




                                                                                          44
  Table III. Mean Values of Compensation Structure and Firm Characteristics by Year, 1996-2000

This table presents the mean values of our various measures of compensation structure and the main control variables (arrayed by
year from 1996-2000). PCTEQ is the percentage of total compensation comprised of stock option and restricted stock
compensation. PCTOPT and PCTREST are, respectively, ratios of stock option compensation and restricted stock compensation
to total compensation. Total compensation is the sum of cash compensation, stock option compensation, restricted stock, other
compensation, and long-term incentive compensation. Market-to-Book is the book value of total assets less the book value of
equity plus the market value of equity divided by the book value of total assets. Size is the natural logarithm of the firm’s market
capitalization. Leverage is the ratio of the firm’s total debt to total assets. We express number of observations in firm years.


                                                                                       Market-
           Date          Obs.         PCTEQ          PCTOPT          PCTREST                             Size       Leverage
                                                                                       to-Book
    1996                  103          0.048           0.043            0.005            1.894           7.487         0.313
    1997                  162          0.094           0.064            0.030            1.959           7.523         0.342
    1998                  212          0.069           0.044            0.025            1.963           7.512         0.368
    1999                  251          0.099           0.075            0.023            2.364           7.747         0.346
    2000                  294          0.127           0.097            0.030            2.031           7.823         0.344




                                                                                                                                45
Table IV. Average Values of Compensation Structure and Control Variables by Antidirector Rights
                      Index (ARI), Origin of Legal System, and Rule of Law

This table presents the mean values of Percent Equity (PCTEQ) and the main control variables (arrayed according to the nation’s
Antidirector Rights Index (ARI), its legal origin, and its value for Rule of Law). ARI, Legal Origin, and Rule of Law are from
LaPorta et al. (1997). For the ARI, higher values indicate greater protection. Rule of Law measures how effectively a nation
enforces its laws. Higher values indicate a stronger tradition of law and order. In Panel C, Quartile 1 contains data from countries
with the lowest values of this variable. PCTEQ is percentage of total compensation comprised of stock options and restricted
stock. Free Cash Flow is the ratio of operating income before depreciation less the sum of income tax, interest, and dividends
paid to the firm’s market value of equity. Market-to-Book is the book value of total assets less the book value of equity plus the
market value of equity divided by the book value of total assets. Size is the natural logarithm of the firm’s market capitalization.
Leverage is the ratio of the book values of the firm’s total debt to total assets.


                                    Panel A: Average Values by Antidirector Rights Index (ARI)

                                                              Free Cash        Market-to-
            ARI            No. Obs            PCTEQ                                                   Size           Leverage
                                                                Flow             Book
    0                         114              0.017             0.052             1.812              7.051             0.375
    1                          57              0.037             0.043             2.183              7.351             0.280
    2                         304              0.108             0.045             2.068              8.279             0.339
    3                         241              0.063             0.048             1.918              7.489             0.344
    4                         302              0.159             0.011             2.304              7.445             0.353

                                        Panel B: Average Values by Origin of Legal System

                                                              Free Cash        Market-to-
        Legal Origin       No. Obs            PCTEQ                                                   Size           Leverage
                                                                Flow             Book

    English                   334              0.180             0.009             2.407              7.232             0.344

    German                    119              0.027             0.052             2.020              8.778             0.325

    French                    456              0.066             0.052             1.909              7.698             0.359

    Scandinavian               63              0.071             0.021             2.303              7.689             0.283

    Socialist                  46              0.010             0.076             1.160              6.784             0.373

                                             Panel C: Average Values by Rule of Law
                                                              Free Cash        Market-to-
        Rule of Law        No. Obs            PCTEQ                                                   Size           Leverage
                                                                Flow             Book

    Quartile 1                218              0.047             0.046             1.924              6.940             0.353

    Quartile 2                220              0.032             0.057             1.871              7.351             0.359

    Quartile 3                280              0.148             0.007             2.366              8.349             0.339

    Quartile 4                206              0.159             0.030             2.294              8.121             0.316




                                                                                                                                46
          Table V. Regression Results Explaining Compensation Structure of Non-U.S. Firms

This table presents regression results from Tobit models. Our dependent variables are as follows: PCTEQ is the ratio of total
equity-based compensation (stock option compensation plus restricted stock compensation) to total compensation. PCTOPT and
PCTREST are, respectively, ratios of stock option compensation and restricted stock compensation to total compensation. Total
compensation is the sum of cash compensation, stock option compensation, restricted stock, other compensation, and long-term
incentive compensation. EQMIX, OPTMIX, and RESTMIX are, respectively, total equity-based compensation, stock option
compensation, and restricted stock compensation each scaled by cash compensation. Cash compensation is salary plus bonus.

Independent variables are as follows. In Panels A-D, respectively, we include alternative measures of legal protection through
shareholder rights: Antidirector Rights Index from LaPorta et al. (1997), where higher values indicate greater protection; Revised
AntiDirIndex and AntiSelfDeal are alternative measures of shareholder rights protection (from Djankov et al., 2008), where
higher values of both suggest stronger protection of shareholder rights; and Legal Origin, which is an indicator variable set to one
if the legal system is based on English common law. Our other independent variables (included in all models) are Rule of Law
(measures quality of law enforcement; higher values indicate stricter enforcement), Free Cash Flow (ratio of operating income
before depreciation less the sum of income tax, interest, and dividends paid to the firm’s market value), Market-to-Book (book
value of total assets less the book value of equity plus the market value of equity divided by the book value of total assets), Size
(log of market capitalization), Leverage (ratio of book values of total debt to total assets), and Log GDP per capita for each firm’s
respective country. Noncen is the number of non-censored (at zero) observations for the dependent variables. We include
industry and year controls, but do not report coefficient values. Standard errors robust to serial correlation and heteroscedasticity
are in parentheses.


                                                              Panel A
                         Model 1      Model 2      Model 3     Model 4      Model 5      Model 6
                         PCTEQ        PCTOPT      PCTREST      EQMIX        OPTMIX      RESTMIX
                         N = 762      N = 762      N = 762     N = 762      N = 762      N = 762
                       Noncen = 188 Noncen = 153 Noncen = 58 Noncen = 188 Noncen = 153 Noncen = 58
   Antidirector             0.131***         0.101***          0.275***          0.489***         0.394***          0.759***
   Rights Index             (0.025)          (0.026)           (0.061)           (0.088)          (0.097)           (0.169)
                            0.093***         0.128***          -0.021            0.296***         0.357***          -0.052
   Rule of Law
                            (0.025)          (0.029)           (0.042)           (0.087)          (0.097)           (0.117)
                            -0.551*           -0.121           -0.813*           -1.730            -0.554           -2.010
   Free Cash Flow
                            (0.311)           (0.355)          (0.466)           (1.091)           (1.211)          (1.278)
                            0.058**           0.033*           0.077**           0.189**            0.098           0.198**
   Market-to-Book
                            (0.018)           (0.019)          (0.029)           (0.063)           (0.067)          (0.080)
                            -0.013             0.002           -0.004            -0.037            -0.007           -0.006
   Size
                            (0.014)           (0.015)          (0.024)           (0.048)           (0.051)          (0.066)
                            -0.221*           -0.181           -0.340            -0.808*           -0.653           -0.958*
   Leverage
                            (0.126)           (0.135)          (0.227)           (0.442)           (0.466)          (0.627)
   Log GDP per              -0.012            -0.095           0.219**           -0.057            -0.184           0.595**
   Capita                   (0.053)           (0.058)          (0.107)           (0.186)           (0.202)          (0.296)
                                                             Panel B
                         Model 1      Model 2      Model 3     Model 4      Model 5      Model 6
                         PCTEQ        PCTOPT      PCTREST      EQMIX        OPTMIX      RESTMIX
                         N = 762;     N = 762;     N = 762;    N = 762;     N = 762;     N=762;
                       Noncen = 188 Noncen = 153 Noncen = 58 Noncen = 188 Noncen = 153 Noncen = 58
   Revised                  0.180***         0.142***          0.302***          0.671***         0.557***          0.843***
   AntiDirIndex             (0.032)          (0.035)           (0.063)           (0.114)          (0.122)           (0.171)
                            0.095***         0.132***          -0.033            0.303***         0.372***           0.085
   Rule of Law
                            (0.025)          (0.029)           (0.042)           (0.089)          (0.099)           (0.115)
                            -0.451            -0.059           -0.764*           -1.444            -0.307           -1.836
   Free Cashflow
                            (0.312)           (0.359)          (0.453)           (1.097)           (1.221)          (1.231)




                                                                                                                                 47
Table V. Regression Results Explaining Compensation Structure of Non-U.S. Firms (Continued)

                    Model 1      Model 2      Model 3     Model 4      Model 5      Model 6
                    PCTEQ        PCTOPT      PCTREST      EQMIX        OPTMIX      RESTMIX
                    N = 762;     N = 762;     N = 762;    N = 762;     N = 762;     N=762;
                  Noncen = 188 Noncen = 153 Noncen = 58 Noncen = 188 Noncen = 153 Noncen = 58
                     0.057**      0.032*        0.077**      0.186**       0.095       0.197**
 Market-to-Book
                     (0.018)      (0.019)       (0.028)      (0.063)      (0.068)      (0.077)
                     -0.014        0.001        -0.004       -0.042       -0.012       -0.005
 Size
                     (0.013)      (0.014)       (0.024)      (0.048)      (0.513)      (0.064)
                     -0.249**     -0.199        -0.402*      -0.911**     -0.721       -1.134*
 Leverage
                     (0.125)      (0.136)       (0.222)      (0.441)      (0.466)      (0.610)
 Log GDP per          0.026       -0.067        0.277**       0.085       -0.073       0.747**
 Capita              (0.052)      (0.059)       (0.101)      (0.186)      (0.203)      (0.277)
                                               Panel C
                    Model 1      Model 2      Model 3     Model 4      Model 5      Model 6
                    PCTEQ        PCTOPT      PCTREST       EQMIX       OPTMIX      RESTMIX
                    N = 762;     N = 762;     N = 762;    N = 762;     N = 762;     N = 762;
                  Noncen = 188 Noncen = 153 Noncen = 58 Noncen = 188 Noncen = 153 Noncen = 58
                     0.648***     0.508***      1.203***     2.463***     1.988***     3.359***
 AntiSelfDeal
                     (0.107)      (0.115)       (0.231)      (0.375)      (0.396)      (0.630)
                     0.112***     0.146***      -0.006       0.369***     0.426***     -0.012
 Rule of Law
                     (0.026)      (0.029)       (0.042)      (0.090)      (0.101)      (0.115)
                     -0.417       -0.024        -0.743*      -1.260        0.170       -1.768
 Free Cash Flow
                     (0.310)      (0.360)       (0.451)      (1.083)      (1.220)      (1.224)
                     0.051**       0.028        0.067**      0.162**       0.078       0.168**
 Market-to-Book
                     (0.017)      (0.019)       (0.028)      (0.062)      (0.067)      (0.078)
                     -0.007        0.005         0.049       -0.017        0.008        0.021
 Size
                     (0.013)      (0.014)       (0.024)      (0.047)      (0.050)      (0.065)
                     -0.251**     -0.203        -0.396*      -0.923**     -0.742*      -1.118*
 Leverage
                     (0.124)      (0.135)       (0.220)      (0.434)      (0.462)      (0.603)
 Log GDP per         -0.069       -0.143**       0.120       -0.277       -0.369**      0.310
 Capita              0.054         0.061         0.100       0.190         0.209        0.273
                                               Panel D
                    Model 1      Model 2      Model 3     Model 4      Model 5      Model 6
                    PCTEQ        PCTOPT      PCTREST      EQMIX        OPTMIX      RESTMIX
                    N = 762;      N=762;      N=762;      N=762;       N=762;       N=762;
                  Noncen = 188 Noncen = 153 Noncen = 58 Noncen = 188 Noncen = 153 Noncen = 58
                       0.493***     0.430***      0.680***     1.802***     1.554***     1.893***
 Legal Origin
                     (0.060)      (0.065)       (0.127)      (0.211)      (0.227)      (0.346)
                       0.133***     0.163***       0.029       0.442***     0.482***      0.086
 Rule of Law
                     (0.025)      (0.029)       (0.042)      (0.089)      (0.100)      (0.116)
                     -0.257        0.166        -0.780*      -0.756        0.386       -1.872
 Free Cash Flow
                     (0.296)      (0.350)       (0.451)      (1.041)      (1.194)      (1.220)
                     0.042**       0.019        0.058**      0.130**       0.048       0.143*
 Market-to-Book
                     (0.017)      (0.019)       (0.028)      (0.061)      (0.067)      (0.078)




                                                                                                    48
Table V. Regression Results Explaining Compensation Structure of Non-U.S. Firms (Continued)

                          Model 1      Model 2      Model 3     Model 4      Model 5      Model 6
                          PCTEQ        PCTOPT      PCTREST      EQMIX        OPTMIX      RESTMIX
                          N = 762;      N=762;      N=762;      N=762;       N=762;       N=762;
                        Noncen = 188 Noncen = 153 Noncen = 58 Noncen = 188 Noncen = 153 Noncen = 58
                             -0.004              0.008      0.008    -0.006      0.017     0.030
     Size
                             (0.013)            (0.014)    (0.023)   (0.046)    (0.049)    (0.065
                            -0.285**            -0.239*    -0.388*   -1.033**   -0.853*    -1.091*
     Leverage
                            (0.119)             (0.130)    (0.220)   (0.418)    (0.449)    (0.602)
     Log GDP per             -0.094*            -0.164**   0.111     -0.352*    -0.424**   0.290
     Capita                   0.051              0.058     0.096      0.181      0.200     0.261
 ***
       significant at the 0.01 percent level
 **
       significant at the 0.05 percent level
 *
        significant at the 0.10 percent level




                                                                                                      49
 Table VI. Regression Results Explaining Compensation Structure of Non-U.S. Firms: Robustness
                                Tests Using Alternative Control Variables

This table presents the regression results from Tobit models. Our dependent variable in all models is PCTEQ (ratio of total
equity-based compensation to total compensation). Independent variables included in all models are: Antidirector Rights Index
(measures strength of legal protection for shareholders; higher values indicate stronger protection), Rule of Law (measures
quality of law enforcement; higher values indicate stricter enforcement), Free Cash Flow (ratio of operating income before
depreciation less the sum of income tax, interest, and dividends paid to the firm’s market value), Market-to-Book (book value of
total assets less the book value of equity plus the market value of equity divided by the book value of total assets), Size (log of
market capitalization), Leverage (ratio of book values of total debt to total assets), and Log GDP per capita for each firm’s
respective country. Model 1 includes an alternative measure of institutional environment. Polity is from CIDCM and is a measure
of the political environment within each country (higher values indicate a more democratic political system). Models 2 and 3
include additional measures of the nation’s managerial labor market. Listed Firms is the number of firms traded on the national
stock exchange. Family is a measure of the prevalence of family controlled businesses in the economy (higher values indicate a
larger proportion of family controlled firms). This variable is from LaPorta et al. (1998). Models 4 and 5 include alternative
measures of firm size. Sales is the natural logarithm of a firm’s sales. Assets is the natural logarithm of a firm’s total assets.
Noncen is the number of non-censored (at zero) observations for the dependent variables. We include industry and year controls,
but do not report coefficient values. Standard errors robust to serial correlation and heteroscedasticity are in parentheses.

                            Model 1      Model 2      Model 3      Model 4      Model 5
                            N = 698      N = 756      N = 636      N = 755      N = 756
                          Noncen = 156 Noncen = 188 Noncen = 174 Noncen = 188 Noncen = 188
     Antidirector             0.150***              0.113***   0.104***         0.129***          0.130***
     Rights Index             (0.028)               (0.026)    (0.027)          (0.024)           (0.024)
                                0.039               0.104***   0.101***         0.093***          0.092***
     Rule of Law
                               (0.027)              (0.025)    (0.028)          (0.025)           (0.025)
                               -0.565               -0.580*    -0.062           -0.480            -0.551*
     Free Cash Flow
                               (0.352)              (0.311)    (0.318)          (0.321)           (0.312)
                               0.068**              0.057**    0.052**          0.051**           0.052**
     Market-to-Book
                               (0.019)              (0.017)    (0.018)          (0.017)           (0.017)
                               -0.016               -0.009     -0.021
     Size
                               (0.016)              (0.013)    (0.013)
                               -0.245*              -0.228*    -0.272*          -0.200            -0.214*
     Leverage
                               (0.138)              (0.125)    (0.129)          (0.129)           (0.129)
     Log GDP per                0.009               -0.054     -0.244**         -0.010            -0.012
     Capita                    (0.058)              (0.057)    (0.124)          (0.053)           (0.053)
                               0.064**
     Polity
                               (0.029)
                                                     0.003
     Listed Firms
                                                    (0.001)*
                                                                -0.512
     Family
                                                               (0.176)**
                                                                                -0.014
     Sales
                                                                                (0.013)
                                                                                                  -0.010
     Assets
                                                                                                  (0.014)

     ***
           significant at the 0.01 percent level
     **
           significant at the 0.05 percent level
     *
            significant at the 0.10 percent level




                                                                                                                               50
Table VII. Regression Results Explaining Compensation Structure of Non-U.S. Firms: Robustness
                                 Tests Using Various Samples

This table presents the regression results from Tobit models based on subsamples of our data. Reported subsamples are: 1) all
firms excluding those from the UK, 2) all firms excluding those from G7 countries, 3) all firms and all U.S. firms during the
sample period, and 4) all firms with each matched with a U.S. firm (paired by industry and size).

The dependent variable in all models is PCTEQ (total equity-based compensation to total compensation). Independent variables
are Antidirector Rights Index, Rule of Law (higher values indicate stricter enforcement) Free Cash Flow (ratio of operating
income before depreciation less the sum of income tax, interest, and dividends paid to the firm’s market value), Market-to-Book
(book value of total assets less the book value of equity plus the market value of equity divided by the book value of total assets),
Size (log of market capitalization), Leverage (ratio of book values of total debt to total assets), and Log of GDP per capita for
each firm’s country. Noncen is the number of non-censored (at zero) observations for the dependent variables. We include
industry and year controls, but do not report coefficient values. Standard errors robust to serial correlation and heteroscedasticity
are in parentheses.

                                          Model 1
                                                               Model 2         Model 3               Model 4
                                          Non-UK
                                                               Non-G7            Add                 Matched
                                           Firms
                                                                Firms         U.S. firms            U.S. firms
                                         N = 625;
                                                               N = 471;       N = 3,805              N = 1,472
                                         Noncen =
                                                             Noncen = 97    Noncen = 2,906         Noncen = 963
                                             120
                   Antidirector            0.098**             0.151***           0.175***             0.172***
                   Rights Index            (0.031)             (0.037)            (0.021)              (0.019)
                                           0.100**             0.073**            0.120**               0.114**
                   Rule of Law
                                           (0.027)             (0.026)            (0.038)               (0.034)
                                          -1.041**             -1.556**           -0.215**             -0.347**
                   Free Cash Flow
                                          (0.547)              (0.585)            (0.070)              (0.109)
                   Market-to-              0.057**              0.016              0.006                 0.002
                   Book                    (0.025)             (0.028)            (0.005)               (0.008)
                                           -0.015               0.021             0.050***             0.051***
                   Size
                                           (0.017)             (0.020)            (0.004)              (0.007)
                                           -0.168              -0.316             -0.093**             -0.086I*
                   Leverage
                                           (0.203)             (0.208)            (0.031)              (0.052)
                   Log GDP per             -0.045              -0.053             -0.119                -0.115*
                   Capita                  (0.059)             (0.059)            (0.074)               (0.067)

               ***
                     significant at the 0.01 percent level
               **
                     significant at the 0.05 percent level
               *
                     significant at the 0.10 percent level




                                                                                                                                 51