Three Essays in Corporate Governance

Wilfrid Laurier University
Scholars Commons @ Laurier

Theses and Dissertations (Comprehensive)
2010
Three Essays in Corporate Governance

Vishaal Rabindranauth Anand Baulkaran
Follow this and additional works at: http://scholars.wlu.ca/etd

Part of the Management Sciences and Quantitative Methods Commons
Recommended Citation
Baulkaran, Vishaal Rabindranauth Anand, "Three Essays in Corporate Governance" (2010). Theses and Dissertations (Comprehensive).
Paper 1121.
This Dissertation is brought to you for free and open access by Scholars Commons @ Laurier. It has been accepted for inclusion in Theses and
Dissertations (Comprehensive) by an authorized administrator of Scholars Commons @ Laurier. For more information, please contact
scholarscommons@wlu.ca.
1*1
Library and Archives

Canada
Bibliotheque et
Archives Canada
Published Heritage
Branch
Direction du
Patrimoine de I'edition
395 Wellington Street

Ottawa ON K1A 0N4
Canada
395, rue Wellington

Ottawa ON K1A 0N4
Canada
Your file Votre reference
ISBN: 978-0-494-75409-2
Our file Notre reference
ISBN: 978-0-494-75409-2
NOTICE:
AVIS:
The author has granted a nonexclusive license allowing Library and

Archives Canada to reproduce,
publish, archive, preserve, conserve,
communicate to the public by
telecommunication or on the Internet,
loan, distribute and sell theses
worldwide, for commercial or noncommercial purposes, in microform,
paper, electronic and/or any other
formats.
L'auteur a accorde une licence non exclusive

permettant a la Bibliotheque et Archives
Canada de reproduire, publier, archiver,
sauvegarder, conserver, transmettre au public
par telecommunication ou par I'lnternet, preter,
distribuer et vendre des theses partout dans le
monde, a des fins commerciales ou autres, sur
support microforme, papier, electronique et/ou
autres formats.
The author retains copyright

ownership and moral rights in this
thesis. Neither the thesis nor
substantial extracts from it may be
printed or otherwise reproduced
without the author's permission.
L'auteur conserve la propriete du droit d'auteur

et des droits moraux qui protege cette these. Ni
la these ni des extra its substantiels de celle-ci
ne doivent etre imprimes ou autrement
reproduits sans son autorisation.
In compliance with the Canadian

Privacy Act some supporting forms
may have been removed from this
thesis.
Conformement a la loi canadienne sur la

protection de la vie privee, quelques
formulaires secondaires ont ete enleves de
cette these.
While these forms may be included

in the document page count, their
removal does not represent any loss
of content from the thesis.
Bien que ces formulaires aient inclus dans

la pagination, il n'y aura aucun contenu
manquant.
1+1
Canada
THREE ESSAYS IN CORPORATE GOVERNANCE
by
Vishaal Rabindranauth Anand Baulkaran

BBA, University of New Brunswick, 2004
MBA, University of New Brunswick, 2006
DISSERTATION
Submitted to the School of Business and Economics
in partial fulfillment of the requirements for
Doctor of Philosophy in Management
2010
Vishaal Baulkaran 2010
ABSTRACT
This dissertation comprises three related but different essays on corporate
governance issues. The essays are preceded by an overview of the major areas of corporate
governance research. The first essay investigates whether the valuation discount of dual
class firms reported in the literature can be explained by three channels through which
private benefits can be extracted -excess compensation, excess cash and excess capital
expenditure. With a propensity score matched sample of S&P 1500 dual class and single
class companies, I provide evidence that excess compensation and excess cash holdings of
dual class companies lead to a larger discount that investors apply to the value of dual versus
single class companies. However, capital expenditure is not statistically significant in
explaining the dual class discount.
The second essay examines the impact of concentrated control under dual and single
class share structure on dividend policy. Three potential dividend policy hypotheses extraction of private benefits, family legacy and managerial reputation - are proposed and
tested. The results indicate that in the U.S. dual class firms pay less dividends and cash
dividend and total distribution decrease as the divergence of voting and cash flow rights
widens. This is consistent with the extraction of private benefits and the family legacy
hypotheses. However, using excess CEO compensation to disentangle these two hypotheses,
the payment of lower dividends in dual class firms is consistent with the extraction of
private benefits hypothesis.
The third essay investigates the impact of managerial and board entrenchment on
dual class discount. Dual class ownership structure is arguably the most effective antitakeover mechanism as it allows controlling shareholders to maintain concentrated voting
ii
positions even if additional equity needs to be issued. Thus, management is insulated from
hostile takeover and is able to become entrenched. Investors, knowing that dual class
structure can result in entrenchment, will discount the value of dual class firms. I provide
evidence that the larger the degree of entrenchment the larger the dual class discount. The
results also show that anti-takeover defenses in dual class firms, such as classified boards,
serve to entrench managers.
Keywords:
Dual Class Discount, Executive Compensation, Cash Holdings, Family
Ownership and Control, Ownership Structure, Dividend Policy, Agency Theory, Private
Benefits of Control, and Managerial Entrenchment
ACKNOWLEDGMENTS
I would like to express my sincere gratitude to my dissertation committee chair,

Professor Ben Amoako-Adu. Without his insightful feedback, guidance and encouragement
this dissertation would not have been possible. In addition, I am indebted to my dissertation
committee members, Professor Brian F. Smith and Professor Phelim Boyle for their
guidance, insightful comments and support over the past two research years. Also, I am
grateful to the Laurier Finance faculty and my fellow PhD students for their feedback and
support.
I am grateful to Professor Randall Wigle for acting as the internal/external at my
dissertation defence. I am thankful to Professor Stephen Foerster, who generously accepted
the role of external examiner, for his insightful recommendations and sincere
encouragement.
I would like to thank my parents Dwarka and Devikarai Baulkaran for their support
throughout this process. Finally, I would like to acknowledge the support and help of my
fiance Jessica Johnson throughout my graduate studies.
IV
VITA
September 16, 1979
Born - New Amsterdam, Berbice, Guyana
2004
Bachelor of Business Administration, University of New Brunswick, Canada
2006
Master of Business Administration, University of New Brunswick, Canada
FIELDS OF STUDY
Major Field: Business Administration

Concentration: Finance
TABLE OF CONTENTS
ABSTRACT
II
ACKNOWLEDGMENTS
IV
VITA
TABLE OF CONTENTS
VI
LIST OF TABLES
IX
LIST OF FIGURES
XI
CHAPTER 1
AN OVERVIEW OF CORPORATE GOVERNANCE
1.0.0
Conceptual Framework of Corporate Governance
2.0 CORPORATE GOVERNANCE MECHANISMS

2.1.0
2.2.0
2.3.0
2.4.0
2.5.0
Laws and Regulations

Hostile Takeovers
Blockholders and Institutional Investors
Board of Directors
Executive Compensation
1
1
5
5
5
6
8
10
3.0
BRIEF HISTORY OF THE DUAL CLASS MARKET IN THE UNITED STATES
11
4.0
FURTHER AREAS OF RESEARCH AND UNRESOLVED QUESTIONS
15
5.0
REFERENCES
20
CHAPTER 2
DUAL CLASS DISCOUNT, AGENCY PROBLEMS AND WEALTH EXPROPRIATION 25
1.0
INTRODUCTION
25
2.0
DUAL CLASS DISCOUNT
31
3.0
LITERATURE REVIEW AND HYPOTHESIS DEVELOPMENT

3.1.0
Ownership Structure and Compensation
3.2.0
Family Executives and Compensation
3.3.0
Dual Class Discount and Corporate Cash Holdings
3.4.0
Dual Class Discount and Capital Expenditure
3.5.0
Dual Class Discount and Excess Compensation
33
33
35
37
39
39
4.0
METHODOLOGY AND DATA

4.1.0
Methodology
4.1.1
4.1.2
4.1.3
4.1.4
4.1.5
40
40
40
43
45
47
47
VI
4.2
DATA
4.2.1
4.2.2
49
49
50
Sample Construction
Propensity Score Matching
5.0
RESULTS
5.1.0
Descriptive Statistics
5.2.0
Regression Analysis
5.2.1
5.2.2
5.2.3
5.2.4
5.2.5
51
51
55
55
57
58
60
60
5.3
ENDOGENEITY
5.3.1
Simultaneity and Two-Stage Least Square Estimation
5.3.2
Heckman Sample Selection
5.3.3
Robustness Check
62
62
64
64
6.0
CHAPTER SUMMARY AND CONCLUSIONS
65
7.0
REFERENCES
68
8.0
APPENDIX A:
100
CHAPTER 3
OWNERSHD? STRUCTURE, AGENCY PROBLEMS AND DIVD3END POLICY
102
1.0
INTRODUCTION
102
2.0
LITERATURE REVIEW
106
2.1
AGENCY THEORY, PRIVATE BENEFITS AND DIVDDEND POLICY

2.1.1
Private Benefits and Family Legacy Explanation for Dividend Policy
107
107
2.1.2
110
Managerial Reputation Explanation for Dividend Policy
2.2
WHY STUDY DUAL CLASS SHARE STRUCTURES OF FffiMS IN THE U.S.?
112
3.0
HYPOTHESES, METHODOLOGY AND DATA

3.1.0
Hypotheses
3.2.0
Methodology
3.3.0
Data
3.3.1
Sample Construction
3.3.2
RESULTS
4.1.0
4.2.0
Regression Analysis
4.3.0
Robustness
4.4.0
Simultaneity and Two-Stage Least Square Estimation
113
113
116
120
120
122
123
123
125
129
130
5.0
131
6.0
REFERENCES
134
4.0
vn
CHAPTER 4
THE VALUATION EFFECTS OF DUAL CLASS STRUCTURE AND MANAGERIAL
ENTRENCHMENT
15 5
1.0
INTRODUCTION
155
2.0
LITERATURE REVIEW AND HYPOTHESIS DEVELOPMENT

2.1.0
Literature Review
2.2.0
Hypothesis
METHODOLOGY AND DATA
3.1
Methodology
3.2
Data
3.2.1
Sample Construction
3.2.2
15 8
15 8
162
162
162
165
165
166
4.0
RESULTS
4.1.0
4.2.0
Regression Analysis
4.3.0
Entrenchment Conditional on Past Performance
4.4.0
Robustness
168
168
170
172
17 5
5.0
177
6.0
REFERENCES
179
3.0
CHAPTER 5
CONTRD3UTIONS AND CONCLUSIONS
196
viii
LIST OF TABLES
CHAPTER 2
Table 1: Summary statistics of executive compensation (2001-2007)
74
Table 2: Number of dual class firms by two-digit SIC Industry code
77
Table 3: Summary statistics of governance characteristics, firm characteristics

and executive characteristics
78
Table 4: Regression of executive compensation on management voting leverage,

firm and governance characteristics
81
Table 5: Panel regression of executive compensation on family executives, management voting

leverage, firm and governance characteristics
83
Table 6: Panel regression for a sub-sample of family executives
85
Table 7: Panel regression of dual class discount on excess corporate cash holdings
87
Table 8: Panel regression of dual class discount on capital expenditure
89
Table 9: Fixed effects regression of dual class discount on excess compensation

and excess corporate cash holdings
90
Table 10: Pair-wise test of separation of voting and cash flow rights on executive compensation:
Fixed effects and 2SLS
93
Table 11: Total compensation based on the Heckman two stage treatment effects
model
95
Table 12: Panel regression of total compensation on management voting leverage,

age and equity ownership
97
CHAPTER 3
Table 1: Descriptive statistics
140
Table 2: Regression of dividend payout on management voting leverage
142
Table 3: Regression of total distribution
143
IX
Table 4: Descriptive statistics for contronlling and non-controlling shareholdersexecutives
144
Table 5: Tobit regression of dividend payout ratios
146
Table 6: Tobit regression of total distribution
147
Table 7: Tobit regression for a sub-sample of family firms only
149
Table 8: Tobit regression for a sub-sample of dual class family firms only
151
Table 9: Tobit regression of industry adjusted cash dividends and total distribution
153
Table 10: Tobit regression with predicted management voting leverage
154
CHAPTER 4
182
Table 2: Effects of entrenchment on discount of dual class Q ratio to that of matching

single class firms
184
Table 3: Effects of entrenchment on discount of dual class Q ratio versus the industry average
185

single class firms' conditional on past performance
186
Table 5: Effects of entrenchment on discount dual class Q ratio versus the industry conditional on
past performance
188
Table 6: Entrenchment effect on dual class discount for a sub-sample of firms with controlling
shareholders-executives
190
Table 7: Entrenchment effect on dual class discount for a sub-sample of firms with
non-controlling shareholders-executives
192
Table 8: Effects of predicted entrenchment on dual class discount conditional

on past performance
194
LIST OF FIGURES
CHAPTER 2
Figure 1: Median total compensation of the top three highest paid executives over
the period 2001-2007
98
Figure 2: Average discount of dual class firms with both classes of shares trading
over the period 1998-2007
99
XI
Chapter 1
An Overview of Corporate Governance
1.0.0
Conceptual Framework of Corporate Governance

In recognition of the key role played by owners, directors and senior officers in the
financial decisions of a company, corporate governance has grown to be an important

subject area within finance. It is a field of research which is multifaceted as it involves
issues related to the board of directors, executive compensation, ownership structures,
market for corporate control, shareholder activism and regulations.
There is no precise definition for corporate governance. The definition depends on
the conceptual framework within which it is being examined. Shleifer and Vishny (1997)
define corporate governance as the ways in which suppliers of capital to corporations assure
themselves of getting a return on their investments. l The Ontario Teachers' Pension Plan
define corporate governance is defined as a system by which companies are directed,
controlled and evaluated. Hence, the primary responsibility for corporate governance lies
mainly with the board of directors. In most definitions of corporate governance several
common concepts tend to resonate, such as accountability, direction, control and evaluation.
Solomon and Solomon (2004) distinguish between the narrow definition of
corporate governance as corporate accountability to shareholders and a broader definition as
a system of checks and balances, both internal and external, which ensure that companies
Denis (2001) defines corporate governance as the set of institutional and market mechanisms that induce selfinterested managers to maximize the value of the residual cash flows of the firm on behalf of its shareholders.
discharge their accountability to shareholders and act in a socially responsible way in all
areas of their business activities. Thus, the breadth of the definition of corporate governance
relates to whether the company should operate for the sole interest of the shareholders or
whether it should safeguard the interest of those that are impacted by it (employees,
customers, the public, government and other stakeholders). From the perspective of nearly
all finance literature and from that of institutional investors, the narrow definition is
preferred. It is the narrow one that will be applied to this dissertation.
Although the term was not explicitly used before the 1960s, corporate governance
issues have been around since the influential writings of Adam Smith. In the 1930s, Berle
and Means (1932) and Dodd (1932), shed light on corporate governance by looking at the
separation of ownership from control. The rise of professional managers and modern
corporations (Chandler, 1962 and 1977) further exacerbate the separation of ownership and
control. Coase (1937) argues that the presence of transaction costs leads to incomplete
contracting between managers and shareholders. Since all possible eventualities cannot be
contracted, managers must use their discretion when dealing with future eventualities. In this
situation, managers can act in the interest of shareholders or in their own interest. Hence
corporate governance mechanisms, such as board of directors, are required to ensure that
managers act in the interest of shareholders.
Agency problems arise from the separation of ownership and control as well as the
limited liability of the modern corporation. As firms grow, the need for capital as well as
professional managers to run these large corporations results in further separation of
2
For further discussion on the rise of professional manager as a dominant force in Modern Corporation, see the
writings of Chandler (1962, 1977).
3
Transaction costs include: i) costs of thinking about all eventualities that can occur during contracting
relationships, ii) negotiating costs and iii) costs of writing contracts which are enforceable by the courts (Hart,
1995).
ownership and control and the accompanying agency costs. Managers (agents) do not
always act in the best interest of the shareholders (principal). Agency costs arise because
shareholders attempt to monitor managers, use incentives and contracts to align managers'
and shareholders' interests (Solomon and Solomon, 2004). The potential conflicts of interest
between managers and shareholders can arise from managers' resistance to being replaced
by a more competent successor, difference in risk borne by shareholders and managers and
diversion of "free cash flow" for the benefit of managers.
Managers and shareholders may face different degrees of risk. This is particularly
true when companies are not closely controlled. In such companies, shareholders normally
have a diversified portfolio of stocks. In comparison, managers' face undiversified risk.
Managers have large human capital tied up in the firm, and often a significant equity
interest. Hence, managers are expected to be more conservative in their decision making
since they have a great deal more at stake. Managers' conservative decision making may
conflict with shareholders' goal of profit maximization. Therefore, managers may be
unwilling to take on a project that is worthwhile from the shareholders' point of view
because they have a lot to lose if the project fails (Denis, 2001). On the other hand, incentive
based compensation such as stock options has the potential to induce unnecessary risktaking behavior by managers. With stock options, managers can benefit greatly from high
risk projects that succeed while facing limited downside risk if the project fails. This is
particularly true if executives have limited equity interest in the firm. Hence, incentive
compensation schemes can serve to increase the alignment of interests between managers
and shareholders or increase risk-taking by executives.
Jensen (1986) was the first to identify the agency costs of "free cash flow". He
argued that managers could use "free cash flow" to pursue value-destroying projects rather
than return the free cash flow to shareholders. Managers may mistakenly believe that the
project is worthwhile or may wish to maximize assets under management (Denis, 2001). As
assets under management increase, managers' gain power and prestige as well as increase
compensation which is usually a function of the firm's size. Managers may also use free
cash flow to undertake "pet" projects which are often value-destroying. Hence, by pursuing
these projects, managers waste "free cash flow" rather than distributing it to shareholders.
There are several potential solutions to agency problems which result from
separation of ownership and control. These include: bonding by the managers, monitoring of
the managers and incentives which align the interests of managers with those of
shareholders.4 The effectiveness of these potential solutions is greatly debated in the
corporate governance literature. In particular, monitoring and alignment of interests occurs
through various corporate governance mechanisms that may help to alleviate agency
problems. For example, legal and regulatory mechanisms as well as incentive based
compensation can align the interests of managers and shareholders. The debate is ongoing
and it is my hope that this dissertation will help to understand the critical role that ownership
and voting control play in determining the value of the firm.
Jensen and Meckling (1976) provide a detailed discussion on how owner-managers can effectively bond
themselves in order to reduce agency costs.
2.0
Corporate Governance Mechanisms
2.1.0
Laws and Regulations
Legal and regulatory corporate governance mechanisms are the most basic external
mechanisms that exist. Regulations and laws have received increasing attention from the
press, researchers and practitioners. One such regulation in the U.S. involved disclosures of
top level executives' compensation by the Securities and Exchange Commission. This has
led to a growing literature that treats executive compensation as both a source and a
potential solution to agency problems. Corporate failures in the U.S. such as Enron and
WorldCom and the financial crisis in several countries around the world have led to an
increasing call for legal and regulatory reforms and oversight. Regulations have been
introduced in many countries to protect minority shareholders. This is especially true in
countries where concentrated control is the dominant form of corporation. Shleifer and
Vishny (1997) argue that in many large corporations around the world the fundamental
agency problem is not the traditional agency problem (between managers and shareholders)
but it is the conflict between outside investors and controlling shareholders. Hence, how best
to protect minority shareholders is an important governance issue.
2.2.0
Hostile Takeovers
Hostile takeovers can be an effective external disciplinary mechanism when

managers are inefficient at running the corporation. The bidding firm or corporate raider
often takes control of the target firm and replaces "inefficient" managers with their own
slate of new executives. In a hostile takeover, the bidder thinks that they can do a better job
than the current management and hence, improve efficiency and increase firm value.
However, researchers are questioning the effectiveness of hostile takeovers as a corporate
governance mechanism due to the sharp decline in takeover activities after 1989. The
decline in hostile takeovers in the U.S. is due to managers lobbying for protection from the
market for control and the demise of the junk bond market (Becht et al., 2003). Also, studies
have shown that the overall gains (target and bidder) are not different from zero (Becht et
al., 2003). This finding is contrary to the popular notion that hostile takeovers correct
managerial inefficiency and enhance value. Often the target shareholders gain at the expense
of the bidder's shareholders. Therefore, hostile takeovers can be ineffective due to the
following: the free-rider problem, the raider may face competition from other bidders and
incumbent management (Hart, 1995).5
2.3.0 Blockholders and Institutional Investors
Outside blockholders are a potential solution to agency problems that arise due to
separation of ownership and control. With large equity interest, blockholders can monitor
managers and influence their decisions by electing board members to act on their behalf. For
example, Chen and Yur-Austin (2007) provide evidence that outside blockholders are more
efficient in reducing managerial discretionary expenses while inside blockholders, especially
managerial blockholders significantly increase the underinvestment problems.6 Several
other studies provide evidence that institutional investors play an important monitoring role
Small shareholders who believe that their decisions are unlikely to affect the success of the bid have an
incentive not to tender to the raider, since they may be able to obtain a pro-rata fraction of the capital gain by
holding on to their shares (Hart, 1995)
6
Chen and Yur-Austin (2007) used Selling, General and Administrative expenses to proxy for managerial
discretionary spending and Market-to-Book value as a proxy for underinvestment.
(Hartzell and Starks, 2003; Bushee, 1998; Wahal and McConnell, 2000; Brickely et al.,
1988; and Agrawal and Mandelker, 1992).
In contrast, blockholders may not be effective monitors if they are able to sell the
holdings in liquid secondary markets. It can be argued that the highly liquid nature of the
U.S. secondary markets, along with regulations, make it difficult for large shareholders to
monitor managers (Mayer, 1988 and Coffee, 1999). Evidence to support this argument is
provided by Kahn and Winton (1998). In addition, Parrino et al., (2003) find that aggregate
institutional ownership and the number of institutional investors decline in the year prior to
forced CEO turnover. This implies that institutional investors prefer to sell their holdings
rather than to monitor and remove poorly performing managers. Furthermore, large
shareholders may use their voting power to improve their own position, agree to leave
managers alone in exchange for having their share repurchased at a premium or they may
take over management of the firm for themselves (Hart, 1995). For these reasons, the
effectiveness of large shareholders as monitors is questionable.
The literature on block ownership is surveyed by Holderness (2003). The evidence
to date indicates that outside blockholders are not that uncommon. First, blockholders seek
to increase value and enjoy private benefits of control not available to other shareholders.
Second, there is some evidence that larger ownership by blockholders has a positive impact
on a firm. For example, the presence of an external blockholder on the board increases the
likelihood of a change in control.
2.4.0
Board of Directors
As an internal governance mechanism, the board of directors performs four major
functions. These include: hiring top management, compensating and evaluating the
performance of top management; voting on major operating proposals (for example, large
capital expenditures); voting on major financial decisions (for example, issuing dividends
and share repurchases) and offering expert advice to management (Kim and Nofsinger,
2004). Another critical role of directors, at least in theory, is the monitoring function of the
board. Directors are expected to monitor managers and fulfill their fiduciary duty by
ensuring that managers act in the interests of shareholders. However, the board of directors
can often be a "rubber stamp" for CEOs. This is evident in some of the most infamous
corporate failures such as Enron and WorldCom. Similarly, when CEOs control the
nomination process and influence the choice of directors, the board of directors can be
ineffective monitors because potential directors will be ones who are more likely to support
the CEO (Becht et al, 2003).
Furthermore, the size of the board has an impact on its effectiveness.7 Larger boards
may have the breadth and expertise to deal effectively with issues confronting the business
because the talent pool is deeper as boards become larger. However, it can be difficult to
keep a larger number of people involved and working efficiently as a team (Colley et al,
2005). With smaller boards, directors can become more focused and can collaborate more
easily. Most of the recent regulations and exchange rules focus on board effectiveness and
independence.
Board size may be a function of complexity of the firm's operations and the type of industry in which it
operates. The typical board size ranges from 8 to 16 directors.
Rules and regulations, such as Sarbanes-Oxley Act, and the NYSE and NASDAQ
corporate governance guidelines, are aimed at improving independence and effectiveness of
the board in the U.S. The Dey report (1994) in Canada and the Cadbury report (1992) in the
U.K. contain similar recommendations. Findings on the effectiveness of independent
directors are mixed. Several studies provide evidence of a negative relationship between the
number of outside directors and company performance in the U.S. (Yermack, 1996; Klein,
1998; Agrawal and Knoeber, 1998). On the other hand, other studies provide evidence that
there is a positive relationship between outside directors and firm performance (Fama, 1980;
Fama and Jensen, 1983 and Rosenstein and Wyatt, 1990). The conflicting result could be
driven by the fact that directors who are considered as outside directors are not truly
independent directors. For example at Disney, several of the outside directors have personal
and related business ties to Michael Eisner, the CEO.8
Hermalin and Weisbach (2003) survey the literature on the board of directors and
highlight some of the common findings. First, smaller boards and boards with a greater
proportion of outside directors result in management taking actions more in line with
shareholders' interest. Second, smaller boards and boards with more "independent" directors
are more likely to remove poorly performing managers. Third, CEO compensation packages
are also more sensitive to performance in firms with smaller boards. Fourth, boards with a
greater proportion of outside directors seem to make better acquisition related decisions.
However, the bulk of the evidence indicates that the proportion of outside directors is not
According to Kim and Nofsinger (2004), Disney claims that 13 of the 16 directors are independent members
of the board. However, of these "outside" directors, Reveta Bowers, is the headmaster of the school Eisner's
children attended. Also, Leo O'Donovan is president of Georgetown University, which one of the Eisner
children attended and which received donations from Eisner. Other "outside" directors such as Irwin Russell is
Eisner's personal attorney and Robert Stern is an architect for several of Disney's projects. It is evident that
these members are not true independent directors since they have ties to the CEO.
related to firm performance. In terms of firm characteristics and board of directors, larger
firms, older firms and firms with small inside ownership stakes have more outside directors.
Closely held firms have insider dominated boards (Hermalin and Weisbach, 2003).
2.5.0
Executive Compensation
Executive compensation in the U.S. and in most countries has increased dramatically
over the past few decades. The primary reasons include: the bull market throughout most of
the 1980s and 1990s, the increased use of stock options and ways in which compensation
packages are determined according to market standards and benchmarking (Becht et al.,
2003). According to the agency theory, incentive based compensation should result in a net
increase in shareholders' wealth. Habib and Ljungqvist (2005) provide evidence to the
contrary. They find that Q-ratio falls as CEO's option holdings increase. However, evidence
in favor of the agency theory comes from event studies by Larcker (1983), DeFusco et al.,
(1990), and Morgan and Poulsen (2001).
Surveys of the executive compensation literature are provided by Murphy (1999)
and Core et al., (2003). The major findings of the papers surveyed by these authors are as
follows: (i) sensitivity of pay-performance has increased over time due to pay packages with
more incentive based compensation which is predominantly option based compensation; (ii)
the evidence on the relationship between managerial equity ownership and performance is
mixed and (iii) stock options are the fastest growing component of CEO compensation. The
authors argue that an increase in option based compensation is driven by several factors.
First, options do not require cash outlay by the firm. Second, options and stocks are treated
as deferred compensation providing tax advantages for both the firm and executives. These
10
findings have important implications for compensation packages which are being used as a
corporate governance mechanism to align the interests of managers with those of
shareholders. Murphy (1999) and Core et al., (2003) argue that we do not fully understand
why stock options have become an increasingly important part of executive compensation
packages.
3.0
Brief history of the dual class market in the United States
In the 1920s, firms in the U.S. began issuing two classes for common shares. This
allowed one group of shareholders to control firms with a disproportionately large number
of voting rights relative to equity (cash flow rights) ownership. In 1926, there were 183
firms with a dual class share structure (Dewing, 1953).9 The New York Stock Exchange
(NYSE) was concerned with the separation of voting and cash flow rights and therefore,
prohibited the listing of non-voting securities in 1926. This prohibition remained in place
until 1985. A few companies such as, Ford Motor Company were able to get around the
NYSE prohibition by issuing a class with inferior voting rights rather than non-voting shares
(Howell, 2009). On the other hand, the American Stock Exchange (AMEX) and National
Association of Securities Dealer (NASD) did not prohibit the listing of firms with dual class
share structure (Bayley, 1989).10 AMEX permitted companies to issue multiple classes of
common stock but only listed designated classes. However, NASD imposed no limitation on
the use of multiple classes of common shares (Bayley, 1989).
For a detailed history of dual class share structure in the U.S. see Howell (2009)
There were only 10 NYSE listed firms with dual class structure in 1985 compared to 60 firms (7% of listed
firms) on AMEX and 110 firms (2.7%) with dual class structure on NASDAQ (Seligman, 1988).
10
11
Because of the trend in dual class capitalization, the NYSE felt increasing
competitive pressure from AMEX and NASD to alter its prohibition on dual-class
mechanisms (Bayley, 1989). Therefore, in September 1986, NYSE sought to suspend its
enforcement of the one-share one-vote standard. However, the Securities Exchange
Commission (SEC) did not approve the NYSE proposed rule change but instead solicited
public comments and held hearings in December, 1986 (Bayley, 1989). As a result, in 1988,
the SEC formally adopted Rule 19c-4. Prior to rule 19c-4, companies commonly introduced
dual class share structures through a recapitalization. However, the Rule 19c-4 prohibits the
creation of dual class structure through recapitalization.11 Rule 19c-4 allows for the creation
of dual class firms via initial public offerings and in 1988 and 2007, 6.7% and 7.2%> of
publicly listed firms used dual class share structures (Howell, 2010). Listing of dual class
firms in the U.S. has remained fairly constant. For example, in 2002, about 6% of all
publicly traded firms in the U.S. (or 362 ) were dual class firms (Gompers, et al., 2010).12
The following characteristics pertain to dual class firms in the U.S. Gompers et al.
(2010) provide evidence that the typical voting structure for firms with two classes of
common share is at a ratio of 10 to 1. In addition, 13% of dual class firms in the U.S. give
the inferior voting class a higher dividend. Typically, the superior voting shares do not trade
but are held by the controlling shareholders. In the U.S., 15% of dual class firms list both
classes for shares. In addition, insiders of dual class firms own a majority of the voting
rights (60%>) and a significant minority of the cash flow rights (40%).
11
Amoako-Adu and Smith (2001) argue that there are few regulations for dual class firms in Canada relative to
other countries. Apart from the coattails provisions, there are no other specific Ontario Securities Act
restrictions and regulations for dual class firms.
12
Howell (2009) provides evidence that dual class firms' listing increased slightly from 6.7% in 1988 to 7.2%
in 2007. In 2002, 9.97% of TSX-listed firms in Canada were dual class firms compared to 5.7% in 2007.
12
A survey by Rydqvist (1992) of the dual-class literature identifies several common

characteristics among dual class firms. Most dual class firms are family-controlled firms
(Smith and Amoako-Adu, 1999 and Cronqvist and Nelisson, 2005) and hence, these firms
are reluctant to undertake seasoned equity offerings. Worldwide, dual class firms are usually
smaller and younger than single class firms (Rydqvist, 1992). However, dual class firms in
the U.S. and Canada are usually larger and older (Smith, et al., 2009 and Gompers et al.,
2010). In addition, dual class firms are more highly leveraged and have a lower value than
single class firms in both Canada and the U.S. (Smith, et al, 2009 and Gompers et al.,
2010).
In the case where both classes of shares are trading, superior voting shares (SV)
trade at a premium. The general argument for the price differential between superior and
restricted voting shares is that voting premiums are related to potential takeover premiums
(Levy, 1982; Horner, 1988; Megginson, 1990; and Smith and Amoako-Adu, 1995). Voting
premiums are significantly reduced when firms are cross-listed in the U.S. (Doidge, 2004).
Voting control has economic value and voting premiums will depend on the voting
differential of the two classes of stocks and the properties of each class of shares in the
firm's equity structure. (Levy, 1982) Voting rights in many countries are worth between
10% and 20% of the value of common equity (Zingales, 1994). Zingales (1994) provides
evidence that voting shares with inferior dividend rights trade at an average premium of
82%) above nonvoting shares in Italy. He argues that, if there were no private benefits, there
would be no reason to hold large blocks of shares in any company. Therefore, concentration
itself is an indication of a large value of control. In a study of British firms, Megginson
(1990) finds that superior voting (SV) shares trade at a premium (13.3%) relative to
13
restricted voting (RV) shares. Similarly, Lease et al., (1983) test whether control is valued
by examining firms with two classes of shares outstanding. Both classes of shares have
identical dividend and distribution rights as well as liquidation rights in the event of
insolvency. They only differ in terms of their voting rights. Price difference reflects
differences in "future benefits". A majority of the firms with superior voting (SV) shares
trade at a premium relative to restricted voting (RV) shares.
The evidence that dual class share structure is detrimental to minority shareholders is
somewhat mixed. For example, Partch (1987) compares managerial ownership before and
after the creation of a class of limited voting common stock for 44 publicly traded firms
between 1962 and 1984, and examines whether the event affects the wealth of current
shareholders. Partch (1987) argues that there is no evidence that current shareholders are
harmed by the creation of limited voting common stock. Similar evidence is provided by
Ang and Megginson (1989) and Cornett and Vetsuypens (1989). On the other hand, Jarrell
and Poulsen (1988) find negative abnormal stock price returns at the announcement of the
dual class recapitalization. In the case of Canada, Jog and Riding (1986) provide similar
evidence. Amoako-Adu and Smith (2001) present numerous legal disputes between
controlling and outside shareholders. Such disputes arise because of the agency problems
which tend to be associated with dual class shares. Agency problems are a major factor
which may explain the observed fact that dual class firms sell at a valuation discount
compared to a control sample of single class firms with concentrated ownership.
14
4.0
Further areas of research and unresolved questions
Holderness (2003) argues that we need to fully understand the relationship between
control and fractional ownership, whether by inside or outside blockholders. I attempt to
address this issue from the perspective of concentrated control and ownership structure.
There is substantial and fairly consistent evidence that dual class structure is associated with
lower valuation. I want to address whether this discount is associated with the extraction of
private benefits of control by examining channels through which private benefits can be
extracted. Also, dual class ownership structures can allow managers to become entrenched.
Therefore, investors discount the value of dual class firms relative to single class companies
with concentrated control.
The existing evidence on many of the corporate governance mechanisms fails to
establish a convincing link between these mechanisms and firm value. One possible reason
is that corporate governance systems are not important enough to have a meaningful impact
on firm value. However, it can be argued that various corporate governance mechanisms
may interact in complicated ways with each other and other aspects of the firms. Therefore,
a particular mechanism valuable for one group of firms may not have an effect on another
group. Hence, it is difficult to identify certain relationships by looking at a broad crosssection of firms (Denis (2001). I propose to take a narrower focus by separating firms into
different ownership structures such as dual and single class with concentrated control and
investigate how differences in ownership structure and control affect dividend policy. In this
way, I will hold constant the impact of concentrated control and isolate the impact of dual
versus single class share structure.
15
In addition, we do not know whether dual class firms possess other countervailing
governance mechanisms such as outside directors, stronger pay-for-performance or strong
monitoring by outside blockholders, which may decrease potential agency conflicts. By
pursuing research on closely controlled companies, we can increase our understanding of
this particular group of firms and increase our ability to better regulate large or active
shareholders to obtain the right balance between managerial discretion and minority
shareholder protection.
The first dissertation essay titled, Dual Class Discount, Agency Problems and
Wealth Expropriation, is presented in Chapter 2. The aim of the essay is to address the
following question: Is the valuation discount of dual class companies compared to
concentrated control single class firms due to greater extraction of private benefits? The
valuation discount of dual class companies is documented in the literature (King and Santor,
2008; Gompers et al., 2010; and Smith et al, 2009). The general conclusion of these and
other studies is that the discount reflects the extraction of private benefits. However, there is
no systematic evidence showing the channels of extraction of private benefits which may
result in the valuation discount between dual class companies and single class companies
with concentrated control. In addressing the above question, I examine three channels
through which private benefits of control can be extracted using a panel dataset of 792
firm-year observations from the S&P 1500 group of firms that have dual class share
structure and a propensity score matched sample of single class firms with concentrated
ownership. These channels are: excess compensation to all senior executives, excess cash
holdings and capital expenditure.
16
Using a panel regression with industry and year fixed effects, the main result is
that investors believe that controlling shareholders and managers in dual class firms are
using their voting control to extract private benefits and therefore, investors discount the
value of dual class companies. I demonstrate that the greater the excess compensation and
excess cash holding, the larger the valuation discount of dual class companies. First, after
controlling for firm characteristics and governance determinants of executive compensation
based on prior literature, I find that the separation of voting and cash flow rights leads to
higher executive compensation for all executives. Furthermore, family executives received
significantly higher compensation compared to non-family executives in the same position.
In a sub-sample of family executives only, I show that family executives in dual class firms
extract higher compensation relative to family executives in single class concentrated
control firms. Second, I find that dual class firms retain more cash holdings compared to
single class firms. Excess cash holdings, in turn, lead to a larger valuation discount of dual
class companies. The results are consistent with investors' belief that managers and
controlling shareholders of dual class firms are using excess cash in pursuit of private
benefits.
The second dissertation essay titled Ownership Structure, Agency Problems and
Dividend Policy is presented in Chapter 3. In this essay, I investigate how concentrated
control through dual versus single class share structure affects dividend policy. I use a
propensity score matched sample of dual and single class companies with concentrated
control. I propose three competing explanations of dividend policy in firms with
concentrated control. The extraction of private benefits hypothesis, the family legacy
hypothesis and the managerial reputation hypothesis are proposed and tested. The
17
extraction of private benefits hypothesis states that managers of dual class firms set a low
payout policy in order to retain resources within firms which can be extracted as private
benefits. The family legacy hypothesis predicts that controlling shareholders in dual class
firms are more likely to maximize firm value and hence, are less likely to extract wealth
from the firm for their own benefits. Therefore, firms that are family controlled may retain
resources in order to ensure survival and growth of the firm which is beneficial to several
generations of family members. The managerial reputation hypothesis states that investors
are concerned about the extraction of private benefits. Therefore, higher dividend payout
commits the firm to raise capital more frequently and hence, the firm is subjected to
increased scrutiny by investment professionals, investors and the capital market (Rozeff,
1982 and Easterbrook, 1984).
Using Tobit estimation and panel regression with industry and fixed effects, the
major findings are as follows: first, dual class firms tend to payout less dividend and
repurchase fewer shares. Second, the greater the divergence between voting and cash flow
rights, the lower the cash dividends and total distribution. This is consistent with both the
extraction of private benefits and family legacy hypotheses. Third, using excess
compensation of the controlling shareholders-executives, I show that lower payout policy is
consistent with the extraction of private benefits of control hypothesis.
The third dissertation essay titled, The Valuation Effects of Managerial
Entrenchment on Dual Class Discount is presented in Chapter 4. In this essay, I
investigate the relationship between managerial entrenchment and the documented dual
class discount. Managerial entrenchment is expected to reduce the value of the firm as there
is a lack of discipline on managers to address the poor performance. Several studies provide
18
evidence indicating that dual firms are discounted compared to single class firms (King and
Santor, 2008, Gompers et al., 2010, Smith et al., 2009). One possible explanation for the
dual class discount is managerial entrenchment. Entrenched managers are more likely to
extract private benefits of control and therefore, investors are likely to discount dual class
firms relative to single class concentrated control companies.
Using a sample of dual class firms and a propensity matched sample of single class
companies with concentrated control, I show that CEOs and directors in dual class firms are
more entrenched. CEOs and directors of dual class firms tend to have longer tenure
compared to their counterparts in single class companies. Furthermore, Entrenchment
should be defined in the context of poor performance by management and hence, it is
important to account for this performance. Conditional on poor performance, I show through
panel regression that dual class firms with excess CEO tenure, E-index or excess directors'
tenure are discounted more by investors. This implies that investors price the impact of
managerial entrenchment in firms with dual class ownership structure.
19
5.0
References
Agrawal, A., and Knoeber, C. R., (1998), Managerial compensation and the threat of
takeover, Journal of Financial Economics, 47, 219-239
Agrawal, A., and Mandelker, G., (1992), Shark repellents and the role of institutional
investors in corporate governance, Managerial and Decision Economics, 13, 15-22
Amoako-Adu, B., and Smith, B. F., (2001), Dual class firms: capitalization, ownership
structure and recapitalization back into single class, Journal of Banking and
Finance, 25, 1083-1 111
Ang, J. S., and Megginson, W. L., (1989), Restricted voting shares, ownership structure and
market value of dual-class firms, Journal of Financial Research, 12, 301-318
Bayley, K., (1989), Rule 19c-4: The death knell for dual-class capitalization, The Journal of
Corporation Law, 15, 1-25
Becht, M., Bolton, P., and Roell, A., (2003), Corporate governance and control, Handbook
of Economics of Finance, 1, 1-109
Berle, A. and Means, G., (1932), The modern corporation and private property, Macmillan,
New York
Brickely, J., Lease, R., and Smith, C, (1988), Ownership structure and voting in
antitakeover amendments, Journal of Financial Economics, 20, 267-291
Bushee, B., (1998), The influence of institutional investors on myopic R&D investment
behavior, The Accounting Review, 73, 305-333
Cadbury, Sir Adrian, (1992), Study group on directors' remuneration, Final Report
Canada, Toronto Stock Exchange Committee on Corporate Governance in Canada, 1994,
Where were the directors? (Dey Report), Toronto stock Exchange: Toronto,
December 1994
Chandler, A. D., Jr., (1962), Strategy and Structure: Chapters in the history of the American
industrial enterprise. Cambridge, MA: MIT Press
Chandler, A. D., Jr., (1977), The Visible hand, Cambridge, Mass. and London, England:
The Belknap Press of Harvard University Press
Chen, X., and Yur-Austin, J., (2007), Re-measuring agency costs: The effectiveness of
blockholders, The Quarterly Review of Economics and Finance, 47, 588-601
20
Coase, R., (1937), The nature of the firm, Economica, n.s., 4, 386-405
Coffee, J. C, (1999), The future as history: The prospects for global convergence in
corporate governance and its implications, Northwestern University Law Review, 93,
641-708
Colley J. L. Jr., Doyle J. L, Logan G. W. and Stettinius, W., (2005), What is corporate
governance? McGraw-Hill, New York
Core, J. E., Guay, W., and Larcker, D. F., (2003), Executive equity compensation and
incentives: A survey, Economic Policy Review, 9, 27-50
Cornett, M. M., and Vetsuypens, M. R., (1989), Voting rights and shareholder wealth: The
issuance of limited voting common stock, Managerial and Decision Economics, 10,
175-188
Cronqvist, H., and Nilsson, M., (2005), The choice between rights offering and private
equity placements, Journal of Financial Economics, 78, 375-407
DeFusco, R. A, Johnson, R. R., and Zorn, T. S., (1990), The effect of executive stock option
plans on stockholders and bondholders, Journal of Finance, 45, 617-627
Denis, D. K, (2001). Twenty-five years of corporate governance research and counting,
Review of Financial Economics, 10, 191-212
Dewing, A. S., (1953), The financial policy of corporations, New York: The Ronald Press
Company
Dey Report (Dec. 1994), Where were the directors? Toronto Stock Exchange Committee on
Corporate Governance in Canada, Toronto
Dodd, E. M., (1932), For whom are corporate managers trustees?, Harvard Law Review, 45,
1145-63
Doidge, C, (2004), U.S. cross-listing and private benefits of control: Evidence from dualclass firms, Journal of Financial Economics 72, 519-553.
Easterbrook, F., (1984), Two agency cost explanations of dividends, American Economic
Review, 14, 650-659
Fama, E. F., (1980), Agency problems and the theory of the firm, Journal of Political
Economy, 88, 288-307
Fama, E. F., and Jensen, M. C , (1983), Separation of ownership and control, Journal of Law
and Economics, 26, 301-325
21
Freeman, E. R., (1984), Strategic management: A stakeholder approach. Boston: Pitman.

Gompers, P. A., Ishii, J., and Metrick, A., (2010), Extreme governance: An analysis of dualclass firms in the U.S., Review of Financial Studies, 23, 1051-1088
Habib, M. A., and Ljungqvist, A. P., (2005), Firm value and managerial incentives: A
stochastic frontier approach, Journal of Business, 78, 2053-2093
Harris, M , and Raviv, A., (1988), Corporate governance: Voting rights and majority rules,
Journal of Financial Economics, 20, 203-235
Hart, O., (1995), Firms, contracts, and financial structure. Oxford, Oxford University Press
Hart, O., (1997), Corporate governance: Some theory and implications, The Economic
Journal, 105, 678-689
Hartzell, J. C. and Starks, L. T., (2003), Institutional investors and executive compensation,
Journal of Finance, 58, 2351-2374
Hermalin, B. E., and Weisbach, M. S., (2003), Boards of directors as an endogenously
determined institution: A survey of the economic literature, Economic Policy
Review, 9, 7-26
Holdemess, C. G, (2003), A survey of blockholders and corporate control, Economic Policy
Review, 9, 51-64
Holdemess, C. G., and Sheehan, D. P., (1988), The role of majority shareholders in publicly
held corporations: An exploratory analysis, Journal of Financial Economics, 20,
317-346
Homer, M , (1988), The value of the corporate voting right: Evidence from Switzerland,
Journal of Banking and Finance, 12, 69-83
Howell, J., (2009), Essays on the U.S. dual class share structure, PhD Dissertation
University of Georgia
Jarrell G., and Poulsen, A., (1987), Shark repellents and stock prices: The effects of
antitakeover amendments since 1980, Journal of Financial Economics, 19, 127-168
Jensen, M. (1986), Agency cost of free cash flow, corporate finance and takeovers,
American Economic Review, 76, 323-329
Jensen, M. C , and Meckling, W. H., (1976). Theory of the firm: Managerial behavior,
agency costs and ownership structure, Journal of Financial Economics, 3, 305-360
22
Jog, V. M., and Riding A. L., (1986), Price effects of dual-class shares, Financial Analyst
Journal, 86, 58-66
Kahn, C, and Winton, A., (1998), Ownership structure, speculation and shareholder
intervention, Journal of Finance, 53, 99-129
Kim, K. A., and Nofsinger, J. R., (2004), Corporate governance. New Jersey: PearsonPrentice Hall
King, M , and Santor, E., (2008), Family values: Ownership structure, performance and
capital structure of Canadian firms, Journal of Banking and Finance, 32, 2423-2432
Klein, A., (1998), Firm performance and board committee structure, Journal of Law and
Economics, 41, 275-30
Larcker, D. F., (1983), The Association between performance plan adoption and corporate
capital investment, Journal ofAccounting and Economics, 5, 3-30
Lease, R. C , McConnell, J. J., and Mikkelson, W. H., (1983), The market value of control
in publicly-traded corporations, Journal of Financial Economics, 11, 439-491
Levy, H., (1982), Economic evaluation of voting power of common stock, The Journal of
Finance, 38, 79-93
Mayer, C , (1988), New issues in corporate finance, European Economics Review, 32, 11671183
Megginson, W., (1990), Restricted voting stock, acquisition premiums and the market value
of corporate control, The Financial Review, 25, 175-198
Morgan, A. G., and Poulsen, A. B., (2001), Linking pay to performance: Compensation
proposals in the S&P 500, Journal of Financial Economics, 62, 489-523
Murphy, K., (1999), Executive compensation. In. O. Ashenfelter and D. Card (eds),
Handbook of labor economics, 3, Amsterdam: North-Holland
Parrino, R., Sias, R.W., and Starks, L. T., (2003), Voting with their feet: Institutional
ownership changes around forced CEO turnover, Journal of Financial Economics,
68, 3-46
Partch, M., (1987), The creation of a class of limited voting common stock and shareholder
wealth, Journal of Financial Economics, 18, 313-339
Rosenstein, S., and Wyatt, J. G., (1990), Outside directors, board independence and
shareholder wealth, Journal of Financial Economics, 26, 175-191
23
Rozeff, M. S., (1982), Growth, beta and agency costs as determinants of dividend payout
ratios, Journal of Financial Research, 5, 249-259
Rydqvist, K., (1992), Dual-class shares: A review, Oxford Review of Economic Policy, 8,
45-57
Seligman, J., (1986), Equal protection in shareholder voting rights: The one common share,
one vote controversy, George Washington Law Review, 54, 687-725
Shleifer, A., and Vishny, R., (1986), Large shareholders and corporate control, Journal of
Political Economy, 95, 461- 488
Shleifer, A., and Vishny, R., (1997), A survey of corporate governance, Journal of Finance,
52, 737-784
Smith B. F., and Amoako-Adu, B., (1995), Relative Prices of dual-class shares, Journal of
Financial and Quantitative Analysis, 30, 223-239
Smith B. F., and Amoako-Adu, B., (1999), Management succession and financial
performance of family controlled firms, Journal of Corporate Finance, 5,341-368
Smith, B. F., Amoako-Adu, B., and Kalimipalli M , (2009), Concentrated control and
corporate value: A comparative analysis of single and dual class structures in
Canada, Applied Financial Economics, 19, 955-974
Solomon, J., and Solomon, A., (2004), Corporate governance and accountability, West
Sussex, England: John Wiley and Sons
Wahal, S., and McConnell, J.J., (2000), Do institutional investors exacerbate managerial
myopia ?, Journal of Corporate Finance, 6, 307-329
Yermack, D., (1996), Higher market value of companies with small board of directors,
Zingales, L., (1994),The value of the voting right: a study of the Milan Stock Exchange
experience, Review of Financial Studies, 7, 125-148
24
Chapter 2
Dual Class Discount, Agency Problems and Wealth Expropriation
1.0
Introduction
A dual class company is a firm with two or more classes of common shares
outstanding (superior voting shares and restricted voting shares).13 A dual class ownership
structure allows an individual or a related group of individuals to control the firm with a
significant proportion of votes, but with a small proportion of the total equity. As a result,
dual class ownership structure can exacerbate the agency conflict between controlling
shareholders-managers and minority shareholders. The agency costs of dual class companies
is due to monitoring costs associated with two groups of non-controlling shareholders,
minority superior voting shareholders and non-voting or restricted voting shareholders.
Similarly, agency problems can arise in single class concentrated control companies due to
monitoring costs and conflicts of interest between controlling and non-controlling
shareholders. However, alignment of interest is stronger compared to dual class firms due to
the larger equity ownership in single class concentrated control companies.14 This may
reduce the impact of agency problems on firm value. Agency costs are expected to be
higher in dual class firms because insiders have to deal with both the minority shareholders
13
Restricted voting shares (RV) are defined as common shares with votes ranging from no vote to one vote per
share. Superior voting shares (SV), on the other hand, are common shares with more votes per share than
restricted shares. A typical superior voting share has a 10:1 ratio that is, 10 votes per share (Gompers et al.,
2010).
14
Dual class ownership structure can lead to weaker alignment of interest among shareholders compared to
single class closely-held companies due to the divergence of voting and cash flow rights.
25
of the multiple voting shares and the shareholders of the restricted voting shares. Given the
greater divergence in interests of controlling and outside shareholders, agency costs
associated with monitoring and disciplining are expected to be higher in dual class firms
(Correia da Silva et. al., 2004). The argument that dual class companies may have bigger
agency problems is made by Grossman and Hart (1988), Bebchuk et al. (2000), Correia da
Silva et al. (2004) and King and Santor (2008).
Controlling shareholders-managers in dual class companies can use their voting
control to influence their compensation, resulting in the extraction of private benefits and
agency costs. These controlling shareholders-managers, by virtue of their voting power, can
influence the composition of the board of directors and therefore influence the board's
effectiveness as a governance mechanism. Furthermore, controlling shareholders-managers
in single class companies with concentrated ownership can also use their equity ownership
to influence the board of directors and therefore, they can influence their compensation and
extract private benefits of control in the form of perquisites and higher compensation.15
However, extraction of private benefits may be greater in dual class companies because of
the smaller proportion of equity ownership relative to control rights. Therefore, as the
divergence between the manager's voting and equity interest widens, the incentive for the
manager to expropriate private benefits from the corporation increases (Grossman and Hart,
1988 and Bebchuk et al., 2000). Hence, the value of dual class firms may be lower
compared to single class concentrated control companies due to weaker alignment of
interest, extraction of private benefits and higher agency costs.
15
Following La Porta et al., (1999) and Claessens et al., (2000), a single class firm with concentrated control is
defined as a firm with ownership of 15% of the shares outstanding by an individual, a family or an institution.
26
The main objective of this essay is to analyze the documented valuation discount of
dual class firms and to relate the discount to three direct avenues through which controlling
shareholders and managers can extract private benefits from dual class companies. The
means through which controlling shareholders of dual class companies can expropriate
wealth from the firm are: excess executive compensation, capital expenditure and excess
cash holdings. Using these mechanisms, the paper investigates whether investors are aware
of the agency costs and potential extraction of wealth by controlling shareholders and
managers from dual class companies and as a result, discount the value of dual class
companies compared to single class concentrated control companies. Therefore, the aim of
this research is to address the following question: Is the valuation discount of dual class
companies compared to concentrated control single class firms due to greater extraction of
private benefits? The valuation discount of dual class companies is documented in the
literature (King and Santor, 2008; Smith et al., 2009 and Gompers et al., 2010). The general
conclusion of these and other studies is that the discount reflects extraction of private
benefits. However, there is no systematic evidence showing the channels of extraction of
private benefits which may result in the valuation discount between dual class companies
and single class companies with concentrated control.
Masulis et al. (2009) are the first to examine the channels through which managers
can pursue private benefits and provide evidence that managerial extraction of private
benefits is influenced by divergence of control and cash flow rights in a sample of dual class
firms. They find that as divergence in voting and cash flow rights increases, corporate cash
holdings are worth less to outside shareholders, CEOs of dual class companies receive
higher levels of total compensation, managers undertake value-destroying acquisition more
27
frequently and capital expenditure contributes less to shareholder value. Hence, managers
with greater control rights in excess of cash flow rights are prone to wasting corporate
resources to pursue private benefits at the expense of shareholders.
My research differs from Masulis et al. (2009) in several important aspects. First, I
directly relate dual class discount to the channels through which controlling shareholdersmanagers can extract private benefits and I test the rational investor hypothesis. Under this
hypothesis, investors expect executives in dual class companies to use their voting power to
extract private benefits and as a result, they discount the value of dual class companies.
Second, I argue that controlling shareholders and managers in single class concentrated
control companies can also use their ownership control to extract private benefits. Therefore,
it is important to examine the channels of wealth expropriation by comparing concentrated
control in dual class companies with concentrated control in single class companies. The
difference between dual class and concentrated control single class is how control is
achieved. In dual class firms, control occurs through superior voting shares with
proportionally less equity ownership. Therefore, controlling shareholders and managers in
dual class companies are more likely to use their voting power to extract private benefits. In
comparison, control in single class firms occurs through significant equity ownership and
hence, executives in single class concentrated control companies are less likely to extract
private benefits. Third, unlike Masulis et al. (2009), I examine all the senior level
executives' compensation, not only the CEO compensation. This is important since family
involvement tends to be higher in dual class companies and excess compensation paid to
these family members would be a perquisite flowing from control. In fact, 83.2% of dual
28
class firms have some degree of family involvement.16 For example, 45.9% of dual class
executives who serve as both CEO and President are family members, compared to 17.6%
in single class companies. It remains to be seen whether family executives in dual class
companies use their voting power to influence their compensation and whether their
compensation is excessive compared to family executives in single class concentrated
control companies.
Hence, excess compensation paid to family executives will provide
further evidence of private benefits of control and agency costs associated with dual class
ownership structure.
Exorbitant CEO pay packages have been widely regarded as a major form of private
benefits and are a symbol of poor governance (Masulis et al., 2009). Therefore, it is possible
that investors are aware of this and as a result, discount the value of dual class companies
compared to single class concentrated control companies. Similarly, excess corporate cash
holdings and excess capital expenditure are other mechanisms through which managers can
pursue private benefits of control. However, it is not clear whether investors, knowing that
controlling shareholders and managers can potentially misuse excess corporate cash
holdings, discount the value of dual class companies. The prior literature does not examine
whether dual class companies compared to single class concentrated control firms retain
excess cash, or pay their executives excess compensation.
Using a propensity score matched sample of dual and single class concentrated
control companies, I provide evidence to support the argument that investors believe that
controlling shareholders and managers in dual class firms are using their voting control to
16
Family involvement typically includes the following: family members as senior executives of the company,
as consultants to the company, as directors of the company or as controlling shareholders.
17
A family executive is defined as an individual who is a member of the senior management team and is related
to the controlling shareholder or founding family, either by blood or marriage.
29
extract private benefits and therefore, investors discount the value of dual class companies. I
demonstrate that the greater the excess compensation and excess cash holding, the larger the
valuation discount of dual class companies after controlling for factors that have been
proven to partially explain the dual class discount. First, after controlling for firm
characteristics and governance determinants of executive compensation based on prior
literature, I find that the separation of voting and cash flow rights leads to higher executive
compensation for all executives. This implies that executive compensation is one avenue
through which managers are extracting private benefits. Since dual class companies
frequently employ family members, excess compensation paid to family executives allows
us to test whether dual class ownership structure and family involvement leads to extraction
of private benefits of control and agency costs. I find that family executives received
significantly higher compensation compared to non-family executives in the same position.
In a sub-sample of family executives only, I show that family executives in dual class firms
extract higher compensation relative to family executives in single class concentrated
control firms.
Second, corporate cash holdings provide managers with the most latitude as to how
and when to spend it (Masulis et al., 2009). Univariate tests show that dual class companies,
on average, tend to retain excess cash compared to single class concentrated control
companies. This excess cash can be easily diverted or misused by controlling shareholders
and managers. Hence, managers may spend part or all of the excess cash in order to pursue
private benefits such as perquisite consumption. Investors, knowing this, may value
corporate cash holdings of dual class firms less compared to cash holdings of single class
companies. Consistent with this prediction, I find that excess cash holdings lead to a larger
30
valuation discount of dual class companies. However, capital expenditure does not explain
dual class discount. This implies that investors do not view capital expenditure as a potential
avenue of extraction of private benefits.
2.0
Dual Class Discount
Using a sample of only dual class firms with both classes of shares trading,
empirical studies generally find that superior voting shares trade at a premium or
alternatively, restricted voting shares trade at a discount, relative to superior voting shares.18
The general argument for the price differential between superior and restricted voting shares
is that voting premiums are related to potential takeover premiums and private benefits of
control (Levy, 1983; Lease et al., 1983; Horner, 1988; Megginson, 1990; Rydqvist, 1992;
Smith and Amoako-Adu, 1995, and Zingales, 1995).
Equally important, other studies examine the impact of dual class share structure on
the overall firm value using a proxy for Tobin's Q as a valuation measure. These studies
find that the greater the divergence between voting and cash flow rights, the lower the value
of dual class firms compared to single class companies (King and Santor, 2008; Gompers, et
al., 2010; Smith et al., 2009). Furthermore, family control of dual class firms can exacerbate
the agency problems and lead to a greater discount. In the Canadian context, King and
Santor (2008) find that family ownership and dual class share structure leads to a 17%
discount relative to other firms. In addition, Smith et al., (2009) examine dual class
companies compared to single concentrated control companies and find that the value of
dual class companies is discounted relative to single class concentrated control companies.
18
The sample size in these studies is generally small as only a small fraction of dual class firms have both
classes of shares trading.
31
This discount can be explained by private benefits of control and agency problems of dual
class ownership structure. Therefore:
& Single Class Concentrated Control
QuualClass + DiSCOUYlt
(1)
Thus, the dual class discount is the valuation wedge between dual class companies
and single class concentrated control firms. Based on the previous literature, it can be argued
that dual class discount is reflective of the agency problems of dual class ownership
structure and potential wealth expropriation in the form of private benefits.
Discount
= \QDuai class ~ QSingle Class Concentrated Control )
d),
where Q is a proxy for Tobin's Q ratio.

Using the definition of "free cash flow" and considering the cash flow differential
between dual and single class firms,
Dual Class Discount = /{[(R - OC - D)(\ -1) + D\-ACapex-
AWorkingCapital}
(3),
where R is defined as total revenue, OC is operating costs including executive

compensation, D is depreciation, t is corporate tax rate, ACapex and AWorking Capital are
allocations for capital expenditure and net working capital, respectively. From the above
definition, we can identify potential avenues through which private benefits can be extracted
from dual class companies by insiders. These channels are: executive compensation which
32
is part of the operating costs, capital expenditure (Capex) and cash and near cash which are
part of the working capital.
Alternatively, following the several studies on diversification discount such as Lang
and Stulz (1994), Berger and Ofek (1995) and Villalonga (2004), I estimate the value effects
of dual class ownership structure based on the industry adjusted difference between the dual
class firm's Q ratio and the industry average Q ratio.19 Following Villalonga (2004), I
compute industry means and medians using 4-digit and 3-digit SIC codes for which there is
a minimum of five single class firms in the industry for given year.20
3.0
Literature Review and Hypothesis Development
3.1.0
Firms with concentrated control may be subject to agency costs arising from
conflicts of interest between minority and majority shareholders (Cheung et al., 2005).21
Controlling shareholders can use their influence to extract private benefits of control at the
expense of minority shareholders. A potential avenue for such an extraction is through
executive compensation because controlling shareholders-executives or owners-directors
effectively set their own compensation. Empirical evidence relating equity ownership and
19
Studies by Lang and Stulz (1994), Berger and Ofek (1995) and Villalonga (2004) examined the
diversification discount, that is the excess value between multi-segment (diversified) firms and single-segment
firms.
20
In 9% of the cases, I relied on 2-digit SIC codes to compute the industry averages, in 33% of the cases I used
3-digit SIC codes and in 68% of cases I used 4-digit SIC codes.
21
Minority shareholders are defined as dispersed shareholders with a small fraction of the outstanding shares or
where their total votes are less than 15%. A controlling shareholder is a shareholder with more than 15% equity
ownership or voting rights.
33
executive compensation is mixed.
Prior studies show that top executives owning a
majority block of common shares receive higher salaries and bonuses than top executives in
similar sized, but diffusely held firms (Holderness and Sheehan, 1988). Cheung et al. (2005)
find a positive relationship between cash emoluments received by the CEO and the
Chairman and their respective shareholding for levels of ownership of up to 35% in small
firms and up to 10% in large firms.23
On the other hand, several studies examining ownership and executive
compensation find that greater ownership concentration lowers the ability of executives to
extract higher levels of compensation (Dyl, 1988; Goldberg and Idson, 1995; Hartzell and
Starks, 2003; Haid and Yurtoglu, 2006).24 For example, Dyl (1988) provides evidence that
corporate control in firms with concentrated ownership is negatively related to total CEO
compensation.
This implies that in these firms, major shareholders engage in monitoring
activities which reduce agency costs. In addition, Hambrick and Finkelstein (1995) argue
that when externally-controlled owners have direct or indirect influence over CEO pay, they
will have a pay philosophy which is to minimize CEO pay subject to the ability to attract
and retain a "satisfactory" CEO. Ownership concentration in these studies predominantly
occurs through significant equity ownership in single class firms. There is little evidence
examining concentrated control where control occurs through voting power with limited
equity investment (dual class share structure).
22
See Murphy (1999) and Core et al. (2003) for surveys of the executive compensation literature.
Cheung et al., (2005) examine cash emoluments and dividends as income to CEOs and Chairmen. They did
not examine other senior level executives or other forms of compensation such as options.
24
The firms examined by Haid and Yurtoglu (2006) are closely-held with an average ownership of 53% by the
largest shareholders and 65% by the three largest shareholders. The derivation of voting rights and cash flow
rights in their sample occurs primarily through pyramid structure. Hartzell and Starks (2003) provide evidence
that institutional ownership is negatively related to the level of compensation for the top management as a
group.
25
Dyl (1988) utilizes the percentage of the total shares outstanding owned by the five largest shareholders as a
measure of concentrated control.
23
34
Denis (2001) argues that it may be difficult to identify the effects of various
corporate governance mechanisms because they interact with other firm characteristics in
complicated ways. In studies of corporate governance, one of the fundamental
distinguishing characteristics among firms is whether they are closely controlled or widely
held. Along this line of argument, instead of analyzing executive compensation in a broad
cross-section of firms, by examining only closely controlled companies, I can better identify
the effects of dual and single class ownership structures on extraction of private benefits in
the form of excess executive compensation. It is not clear whether executives in dual class
firms pay themselves more than executives in single class firms with concentrated control.
According to the private benefits of control hypothesis, executives in dual class firms will
tend to extract private benefits. This can be in the form of excess compensation relative to a
control group of firms.
HI:
In dual class firms, executives have the ability to extract excess
compensation. After controlling for firm characteristics and governance attributes, I
expect executives in dual class firms to have higher compensation compared to
executives in single class concentrated control firms.
3.2.0 Family Executives and Compensation
The ownership concentration and management in the hands of a family gives a lot of
power to the family executives. It enables them to take actions that are beneficial to
themselves which may be detrimental to minority shareholders. Family executives, because
of their control, are more likely to receive higher compensation compared to non-family
executives due to their influence over the board of directors. They may use their power to
26
Following Anderson and Reeb (2003) and Villalonga and Amit (2005), a family firm is defined as one in
which the founder or a member of his or her family, by either blood or marriage, is an officer, director or
35
influence the remuneration process thereby extracting benefits at the expense of

shareholders (Bebchuk, et al., 2002). On the other hand, in family firms, the firm symbolizes
the heritage and tradition of their family and is often a part of their identity. Therefore,
family executives are unlikely to act against the interests of the company because in doing
so, they may also harm themselves. It is not clear whether the same argument holds for both
dual class and single class firms. Family executives in firms with concentrated ownership
and control are in a position to exert their influence on the composition of the board and its
various committees. Therefore, concentrated control may allow executives to extract private
benefits. Hence, family ownership and control can affect family and non-family executive
compensation in very different ways.
The evidence on the relationship between family ownership and compensation is
mixed. There is evidence that family CEOs earn less compensation compared to non-family
CEOs in a group of family firms (McConaughy, 2000 and Gomez-Mejia et al., 2003).
Several studies, on the other hand, provide evidence that family executives earn higher
compensation compared to executives in non-family controlled firms (Cheung et al., 2005;
Basu et al., 2007; and Cohen and Lauterbach, 2008).27 Cheung et al. (2005) provide
evidence that family control is associated with higher executive compensation when the
CEO and Chairman have significant shareholdings in the firm and with lower executive
compensation when they hold a small percentage of shares outstanding.
blockholder, either individually or as a group. In addition, if there are firms with co-founders, I will consider
these firms as family firms.
27
Basu et al., (2007) provide evidence that the founding families retain strong ownership and control rights in
Japanese firms. Top executives in Japanese firms earn more in family-controlled firms than firms lacking such
influence. In a study of Israeli firms, Cohen and Lauterbach (2008) find that CEOs who belong to a family or a
business group that own most of the firm shares receive significantly (about 50%) higher pay than professional
CEOs who do not belong to the control group.
28
Cheung et al., (2005) used an indicator variable to identify whether a firm has family ownership greater than
a certain percentage. For example, when family ownership is over 30% or over 50% and executive ownership is
36
The few studies analyzing dual class share structure and compensation did not
separately examine executive compensation of family versus non-family executives in these
companies. Family executives in dual class companies are expected to behave differently
because they control their companies with a larger proportion of voting rights relative to
their equity ownership. Due to the proportionately smaller cash flow consequence to
themselves, executives who are family members in dual class companies will find it more
attractive to use their control to extract resources from the firm. One possible way of
extracting resources is through excess compensation. If the private benefits of control
hypothesis is true and family executives extract resources in the form of excess
compensation, then I expect (1) higher compensation in firms with family executives
relative to other executives in a similar position and (2) higher compensation for family
executives in dual class firms relative to family executives in single class concentrated
control firms. Therefore:
H2a: Family members of the controlling shareholder who are executives are
expected to earn higher compensation compared to non-family executives in similar
positions in non-family controlled firms.
H2b: Family members who are executives in dual class firms are expected to pay
themselves higher compensation relative to family members in single class
concentrated control firms.
3.3.0 Dual Class Discount and Corporate Cash Holdings
Agency theories can be used to explain why corporate cash holdings are valued less
in countries with poor shareholder protection. Managers and controlling shareholders in
10% to 35% the interaction term between family dummy variable and executive ownership is positively related
to compensation. One possible reason is that family members in family firms are also executives in these firms.
However, Cheung et al., (2005) did not explicitly examine compensation of family executives in their study.
Furthermore, it is possible that outside CEOs are appointed to run family firms.
37
these countries can extract private benefits using corporate cash holdings and use "free cash
flow" to pursue value-destroying activities such as acquisition (Jensen, 1986) or to obtain
perquisite consumption at the expense of minority shareholders. Harford (1999) provides
evidence that cash-rich firms are likely to undertake value-destroying acquisition.
Therefore, when managers control the firm, either by equity ownership or by votes,
corporate cash holdings can influence firm value. The evidence in the literature shows that
the contribution of corporate cash holdings to firm value is lower in countries with weak
investor protection (Pinkowitz et al., 2006; Kalcheva and Lins, 2007; and Harford et al.,
2008). However, Kalcheva and Lins (2007) also find that when external shareholder
protection is strong, cash held by controlling managers is unrelated to firm value. In
contrast, Masulis et al. (2009) examine the contribution of one extra dollar of cash to firm
value in a sample of dual class companies in the U.S. where investor protection is more
stringent than in other countries. They find that the marginal value of cash is decreasing in
the divergence between voting and cash flow rights. This is consistent with the evidence
provided by Dittmar and Mahrt-Smith (2007) that an extra dollar of cash is valued less by
shareholders in poorly governed companies.
Most of the evidence in the prior literature is consistent with the agency theory of
"free cash flow" (Jensen, 1986). Shareholders believe that managers are more likely to
misuse corporate cash holdings and therefore, shareholders' place a lower value on
corporate cash holdings. In keeping with this argument, I investigate the link between the
dual class discount and excess cash holdings. Managers and controlling shareholders of dual
class firms are likely to misuse corporate cash holdings in pursuit of private benefits. Hence
investors, anticipating misused cash holdings by dual class managers, are more likely to
38
discount the value of dual class companies compared to single class companies with
concentrated control.
H3:
The higher the excess cash holding, the greater the valuation discount of
dual class firms relative to single class concentrated control firms after controlling
for size, dividend differences, conversion rights andfinancial leverage.
3.4.0
Capital expenditure can have a positive effect on firm value if shareholders believe
that the firm's capital expenditures are related to positive net present value projects.
McConnell and Muscarella (1985) provide evidence that mangers seek to maximize the
market value of the firm in making their corporate capital expenditure decisions. Also,
Masulis et al. (2009) find that change in capital expenditure has a positive impact on returns
for dual class firms. However, insiders' excess control rights reduce the contribution of
capital expenditure to shareholder value in dual class firms. Therefore, shareholders believe
that insiders' in dual class companies are using capital expenditure to pursue private
benefits. As a result, I formulate the following hypothesis:
H4:
Excess capital expenditure in dual class firms relative to single class firms
increases the dual class discount.
3.5.0
If investors believe that excess compensation is a channel through which managers

extract private benefits, then the rational investor hypothesis can add to our understanding of
the valuation discount of dual class companies by relating excess executive compensation to
the valuation discount. Investors, knowing that executives in dual class firms have the ability
and incentive to expropriate or extract resources without facing the cash flow consequences,
39
will discount dual class companies. Therefore, the larger the excess compensation, the
greater the dual class valuation discount. This leads to the following hypothesis:
H5:
The higher the excess executive compensation, the greater the valuation
discount of dual class firms relative to single class firms.
4.0
4.1.0
Methodology and Data

Methodology
The fixed effects regression specification will be utilized to test the various
hypotheses in this paper. The fixed effects approach is a common technique in panel data
setting. By incorporating industry fixed effects, I account for fixed differences in executive
compensation across different industries within the cross-section. Year fixed effects are
included to control for any year specific shocks to executive pay and concentrated control.
4.1.1
The following equation will be estimated in order to test hypothesis 1.
Ln(Compensation)l
= a ; + <|>r Mgmt.Vote
+ X r| (^Fz'rm Characteristicsk
k=\
+ Z S , kiGovernanceb k J ,,1 , + Y y , Position

t
k x[
~
'
'-
- ''-
^ 'i,k
k = 1
k,j
+s
I.J.I
(4),
''
where i is equal to different components as well as total compensation for each executives'
compensation package for firm/ at year t and management voting leverage (Mgmt. Vote) is
40
, _ .
defined
as
% of total votes controlled by management and directors ?9

.
^.
Firm
% of total equity held by management and directors

characteristics include: size, profitability, growth, risk and financial leverage. The
governance variables are as follows: board size, percentage of independent, busy and grey
directors, percentage of institutional ownership and percentage of family members on the
board of directors.30 Position is a dummy variable equal to 1 for different executive
positions and zero otherwise. For a complete definition of all the variables see Appendix A,
Table 1A.
Management voting leverage ratio is an indication of concentrated control relative to
equity ownership by senior level executives and directors. It is expected to be positively
related to executive compensation, hence, <j>. is expected to be positive. A high management
voting leverage ratio indicates that managers and directors control a significant proportion of
the voting shares relative to their equity stake. Therefore, executives in firms with high
management voting leverage are able to exert control over the board and extract more
resources or private benefits at the expense of minority shareholders.
Several firm characteristics based on previous studies are used as control variables
such as size, performance, investment opportunities, risk and financial leverage. It is
expected that larger firms (log of sales), and those with more complex operations (proxy by
size) will tend to pay their executives more in order to attract high quality executives (Core
et al., 1999 and Chalmers et al, 2006). Board of directors often reward executives for
29
Instead of using management voting leverage ratio as an indication of concentrated control, I re-estimate
equations (4) to (6) replacing management voting leverage with the "wedge" variable. Wedge is defined as the
difference in the percentage of votes and the percentage of cash flow rights held by management and directors.
The results (not reported) are similar to those presented in Tables 4-6 using wedge instead of management
voting leverage.
30
A busy director is defined as a director with more than four board memberships and grey directors are
defined as outside directors who are related to the company through a transactional relationship.
41
superior performance. Therefore, more profitable firms (ROA or stock return) will tend to
pay their executives higher compensation, especially in the form of bonuses and stock
options. Hence, profitability is expected to be positively related to compensation. Firms with
more growth options will tend to compensate executives more in order to attract executives
capable of turning those options into actual growth. Therefore, growth (geometric mean
growth in total assets over the previous five years) is expected to be positively related to
compensation. Firm risk can be a measure of the firm's operating risk as well as the firm's
information environment. Following (Smith and Watts, 1992; Core et al., 1999 and
Chalmers et al., 2006), I include a measure for firm risk. Firm risk is measured by the
standard deviation of annual returns (Core et al., 1999) and it represents total risk because of
the undiversifiable nature of executive contracts. Based on prior literature, firm risk can
positively as well as negatively influence the level of compensation. Therefore, the
expectation is ambiguous. In addition, several governance variables that explain executive
compensation are controlled for in equation (4).
Governance variables include, board size, independent directors, busy directors, grey
directors, institutional investors and the percentage of family directors. According to Core et
al. (1999), weak governance structure results in greater agency problems and higher
compensation. Therefore, a greater percentage of busy and grey directors signifies a weaker
governance structure. Hence, busy and grey directors are expected to be positively related to
executive compensation. In theory, independent directors are expected to monitor managers
and curb excess compensation and therefore increase board effectiveness. However,
empirical evidence shows a positive relationship between the percentage of independent
directors and CEO compensation (Lambet et al, 1993). One possible explanation is that
42
independent directors are not truly independent directors such as grey directors.31 Therefore,
after controlling for the percentage of grey directors, the percentage of independent directors
is expected to be negatively related to executive compensation. Furthermore, institutional
investors can play an important role in executive compensation by using their voting power
to influence board composition and thereby limiting rent extraction. Empirical evidence
shows that institutional ownership is negatively related to the level of executive
compensation (Hartzell and Starks, 2003). As a result, I account for the effects of
institutional ownership on executive compensation. In addition to the governance variables
used in the literature, since several of the firms are family controlled, I account for the
percentage of family directors on the board. Family directors are expected to monitor nonfamily executives and therefore, reduce compensation of non-family executives. Family
directors, on the other hand, may be influenced by family executives and are likely to
reward these executives with higher compensation.
4.1.2 Family Executives and Compensation
Testing the effects of family executives on compensation in both dual and single
class firms (H2a), the following equation is estimated using a panel regression with year and
industry fixed effects.
31
For example, Disney claims that 13 of the 16 directors are independent members of the board. However, of
these "outside" directors, Reveta Bowers, is the headmaster of the school Eisner's children attended. Also, Leo
O'Donovan, president of Georgetown University, which one of the Eisner children attended, received
donations from Eisner. Other "outside" directors such as Irwin Russell, Eisner's personal attorney and Robert
Stern, architect for several of Disney's projects are clearly not true independent directors since they have ties to
the CEO. They may be considered as "grey directors" (Kim and Nofsinger, 2004).
43
Ln(Compensation)
= a, + 9, kFamilyPositionk
+(j), Mgmt.Vote., {
i,k=i
5
+ Z 1 !, kFirm Characteristicsk
k=\
6
6
n
'
+ Z ( kGovernancek
k=\
+ Ty,,kPositionkj+zlJt
t{
'
(5),
k=\
where / is equal to different components for each executives' compensation package for firm
j at year t. Family positions are indicator variables equal to 1 if, for example, the CEO is a
family member and 0 otherwise. It is expected that family executives will extract more
resources from the firms they control and therefore, executive compensation will be higher
for family members relative to non-family executives.
I will estimate equation (6) below to examine the difference in executive
compensation for family members in dual class firms as compared to family executives in
single class firms (H2b) in a sub-sample of family executives only. I expect dual class
family executives to extract resources in the form of higher executive compensation relative
to family executives in single class firms. To test this prediction, I interact dual class family
executive positions and management voting leverage. The coefficients of interaction terms
are expected to be positive and significant for various executive positions since greater
voting power relative to cash flow stake in family firms can exacerbate the agency conflict
between controlling shareholders and minority shareholders. I also include management
voting leverage as a separate control variable to make sure that the interaction term does not
merely pick up the effect of management voting leverage itself.
44
Ln(Compensation), j t = a, + Z 9 a f DFam.Execu
x Mgmt.VoteyjM)
kj
A=I
+ ^lMgmt.VoteJil_l + Y.r\lkFirm Characteristicsk t_

6
+ H&i,kGovernancek,il_l
fc=i
+ ^ , position
+e
*=i
,y
(6),
'
where DFam.Execu. is an indicator variable equal to 1 for the k family executive position
in the/th dual class firm and zero otherwise.
4.1.3 Dual Class Discount and Corporate Cash Holdings
In order to test the effects of excess corporate cash holdings on the valuation
discount of dual class companies, I estimate the following equation:
Dual Class Discount]t+x
- a + (3j (Excess Cash x Mgmt.Vote)
+ (3 3 Excess Cash-1 + p 4 Financial Leverage
+ $5ConversionRight
+ (36 Size
+ Q^Mgmt. Vote
+ p7D/v.diff
+,,
(7)
There are two measures of dual class discount. In the first measure, dual class
discount is the difference between Tobin's Q ratio for a dual class firm and Tobin's Q ratio
for a matching single class concentrated control firm.32 The second measure of dual class
In order to estimate Tobin's Q ratio, I follow the approach used by Lins (2003) where
' Market Value of Equity + Total Assets - Book Value of Equity *
Gompers et al. (2010 used a similar
Q=
Total Asse ts
definition except they subtract deferred taxes from the numerator. In order to calculate the market value equity
of the dual class companies, I follow Smith et al. (2009). For dual class companies with both classes of shares
trading, the market value equity is the sum of the market value of the restricted voting class plus the superior
voting class. For companies with only the restricted voting class shares trading, I add a premium to the price of
45
discount is the difference in the Q ratio of dual class firms and their SIC industry average Q
ratio. Excess cash is defined as
A cash,
Total assets, , ; , , ,
' ~ ' / Dual class
A cash.
v Total assets,
V
, ,,
, ,
where
~"1 y Single class
the change in corporate cash holdings is cash and marketable securities at the end of year t
minus cash and marketable securities at the end of year t-1. To test whether the divergence
between voting and cash flow rights along with excess cash holdings affect dual class
discount, I interact management voting leverage and excess cash and include the interaction
term as an explanatory variable. The coefficient of the interaction term is expected to be
negative and significant, since excess control rights relative to equity stake can lead to
inefficient use of corporate cash holdings. Management voting leverage (Mgmt. vote) is
expected to be negatively related to the dual class discount. This represents agency problems
associated with dual class ownership structure. Therefore, the greater the divergence
between voting and cash flow rights, the larger the discount. Following Zingales (1995),
conversion right is an indicator variable equal to 1 if superior voting shares can be converted
into restricted voting shares and 0 otherwise, market value of equity (size) is used as a proxy
for the probability of acquisition and dividend difference (Div. diff) is an indicator variable
equal to 1 if the dividend paid or payable to restricted voting shares is greater than that of the
superior voting shares and zero otherwise.
the restricted voting shares in order to estimate the market value of equity for the superior voting shares. I use
6.5% as the voting premium which was established by Zingales (1995). There has been some debate on the
measurement error of Tobin's Q ratio as a proxy for "true" Q (See Whited, 2001, Erickson and Whited 2002
and Erickson and Whited 2006). Erickson and Whited (2006) argue that the measurement error of Q becomes
an issue depending on the use of Q in a regression. If a proxy for Q is used as a regressor to control for the
incentive to invest, then the measurement error is a serious problem. However, if proxy appears as the
dependent variable in a regression, then the measurement error does not bias any slope coefficient. It does,
however make the R2 smaller than in the absence of such error.
33
In 13.5% of the dual class firms, holders of restricted voting (RV) shares are paid more or will receive more
dividends in the future relative to holders of superior voting (SV) shares.
46
4.1.4

In order to test the effects of capital expenditure on the valuation discount of dual
class companies, I estimate the following equation:
Dual Class Discount
t+]
= a +fix(Capex x Mgmt.Vote)
+ p^Financial Leverage
+ 2Mgmt. Vote
+ fi5Conversion Right
+ p 6 Size
+ V7Div.diffJt + eJt
+ fi3Capex
(8),
where capital expenditure (Capex) is defined as:

ACapex, t
Market value of equity t_x
ACapex ( ,
Dual class
Market value of equity t_{ j Single

, class
and the change in Capex is capital expenditure at the end of year t minus capital expenditure
at the end of year t-l. The interaction term between capital expenditure (Capex) and
management voting leverage (Mgmt.Vote) is expected to be negative. This implies that
excess capital expenditure increases the dual class discount.
4.1.5 Dual Class Discount and Excess Compensation
Finally, I examine the effects of excess compensation on valuation discount of dual

class firms (H3b). Following Zingales (1995) and Masulis et al. (2009), I estimate equation
(9) using firm characteristics and governance variables that have been proven to explain
executive compensation and extract the residuals as a measure of excess compensation. This
excess compensation is used to explain the valuation discount in equation (10).
47
Ln(Compensation)
- a, + X r\lkFirm Characteristicsk
,j,
k = ]
,,
,j,
+ I 5lkGovernancek
,_, + I ylkPositionk
ft=l
=1
+e
(9),
'
where z is equal to different components of the executive compensation package for firm 7 at
year /.
Dwa/ CYass Discount
t+x
- a + Pjisxcess Comp.
+<^1{Mgmt.Vote x Excess Cash)
+ $3Mgmt. Votej t + fi4Excess Cash

+ P^Conversion RightJt + (37 Size
t
t
+ fi5Financial Leverage
+ P8D/v. diff
+ Y t Position k
+s ,
(10),
where exce^ cowp. is residual from equation (9) for total compensation and the various
components of executives' pay package such as salary, bonuses, stock options and other
compensation.
After accounting for excess corporate cash holdings, management voting leverage
and other control variables, I expect a higher excess compensation (Excess Comp.) to be
associated with a larger valuation discount of dual class companies. The discount reflects
investors' belief that executives of dual class companies are extracting private benefits of
control through excess compensation. Thus excess compensation is expected to be
negatively related to dual class discount after controlling for factors that partially explain
dual class discount in the prior literature. Management voting leverage and the interaction
term between management voting leverage and excess cash are expected to be negatively
related to dual class discount.
48
4.2
Data
4.2.1 Sample Construction

The dataset used in this research is constructed from a variety of sources.
To construct a sample of U.S. dual class firms, I retrieved a list of firms with dual class
share structure from Corporate Library for 2005-2007. Corporate Library identifies all firms
with dual class common shares as a takeover defense mechanism during the period. I extend
this list of dual class firms to 2001 using a list of dual class IPOs and a list of dual class
firms used in Gompers et al. (2010).34 The list of 1,910 dual class firms over the period of
2001-2007 is merged with Execucomp database to determine whether compensation data is
available for these firms. Execucomp database contains executive compensation data for the
top executives representing the S&P 1500 group of companies. For each dual class company
with compensation data, I retrieve proxy statements from the Securities and Exchange
Commission (SEC) website and check the proxy statement for each firm in the sample to
ensure that they are in fact, dual class companies. Next, using proxy statements, I collect
voting rights per share and the number of superior voting and restricted voting shares owned
by the largest shareholders and management and directors as a group.
For each firm, I collect accounting data from Compustat. I retrieve annual firm-level
information such as total assets, sales, long-term debt, common equity and operating
income. In addition, I collect several governance variables and equity ownership data from
Corporate Library and Execucomp. These include the number of directors, outside related
directors and unrelated directors. I use proxy statements, firm websites and internet search
34
A list of dual class IPO is available on Jay Ritter's IPO website. Andrew Metrick generously provided the list
of dual class companies used in their study.
49
engines such as Lexus Nexus and Google to identify family executives and family directors.
I then calculate the percentage of family members who are directors of the board. In order
to complete the set of control variables, I collect stock return data from CRSP. I obtain
monthly returns to estimate beta and annual returns to compute standard deviation. Finally, I
collect compensation information for all executives available on Execucomp. Total
Compensation (TDC1) is defined as:35
Total Compensation = Salary + Bonuses + Other Compensation + Stock Options
(11),
where other compensation includes the value of restricted stock grants, long term incentive
payouts, contributions to defined pension plans, life insurance premiums, consulting fees
and awards under charitable award programs.
4.2.2 Propensity Score Matching
The list of dual class companies is matched with a list of single class concentrated
control firms using propensity score matching. Propensity score matching methods were
developed in Rosenbaum and Rubin (1983), Heckman and Robb (1986) and Heckman et al.
(1998). Propensity score matching has become a popular matching technique applied to
studies of the financial markets (see Hillion and Vermaelen, 2004 and Villalonga, 2004)
which allows for matching using a larger number of characteristics and hence, reducing or
eliminating potential bias. The greater the overlap in all characteristics of the treated and
35
The Execucomp database value of executive stock options using Black-Scholes call option valuation model
(1973). Alternatively, Johnson, Ryan and Tian (2009) adopt a modified version of the Black-Scholes model
to compute the option values because executives typically exercise their options before maturity. Therefore,
they reduce the contractual option maturity by 30%.
50
control groups, the more comparable the groups are and the smaller the bias (Heckman et
al., 1997 and Heckman et al., 1998).
Using a propensity score algorithm, I estimate a probit model of the determinants of
dual class structure and compute a propensity score for each firm based on several firm and
governance characteristics. The propensity score is then used to match each dual class firm
with a similar single class company. The following firm and governance characteristics are
used in the matching exercise: equity ownership of the largest shareholder, sales, industry,
return on asset, annual stock return, beta, standard deviation of annual returns, market-tobook, debt-to-asset, sales growth, board size, proportion of independent directors, busy
directors, grey directors, institutional ownership, company age, R&D-to-sales and capex-tototal asset. This matching exercise results in a final sample of 792 dual class firm-years over
the period of 2001-2007. This represents an average of 113 dual class firms per year. The
final matched sample is made up of 1,584 firm-year observations and 7,920 executive-firmyear observations.
5.0
Results
5.1.0
Table 1 contains descriptive statistics of executive compensation. Panel A shows
that on average, dual class executives received $2.6 million of total compensation compared
to $2.3 million of total compensation for a similar executive in a single class concentrated
control firm. The test of mean (median) difference between dual and single class executives
is positive and statistically significant at the 5% level. In terms of salary and bonuses, dual
51
class executives are paid more than single class concentrated control executives. Univariate
tests for the difference in mean (median) show that salary and bonuses are significantly
higher for dual class executives (Panel A, Table 1). The CEOs, COOs and VPs of dual class
firms receive significantly higher salary, bonuses and total compensation relative to the
matching group of executives (Panels B, D and E). However, there is no difference in total
compensation for CFOs (Panel C, Table 1). For example, Panel B Table 1, shows that dual
class CEOs and presidents receive an average of $5.2 million of total compensation
compared to matching single class concentrated control CEOs with $4.7 million in total
compensation. The univariate test for difference in mean (median) for CEO total
compensation is positive and significant at the 5% level.
Compared to single class firms, COOs and VPs in dual class firms received $287.35
thousand and $122.98 thousand more in total compensation, respectively (Panel D and E,
Table 1). The test for difference in mean (median) is positive and significant for salary,
bonuses and total compensation for both COOs and VPs. Figure 1 shows the median total
compensation for the top three executives over the period 2001-2007. The graph indicates
that the top three highest paid executives in dual class firms receive more compensation
compared to similar top three executives in single class concentrated control firms.
In terms of family executives, family CEOs and Chairmen in dual class firms are
paid more total compensation ($2.9 million) compared to family CEOs and Chairmen in
matching single class firms (Panel F, Table 1). A family member who is the CEO but not
the Chairman of the board also earns significantly more salary, bonuses and total
compensation compared to a similar family CEO in a single class company (Panel G, Table
1). Finally, family CEOs in general received higher compensation than non-family CEOs
52
(Panel H, Table 1). In fact, family CEOs received 24.6% more in total compensation to
similar non-family CEOs.
In terms of the industry distribution of dual class firms, Table 2 provides descriptive
statistics based on SIC industry codes. The sample of dual class firms in this essay is
generally distributed across all industries. However, business services, communication,
retail services, food and kinder products, printing and publishing and electronic equipment
are the six industries with the greatest number of dual class firms over the period of 20012007.
Table 3 shows descriptive statistics of firm characteristics, governance and executive
characteristics. Panel A summarizes governance characteristics for the sample. In most dual
class firms the family, founder or management as a group, controls a significant proportion
of the voting rights. The mean (median) cash flow ownership by the largest shareholders is
22.5% (18.2%>) compared to 23.6% (19.2%) for similar single class companies. In terms of
the voting rights, the largest shareholder controls on average, 57.8% of the voting rights. As
a group, management and directors in dual class companies, on average control 58.3% of
the votes and only 24.9% of the cash flow rights. This represents a significant divergence
between their voting and cash flow rights. As a result of this divergence, institutional
investors are less likely to hold a significant proportion of dual class restricted voting shares.
According to Li et al. (2008), voting rights are an important determinant of institutional
investment decisions. This argument is consistent with the finding that institutional investors
hold significantly fewer shares in dual class firms (16.9%) compared to single class firms
(21.5%).
53
Univariate tests show that dual class firms have, on average, significantly less
independent directors (59.3%) but significantly more busy directors (16.6%) compared to a
group of matching single class concentrated control firms (68.8% and 13.0%, respectively).
Also, dual class firms have more grey directors. In fact, 10.2% of the board is made up of
outside related directors (grey directors) compared to 9.9% in single class concentrated
control firms. In addition, a large percentage of dual class firms in the sample are considered
as family firms. In fact, 83.2% of dual class firms are family firms compared to only 31.4%
of the single class concentrated control firms. This implies that family influence is greater in
dual class companies. The univariate test shows that a greater percentage of family members
in dual class companies are directors (15.8%) relative to single class firms (5.6%) with the
difference in mean and median being positive and statistically significant.
Table 3, Panel B reports the descriptive statistics for firm characteristics. In terms of
the valuation discount, measured by Tobin's Q ratio, dual class companies have a lower
value compared to a group of matching single class concentrated control companies. This is
consistent with the findings of prior studies such as King and Santor (2008), Gompers et al.
(2010), and Smith et al. (2009). The test for mean (median) difference is negative and
statistically significant at the 1% level. Using median Q ratio, there is a 9.3% discount of
dual class firms compared to single class companies. Figure 2 shows the average discount of
restricted voting shares relative to superior voting for dual class companies with both classes
of shares trading over the period of 1998-2007. The mean (median) discount over this period
is 5.10% (5.98%). In terms of firm size, the univariate test indicates that there is no
difference in the mean (median) size of dual class firms and their matching single class
concentrated control counterpart (Table 3, Panel B). Measured by sales, the mean size of
54
dual class firms is $4.9 billion compared to $4.6 billion for single class concentrated control
firms. Similarly, there is no difference between single class and dual class firms in terms of
risk (standard deviation), growth, performance (ROA) and financial leverage (D/A). This
implies that, based on firm characteristics, the propensity score matching produces a
relatively good group of control firms.
5.2.0
Regression Analysis
5.2.1
The relationship between executive compensation and ownership structure after

accounting for firm characteristics and governance determinants of executive compensation
is examined using fixed effects regression. The dependent variable includes various
components of the executives' compensation package as well as total compensation
(Execucomp item TDC1). The results are presented in Table 4. Under the identification
assumption that variation in concentrated control is exogenous, the regression results are
consistent with the prediction of hypothesis 1. Executives in dual class companies are paid
significantly higher salaries, bonuses, other compensation, options and total compensation.
The results provide evidence that dual class executives are using their voting power to
extract resources from the firm at the expense of minority shareholders. This finding is
consistent with the private benefits of control and the agency costs arguments. Management
voting leverage is positive and significant for each component of compensation package as
well as total compensation. This implies that the greater the voting power relative to cash
flow ownership, the greater the excess compensation paid to dual class executives. The
55
positive relationship between management voting leverage and executive compensation for
all senior level executives corroborates the findings of Masulis et al. (2009) for CEOs only.
Several firm characteristics are positive and significant which are consistent with
previous research findings. As expected, performance is positively associated with total
compensation and all components of executive compensation, except for salary.36 Financial
leverage is negative and significant for total compensation and stock options. The
coefficient on size is positive and significant indicating that larger firms tend to pay their
executives higher compensation. The evidence is consistent with the argument that
compensation related to size, complexity and investment opportunities reflects a demand for
higher quality executives (Core et al., 1999; and Chalmers et al., 2006). Firm risk, measured
by standard deviation of returns, is positively associated with bonuses, stock options and
total compensation. This implies that executives in more risky firms demand higher
compensation. The evidence is consistent with Chalmers et al. (2006). Growth is positive
and significant for total compensation which is primarily driven by bonuses and stock
options. This implies that firms with growth options tend to compensate their executives
with more incentive based compensation compared to fixed salaries.
Table 4 also presents results for several governance variables that have been proven
to explain executive compensation. Larger boards (number of directors) lead to higher
compensation. This evidence suggests that larger boards are less effective in reducing excess
compensation. This is consistent with prior studies such as Core et al. (1999). In theory, a
greater percentage of independent directors should result in lower executive compensation
because independent directors are expected to be better monitors. However, the coefficient
36
1 replaced ROA as a measure of performance with annual stock returns and the results are qualitatively
similar.
56
on independent directors is positive indicating that a greater percentage of independent

directors leads to higher executive compensation. The greater the percentage of busy or grey
directors, the higher the executive compensation. The evidence suggests that boards with a
high proportion of busy and grey directors are less effective in monitoring managers and
acting in the interest of shareholders. Finally, the results presented in Table 4 indicate that
the percentage of family directors on the board leads to lower executive compensation as
predicted. However, the percentage of institutional ownership does not influence executive
compensation.
5.2.2
Table 5 presents the results for various family executives after controlling for
management voting leverage, firm characteristics and several governance variables that
explain executive compensation. Being family members in executive positions in both dual
class and single class concentrated control firms has a positive impact on executive
compensation. The coefficient on family CEOs and chairmen is positive and significant
indicating that family members performing the dual role of CEO and chairman received
significantly higher compensation. Family CEOs who are not the Chairmen also received
higher total compensation. This evidence is consistent with the findings of Cheung et al.
(2005), Basu et al. (2007) and Cohen and Lauterbach (2008). Family members in other
executive positions earn significantly higher compensation compared to non-family
executives in similar positions. Family Chairmen, CFOs and COOs received significantly
higher total compensation compared to non-family executives in similar positions.
57
Table 6 presents regression results for a sub-sample of family executives only in

both dual and single class concentrated control firms. Examining compensation of dual class
family executives compared to single class family executives, total compensation as well as
various components of executive compensation are regressed on several interaction
variables along with firm characteristics and governance determinants of executive
compensation. As predicted, dual class family executives received significantly higher
compensation compared to family executives in similar positions in single class
concentrated control firms, except for dual class family VPs. This provides additional
support for the extraction of private benefits of control and the agency cost arguments. In
terms of the various components of the compensation package, dual class family executives
seem to be extracting private benefits in the form of higher salaries and other compensation.
5.2.3
Dual class firms, on average, retain more cash compared to single class companies
with concentrated control. The mean (median) cash holdings for dual class companies is
$642.3 million ($166.0 million) compared to $431.0 million ($132.8 million) for single class
concentrated control firms (Table 3, Panel B). The univariate test for the difference in mean
(median) is positive and significant at the 1% level. The change in cash, year t-1 to year t, is
larger in dual class companies compared to single class concentrated control companies.
This implies that managers in dual class companies are retaining excess cash which can be
58
used to provide private benefits. In addition, the mean (median) excess cash holdings, scaled
by total assets, is 2.4% (0.35%) and the mean is statistically significant at the 5% level.37
Table 7 provides evidence which links the discount of dual class companies to the
managerial extraction of private benefits of control. I estimate equation (7) and present the
results in Table 7. In models (1) and (2) the dependent variable, dual discount, is the
difference between Q ratio for dual class firms and matching single class concentrated
control firms. In columns (3) and (4), the dual class discount is computed as the difference
between dual class firms and their respective SIC industry average Q ratio. Using both
measures of discount, I find that the interaction term between excess cash holdings and
management voting leverage is negative and significant after controlling for excess cash,
management voting leverage and other control variables. The result is consistent with the
argument that when managers control a significant proportion of votes relative to equity
ownership, excess corporate cash holdings is more likely to be diverted to provide private
benefits and thus investors discount the value of dual class companies. The findings imply
that investors view excess cash in dual class companies as a potential avenue for the
extraction of private benefits. Therefore, excess cash exacerbates the agency conflict
between controlling shareholders and managers and minority shareholders. For the control
variables, the signs are generally consistent with those reported in the literature. For
example, the difference in dividends paid or payable to restricted shareholders is positive
and significant when dual class discount is measured using average industry Q ratio.
37
Excess cash is defined as:
A casht t
A cash,,
Total assets,_ \J'Dual
class
Total assets,_ \) 'Single class
where the change in
corporate cash holdings is cash and marketable securities at the end of year t minus cash and marketable
securities at the end of year t-1.
59
5.2.4
Univariate tests indicate that dual class firms undertake more capital investments
compared to single class closely-held companies. The mean (median) capital expenditure for
dual class firms is $256.5 million ($65.5 million) relative to $138.3 million ($46.5 million)
for single class concentrated control companies (Table 3, Panel B). Given the size of the
capital expenditure for dual class companies, one would suspect that capital expenditure is a
channel through which managers and controlling shareholders can extract private benefits.
However, the multivariate test does not support this view. Table 8 reports the results of
estimating the impact of capital expenditure on both measures of dual class discount. The
results indicate that after controlling for management voting leverage and other factors,
capital expenditure does not affect dual class discounts. This implies that investors do not
view capital expenditure as an avenue through which controlling shareholders and managers
extract private benefits. One possible explanation is that it is difficult to disentangle the
effects of real long-term investments and potential extraction of private benefits through the
use of capital expenditure on firm value. Alternatively, capital expenditure may not be an
avenue to extract private benefits due to its importance as capital investment for corporate
growth.
5.2.5 Dual Class Discount and Excess Compensation
To test the rational investor hypothesis explaining dual class discount firms using
excess compensation, I estimate equation (9) and obtain the excess compensation. I then
estimate equation (10) and the results are presented in Table 9. The results indicate that after
60
controlling for excess corporate cash holdings, the higher the excess total compensation, the
larger the valuation discount of dual class companies (Panel A, Table 9, Column V). Also,
various components of the compensation package such as excess salary, bonuses and other
compensation result in a greater discount of dual class companies. We can argue that
investors believe that executive compensation is a channel through which controlling
shareholders' and managers' can extract private benefits. The interaction term between
excess cash and management voting leverage is negative and significant. This confirms the
rational investor hypothesis that investors expect executives in dual class firms to use their
control to extract resources for the firms they control. Therefore, investors, in turn, discount
the value of dual class firms. The control variables corroborate prior findings. Size is
positive and significant which is consistent with prior studies, such as that of Zingales
(1995).
In column VI (Panel A, Table 9), I examine the relationship between family
executives' excess compensation and dual class discount. Excess compensation paid to
family CEOs (CEO only and CEO and Chairman) in dual class companies leads to an
increase in the valuation discount of dual class firms. This is consistent with the evidence
provided by King and Santor (2008).
In Panel B, Table 9, the dual discount is computed using the industry average Q
ratio. The results for total compensation, salary and other compensation are similar to the
results reported in Panel A, Table 9. As for dual class family excess total compensation,
excess compensation to family CEOs results in a larger dual class discount.
61
5.3
Endogeneity
5.3.1 Simultaneity and Two-Stage Least Square Estimation
The identification assumption crucial to the causal interpretation of the findings

above is that the variation in management voting leverage is unrelated to unobserved factors
which influence executive compensation. This assumption is motivated by the observation
that ownership structure is established at the IPO stage or earlier in the firm's life cycle.
However, there are occasions where dual class ownership and control structure is
established long after IPO. Therefore, it is possible that some unobserved factors correlate
with changes in control structures that affect executive compensation. Therefore, the
regression estimates in Tables 4-6 are likely to be biased and inconsistent due to
simultaneity. In order to address the potential simultaneous process determining executive
compensation and ownership structure or omitted variable bias, a two stage least square
estimation technique is utilized. Executive ownership can influence the level of executive
compensation, while executive compensation, especially stock options, can influence their
ownership level. As a result, the following specification will be jointly estimated with
equation (13):
Mgmt.Votej
=a + \\i Ln( Compensation ) . t
Jt
+v|/ Beta + \\i Lev.t +M? Qt
+ \\f Name + \\i Media+\\i Sizet+Zjt
(12),
where management voting leverage, compensation, financial leverage (Lev.) and size are
defined in Table 1A, Appendix A. Beta is estimated using the previous 5 years of monthly
62
returns. Q is the Tobin's Q ratio. Name is a binary variable equal to 1 if the firm name is
the same as an individual's name. Media is a dummy variable equal to 1 if the firm is in the
media industry.39
Ln( Compensation )t
= a, +(|)( Pr ed.Mgmt.Vote
+ YJr\,k^rm
k=\
k t_x
+ ZS
k=\
Characteristics
'
Governancek
/M
+ Y.J, k-Psônk
'*
k=\
+s
/ n
0-^),
'
where Pred.Mgmt.Vote is the predicted management voting leverage from equation (12)
above.
The results of the second stage regression are presented in Table 10. Column I
shows the estimates for equation (13) for total compensation. Predicted management voting
leverage is positive and significant indicating that executives are using their voting power to
extract higher compensation. Column II presents results for a modified version of equation
(13) which includes family executive indicator variables. The results are similar to those
presented in Table 5. Finally, column III reports the second stage results for a sub-sample of
family executives only. Dual class family executives received significantly higher total
compensation, except for family VPs in dual class firms.
I used leverage, size, beta and Tobin's Q as determinants of management voting leverage following Lins
(2003) and Smith et al., (2009).
39
A firm is defined as a media firm if it belongs to SIC codes 2710-11, 2720-21, 2730-31, 4830, 4832-33,
4840-41, 7810, 7812 and 7820. Gompers et al., (2010) show that name and media variables are important
predictors of dual class ownership structure. Therefore, I include a name and media variable in estimating
equation (12).
63
5.3.2 Heckman Sample Selection
It is possible that dual class firms self-select this type of ownership structure and
therefore, we have to account for the sample selection bias. Using the Heckman (1979) twostage estimation procedure, I account for the potential bias due to self-selection. In the first
stage, I estimate a probit model with the dependent variable equal to 1 if the firm is a dual
class firm and 0 otherwise. Several of the independent variables used in determining dual
class status are based on prior studies such as Gompers et al. (2010). Using the estimated
parameters from the probit model, I compute the inverse Mills ratio which is included as an
additional explanatory variable in the pool regression. The results of the Heckman secondstage estimation are presented in Table 11. The results are similar to those presented in
Tables 4-6. The inverse Mills ratio is not significant indicating that sample self-selection
bias does not affect the results.
5.3.3 Robustness Check
It is often argued in the literature that equity ownership by executives tends to have
an impact on their compensation. Executives having a significant ownership stake in the
firm they manage are less likely to extract excess compensation. As a result, I include
percentage of equity ownership as an additional explanatory variable. The percentage of
equity ownership is negative and significant at the 5% level (Table 12). This corroborates
evidence from prior studies (Core et al., 1999 and Chalmers et al., 2006). However,
management voting leverage is still positive and significant at the 1% level after controlling
for the percentage of equity ownership by each executive, governance variables and firm
characteristics. In addition, the literature on CEO compensation often argues that a CEO's
64
age is a positive determinant of CEO compensation. Therefore, I include executives' age as

a control variable and re-estimate equations 4-6 above. The coefficient on the age variable is
not statistically significant (Table 12).
I also re-examined the valuation discount of dual class firms using only those with
both classes of shares trading on the exchange. There are 18 dual class firms (126 firm-year
observations) that have both classes of shares trading in the sample. The regression results
(not reported) are qualitatively similar to those presented in Tables 7 and 8.
6.0
Chapter Summary and Conclusions
A long standing concern regarding concentrated ownership and control is that it

creates agency problems and also leads to the extraction of private benefits and agency
costs, especially in dual class firms where there is the separation of voting and cash flow
rights. In addition, the channels through which controlling shareholders and managers can
extract private benefits have not been well documented in the literature. Of particular
interest is the relationship between these channels and the observed valuation discount of
dual class companies compared to single class concentrated control companies. Using Q
ratio as a measure of value, I show that there is a significant difference in the valuation
between dual and single class companies. Using mean (median) Q ratio, there is an 8.33 %
(9.3%) discount of dual class firms relative to matching single class companies.
Using a sample of 1584 dual and single class concentrated control companies over
the period 2001-2007 from the S&P 1500 group of companies and utilizing panel regression
techniques with industry and year fixed effects, this essay provides empirical evidence
showing a direct link between dual class discount and excess compensation and excess cash.
65
However, capital expenditure does not explain the dual class discount. Therefore, the
channels of wealth expropriation are excess compensation in general, excess compensation
for family members, and excess corporate cash holdings. The degree of excess
compensation is greatest in dual class companies with executives who are family members.
This essay tests the effects of two different ownership structures with different forms
of concentrated control using a sample of U.S firms. In the first ownership structure,
concentrated control is obtained via majority voting rights (dual class ownership structure).
In the second sample, control occurs through significant equity ownership (single class
concentrated control firms). The tests indicate that executives in dual class companies earn
significantly higher compensation compared to executives in similar single class companies.
The results also show that family CEOs in dual class companies receive higher
compensation compared to family CEOs in single class concentrated control companies.
Family CEOs, in general, are paid more compensation relative to non-family CEOs. I
conducted a series of robustness checks and corrections of potential endogeneity due to
simultaneity and self-selection biases. Using two-stage least square and Heckman sample
selection specification, the results are robust to potential endogeneity biases. Also, I re-run
the analysis with "wedge" instead of management voting leverage. In addition, I include
executives' age and equity ownership and the results remain qualitatively the same.
Using the valuation discount of dual class firms, I provide evidence of private
benefits of control by relating excess compensation to the dual class discount. In particular,
excess total compensation paid to dual class family CEOs increases the dual class discount.
In addition, excess cash holdings lead to a larger valuation discount of dual class companies
relative to single class concentrated control firms. This is consistent with the belief that
66
managers can misuse corporate cash holdings in the pursuit of private benefits and as a
result, investors discount the value of dual class companies. Additional tests indicated that
capital expenditure is not a channel for the expropriation of shareholders' wealth in dual
class companies. Consistent with Masulis et al. (2009), the two channels of extraction of
private benefits from dual class companies are executive compensation and cash holdings.
In summary, the evidence supports the agency cost argument that dual class
controlling shareholders and managers use their control to extract private benefits from the
firm at the expense of minority shareholders. Thus, investors are aware of this perquisite
consumption and discount the value of these companies, accordingly.
67
7.0
References
Anderson, R., and Reeb, D. M., (2003), Founding-family ownership and firm performance:
Evidence from the S&P 500, Journal of Finance, 58, 1301-1329
Basu, S., Hwang, L., Mitsudome, T., and Weintrop, J., (2007), Corporate governance, top
executive compensation and firms performance in Japan, Pacific-Basin Finance
Journal, 15, 56-79
Black F., and Scholes, M , (1973), The pricing of options and corporate liabilities, Journal
of Political Economy, 81, 637-654
Bebchuk, L.A., and Fried L. ML, (2003), Executive compensation as an agency problem,
Journal of Economics Perspectives, 17, 71-92
Bebchuk, L. A., Fried, L. M , and Walker, D. I., (2002), Managerial power and rent
extraction in design of executive compensation, The University of Chicago Law
Review, 69, 751-846
Chalmers, K., Koh, P., and Stapledon, G., (2006), The determinants of CEO compensation:
Rent extraction or labour demand?, The British Accounting Review, 38, 259-275
Cheng, S., and Firth, M., (2006), Family ownership, corporate governance and top executive
compensation, Managerial and Decision Economics, 27, 549-561
Cheung, Y., Stouraitis, A., and Wong, A.W. S., (2005), Ownership concentration and
executive compensation in closely held firms: Evidence from Hong Kong, Journal
of Empirical Finance, 12, 511-532
Chung, K. H., and Pruitt, S. W., (1996), Executive ownership, corporate value and executive
compensation: A unifying framework, Journal of Banking and Finance, 20,11351159
Claessens, S., Djankov, S., and Lang, L. H. P., (2000), The separation of ownership and
control in East Asian corporations, Journal of Financial Economics, 58, 81-112
Cohen, S., and Lauterbach, B., (2008), Difference in pay between owner and non-owner
CEOs: Evidence from Israel, Journal of Multinational Financial Management, 18,
4-15
Core, J. E., and Larcker, D.F., (2002), Performance consequences of mandatory increases in
executive stock ownership, Journal of Finance Economics, 64, 317-340.
Core, J. E., Holthausen, R.W., and Larcker, D.F., (1999), Corporate governance, chief
executive officer compensation and firm performance, Journal of Financial
68
Core, J. E., Guay, W.R., and Larcker D.F., (2003), Executive equity compensation and
incentives: A survey, Economic Policy Review, April, 27-50
DeAngelo, FL, and DeAngelo, L., (1985), Managerial ownership of voting rights: A study of
public corporations with dual classes of common stock, Journal of Financial
Demsetz, H., and Lehn, K., (1985), The structure of corporate ownership: Causes and
consequences, Journal of Political Economy, 93, 1155-1177
Denis, D. K, (2001), Twenty-five years of corporate governance research and counting,
Review of Financial Economics, 10, 191-212
Dittmar, A., and Mahrt-Smith, J., (2007), Corporate governance and the value of cash
holdings, Journal of Financial Economics, 83, 599-634
Doidge, C , (2004), U.S. cross-listing and private benefits of control: Evidence from dualclass firms, Journal of Financial Economics, 72,519-553
Dyl, E. A., (1988), Corporate control and management compensation: Evidence on the
agency problem, Managerial and Decision Economics, 9, 21-25
Erickson, T., and Whited, T.M., (2002), Measurement error and the relationship between
investment and Q, Journal of Political Economy, 108, 1027-1057
Erickson, T., and Whited, T. M., (2006), On the accuracy of different measures of Q,
Financial Management, 35, 5-33
Fama, E., and Jensen, M., (1983), Separation of ownership and control, Journal of Law and
Goldberg, L. G., and Idson, T. L., (1995), Executive compensation and agency effects, The
Financial Review.,30, 313-335
Gomez-Mejia, L., Tosi, H., and Hinkin, T., (1987), Managerial control, performance and
executive compensation, Academy of Management Journal, 30, 51-70
Gomez-Mejia, L. R., Larraza-Kintana, M., and Makri, M., (2003), The determinations of
executive compensation in family-controlled public corporations, Academy of
Management Journal, 46, 226-237
Gompers, P.A., Ishii, J. and Metrick, A., (2010), Extreme governance: An analysis of dualclass firms in the United States, Review of Financial Studies, 23, 1051-1088
Grossman, S., and Hart, O., (1988), One share-one vote and the market for corporate control,
69
Haid, A., and Yurtoglu, B., (2006), Ownership structure and executive compensation in
Germany, Available at SSRN: http://ssrn.com/abstract=948926
Hambrick, D.C., and Finkelstein, S., (1995), The effects of ownership structure on
conditions at the top: The case of CEO pay raises, Strategic Management Journal,
16, 175-193
Harford, J., (1999), Corporate cash reserve and acquisitions, Journal of Finance, 54, 19691997
Harford, J., Mansi, S. A., and Maxwell, W. F., (2008), Corporate governance and firm cash
holdings in the U.S., Journal of Financial Economics, 87, 535-555
Hartzell, J.C., and Starks, L. T., (2003), Institutional investors and executive compensation,
Heckman, J., (1979), Sample selection bias as a specification error, Econometrica, 47, 153161
Heckman, J., Ichimura, H., Smith, J., and Todd P., (1998), Characterizing selection bias
using experimental data, Econometrica, 66, 1017-1098
Heckman, J., Ichimura, H., and Todd P., (1997), Matching as an econometric evaluation
estimator: Evidence from evaluating a job training programme, Review of Economic
Studies, 64, 605-654
Heckman, J., and Robb R., (1986), Alternative methods for solving the problem of selection
bias in evaluating the impact of treatments on outcomes. Drawing Inference from
Self-Selected Samples, ed. H. Wainer. Berlin: Springer-Verlag
Hillion, P., and Vermaelen, T., (2004), Death spiral convertibles, Journal of Financial
Economics, 71 381415
Holderness, C. G., (2003), A survey of blockholders and corporate control, Economic Policy
Review, April 2003, 51-64
Holderness, C. G., and Sheehan, D. P., (1988), The role of majority shareholders in publicly
held corporations: An exploratory analysis, Journal of Financial Economics, 20,
317-346
Horner, M., (1988), The value of the corporate voting right: Evidence from Switzerland,
Journal of Banking and Finance, 12, 69-83
Jensen, M. C , (1986), Agency costs of free cash flow, corporate finance and takeovers,
70
Jensen, M. C , and Murphy, K. J., (1990), Performance pay and top-management incentives,
Journal of Political Economy, 98, 225-264
Johnson, S. A., Ryan, H. E., and Tian, Y. S., (2009), Managerial incentive and corporate
fraud: The sources of incentives matter, Review of Finance, 13, 115-145
Kalcheva, I., and Lins, K. V., (2007), International evidence on cash holdings and expected
managerial agency problems, Review of Financial Studies, 20, 1087-1112
Kim, K. A., and Nofsinger, J. R.., (2004), Corporate governance. New Jersey: PearsonPrentice Hall
King, M., and Santor, E., (2008), Family values: Ownership structure, performance and
La Porta, R., Lopez-de-Silanes, F., and Shleifer, A., (1999), Corporate ownership around the
world, Journal of Finance, 54, 471-517
Lambert, R., Larcker, D., and Weigelt, K., (1993), The structure of organizational
incentives, Administrative Sciences Quarterly, 38, 438-463
Lease, R. C , McConnell, J. J., and Mikkelson, W. H., (1983), The market value of control
in publicly-traded corporations, Journal of Financial Economics, 11, 439-491
Levy, H., (1983), Economic evaluation of voting power of common stock, The Journal of
Finance, 38, 79-93
Li, K., Ortiz-Molina, H., and Zhao, X., (2008), Do voting rights affect institutional
investment decisions? Evidence from dual-class firms, Financial Management^!,
713-745
Lins, K. V., (2003), Equity ownership and firm value in emerging markets, Journal of
Financial and Quantitative Analysis, 38, 159-84
Masulis, R. W., Wang, C, and Xie, F., (2009), Agency problems at dual-class companies,
McConaughy, D., Walker, M., Henderson, G., and Mishra, C , (1998), Founding family
controlled firms: Efficiency and value, Review of Financial Economics, 7, 1-19
McConaughy, D. L, (2000), Family CEOs Vs. non-family CEOs in family-controlled firms:
An examination of the level and sensitivity of pay to performance, Family Business
Review, 13, 121-132
71
McConnell, J. J., and Muscarella, C. J., (1985), Corporate capital expenditure decisions and
the market value of the firm, Journal of Financial Economics, 14, 399-422
Megginson, W., (1990), Restricted voting stock, acquisition premiums and the market value
of corporate control, The Financial Review, 25, 175-198
Mehran, D., (1995), Executive compensation structure, ownership and firm performance,
Murphy, K., (1999), Executive compensation, in O. Ashenfelter and D. Card (eds),
Handbook of Labor Economics, 3, Amsterdam: North-Holland
Nenova, T., (2003), The value of corporate voting rights and controls: A cross-country
analysis, Journal of Financial Economics, 68,325-357
Pinkowitz, L., Stulz, R., and Williamson, R., (2006), Does the contribution of corporate cash
holdings and dividends to firm value depend on governance? A cross-country
analysis, Journal of Finance, 61, 2725-2751
Rosenbaum, P., and Rubin. D., (1983) The Central Role of the Propensity Score in
Observational Studies for Causal Effects, Biometrika 70, 41-55.
Rydqvist, K., (1992), Dual class shares: A review, Oxford Review of Economic Policy, 8,
45-57
Shleifer, A., and Vishny, R. W., (1986), Large shareholder and corporate control, Journal of
Political Economy, 94, 461-488
Smith B. F., and Amoako-Adu, B., (1995), Relative Prices of Dual class Shares, Journal of
Financial and Quantitative Analysis, 30, 223-239.
Smith B. F., and Amoako-Adu, B., (1999), Management succession and financial
performance of family controlled firms, Journal of Corporate Finance, 5,341-368
Smith, B. F., Amoako-Adu, B., and Kalimipalli M., (2009), Concentrated control and
corporate value: A comparative analysis of single and dual class structures in
Canada, Applied Financial Economics, 19, 955-974
Smith, C. W., and Watts, R. L., (1992), The investment opportunity set and corporate
financing, dividend and compensation policies, Journal of Financial Economics, 32,
263-292
Tinaikar, S., (2008), Voluntary disclosure and ownership structure: An analysis of dual class
firms, Working Paper, University of Florida
72
Villalonga, B., (2004), Does diversification cause the "diversification discount"? Financial
Management, 33, 5-27
Villalonga, B., and Amit, R., (2006), How do family ownership, control and management
affect firm value?, Journal of Financial Economics, 80, 385-417
Villalonga, B., and Amit, R., (2009), How are family firms controlled?, Review of Financial
Studies, 22, 3047-3091
Whited, T. M , (2001), Is it efficient investment that causes the diversification discount?,
Zingales, L., (1994), The value of the voting right: A study of the Milan Stock Exchange
experience, Review of Financial Studies, 7,125-148
Zingales, L., (1995), What determines the value of corporate votes? The Quarterly Journal
of Economics, 110,1047-1073
73
Table 1: Summary statistics of executive compensation (2001-2007)

Significance level at the 1%, 5% and 10% is indicated as *, **, and ***, respectively. Test for difference in mean is the t-test and test for
difference in median is the Wilcoxon signed-rank.
Panel A: All Executives
($000)
Salary
Bonuses
Other Compensation
Stock Options
Total Compensation
Dual Class
Mean
Median
503.41
468.76
126.64
1216.94
2597.74
400.48
150.00
19.64
286.50
1454.35
_ "
Dev.
449.61
1280.26
646.37
3349.62
4588.05
Single Class
Obs
Mean
Median
3960
3960
3960
3960
3960
429.08
330.60
122.77
1214.37
2319.55
360.00
135.65
22.98
388.85
1280.87
Panel B: President and Chief Executive Officer

($000)
Dual Class
Salary
Bonuses
Other Compensation
Stock Options
Total Compensation
Mean
Median
779.79
1008.15
221.68
2500.67
5197.12
690.30
351.40
47.23
750.41
2828.54
Std.
Dev.
617.36
2276.79
994.37
5973.59
6060.02
'
Dev.
255.77
808.52
576.06
2631.82
3639.96
Obs
3960
3960
3960
3960
3960
Single Class
Obs
Mean
Median
819
819
819
819
819
685.33
624.52
199.27
2577.54
4743.65
650.00
283.85
46.97
1055.58
2551.79
74
Std.
Dev.
318.20
1326.66
651.05
4387.90
5180.89
Obs
808
808
808
808
808
Test for
difference
in mean
Test for
difference
in median
c 72***
0.28
0.04
2.17**
T 27***
-1.87*
-6.90***
2.08**
Test for
difference
in mean
Test for
difference
in median
2 72***
2.48**
0.16
-3.84***
2.54**
-5 0 7 * * *
414***
I 95**
-0.29
2.18**
Table 1 Cont'd
Panel C: Chief Financial Officer
Dual Class
Salary
Bonuses
Other Compensation
Stock Options
Total Compensation
Mean
Median
394.60
289.82
76.12
866.08
1775.06
344.14
113.18
17.26
276.21
1082.56
Std.
Dev.
228.17
578.07
326.41
1809.93
2389.54
1.89*
-0.39
0.26
0.18
Test for
difference
in median
3 62***
-0.26
-1.10
-2.53**
0.63
352
352
352
352
352
Test of
Mean
difference
3 g2***
2.42**
-0.53
-0.42
1.98**
Test of
Median
difference
3.63***
2.67***
-0.56
-1.87*
2.07**
Std.
Obs
Dev.
151.07 1625
409.66 1625
410.04 1625
1481.36 1625
1862.25 1625
Test for
difference
in mean
4 35***
3 02***
2.42**
0.34
2.04**
Test for
difference
in median
3 80***
2.33**
-5.08***
-5 94***
3 29***
Single Class
Obs
Mean
Median
670
670
670
670
670
349.17
247.56
84.40
819.75
1719.54
319.00
137.23
19.79
343.40
1071.70
Std.
Dev.
169.94
432.65
431.17
1484.91
2050.77
Obs
702
702
702
702
702
Test for
difference
in mean
413***
Panel D: Chief Operating Officer

Dual Class
Salary
Bonuses
Other Compensation
Stock Options
Total Compensation
Mean
Median
555.61
573.62
141.19
1164.24
2835.11
477.50
209.59
23.25
372.64
1696.83
Std.
Dev.
518.82
1406.42
427.81
2244.56
3802.98
Single Class
Obs
Mean
Median
374
374
374
374
374
442.24
359.47
174.65
1235.38
2547.76
403.70
163.90
23.44
519.26
1425.98
Std.
Dev.
208.96
906.01
1125.85
2309.04
3474.51
Obs
Panel E: Vice President

Dual Class
Salary
Bonuses
Other Compensation
Stock Options
Total Compensation
Mean
Median
357.32
229.15
60.70
766.27
1521.43
315.00
208.08
13.62
196.05
957.52
Std.
Obs
Dev.
182.70 1583
596.33 1583
215.48 1583
2142.24 1583
2501.34 1583
Single Class
Mean
Median
329.04
181.92
44.72
743.71
1398.45
300.00
109.16
21.33
274.27
887.61
75
Table 1 Cont'd
Panel F: Family CEO & Chairman
Single Class
Dual Class
Salary
Bonuses
Other Compensation
Stock Options
Total Compensation
Mean
885.67
1264.78
354.13
3982.87
7313.90
Panel G: Family CEO Only

Family CEO
Salary
Bonuses
Other Compensation
Stock Options
Total Compensation
Mean
708.79
1231.96
225.53
2534.48
5195.37
Median
800.00
369.60
103.21
616.63
2987.81
Std.
Dev.
558.35
2788.90
490.99
6889.87
9684.55
Mean
666.37
823.63
89.91
2501.99
4460.25
Obs
201
201
201
201
201
Std.
Dev.
480.71
774.55
718.69
4931.29
5128.17
Std.
Dev.
313.71
2847.06
147.66
4736.53
6346.50
Obs
94
94
94
94
94
Single Class
Dual Class
Median
650.00
420.00
32.66
937.17
2823.63
Median
582.50
278.15
16.91
875.50
1527.34
Obs
175
175
175
175
175
Mean
573.51
359.15
169.12
2623.87
4078.05
Panel H: Family CEO Vs. Non-Family CEO

Family CEO
Std.
Mean
Median
Obs
Dev.
Salary
768.60
577.54 517
675.00
Bonuses
1108.28
341.44 2700.16 517
Other Compensation
241.78
45.86 1189.98 517
3093.68
Stock Options
792.00 5233.53 517
Total Compensation
5806.49
2837.13 6042.82 517
Note: CEO and President are referred to as CEO in the above panels.
76
Mean
716.62
685.31
196.34
2288.06
4658.71
Median
581.75
14.00
48.87
1373.15
2140.08
Std.
Dev. Obs
257.75
47
708.64
47
288.07
47
3781.29
47
4542.40
47
Non-Family CEO
Std.
Median
Obs
Dev.
657.42 450.53 1110
311.00 1322.69 1110
47.94 621.17 1110
821.75 4008.14 1110
2582.60 5636.95 1110
Test for
difference
in mean
2.69***
2.25***
2 57***
1.37
2.22**
Test for
difference
in median
3 22***
1.98**
4 59***
-0.94
3 37***
Test for
difference
in mean
1.86*
2.13**
0.22
-0.11
2.25**
Test for
difference
in median
2.09**
3 go***
-0.86
-1.10
2.06**
Test for
difference
in mean
I 97**
4 23***
1.01
Test for
difference
in median
2.44**
2.47**
-0.66
-1.43
2.86***
2 07***
3 ii***
Table 2: Number of dual class firms by two-digit SIC industry code

This table illustrates the two-digit SIC code industry distribution of 792 firm-year observations over
the period 2001-2007.
SIC Code
Industry description
% of total
10-17
Construction
2.78
20
Food and Kinder Products
7.70
22
Textile Mill Products
2.40
23
Appeal and other Finish Products
1.89
27
Printing and Publishing
8.21
28
Chemicals and Allied Products
4.92
29
Petroleum Refining and Related Industries
0.51
30
Rubber and Misc. Plastic Products
1.39
31
33
Leather and Leather Products
1.89
0.25
34
Primary Metals
Fabricated Metal Product
35
36
Industrial Machinery and Equipment

Electronic and other Electrical Equipment
37
Transportation Equipment
38
39
42
Instrument and Related Products

Misc. Manufacturing Industries
7.58
3.41
3.54
1.52
Motor Freight Transportation and Warehousing

Transportation Services
1.14
0.76
Communications
Wholesale Trade
7.95
2.90
Retail trade
Deposit Institution, and Non-depository Credit Institutions
Security And Commodity Brokers, Dealers, Exchanges, & Services
Insurance
Holding And Other Investment Offices
8.59
1.52
2.27
3.66
1.77
1.52
47
48
50-51
52-59
60-61
62
63
67
72
73
Personal Services
Business Services
0.76
4.04
9.09
79
Motion Pictures
Amusement and Rec. Services
1.77
2.02
80
82
87
Health Services
Educational Services
Engineering and Management Services
1.01
0.51
0.76
78
Note: The industry breakdown for the single class concentrated control companies is exactly the
same as above due to matching by industry group.
77
Table 3: Summary statistics of governance characteristics, firm characteristics and executive characteristics
Significance level at the 1%, 5% and 10% is indicated as *, **, and ***, respectively. Test for difference in mean is the t-test and test for
difference in median is the Wilcoxon signed-rank test.
Panel A: Governance Characteristics
Dual Class
Mean
Median
Single Class
Test for
difference
in mean
Test for
difference
in median
Mean Median
_.
Dev.
Dev.
-0.62
Cash Flow Rights (Largest Shareholder) %
22.50
18.20 16.50
23.60
19.20
13.20 -0.71
34
33***
Voting Rights (Largest Shareholder) %
57.80
54.80 25.70
15.75***
23.60
19.20
13.20
447***
Cash Flow Rights (Management & Directors) %
24.90
19.30 15.60
17.10
16.50
14.10 14.21***
Voting Rights (Management & Directors) %
58.30
57.30 25.20
15.82***
17.10
16.50
14.10 44 25***
Mgmt. Voting Lev.
3.01
2.42
2.15
82.55***
1.00
1.00
0.00
Institutional Ownership %
16.92
13.60 15.66
21.49
18.71
14.96 -13 24*** -15.65***
Board Size
10.89
10.00
3.89
0.08
10.32
10.00
3.48 0.87
Independent directors %
59.29
58.33 15.39
68.81
71.43
16.33 -26.71*** -27.01***
13.00
11.11
15.34 5 j9***
Busy Directors %
16.58
12.50 14.72
6.88***
Grey Directors %
10.23
9.09 11.75
9.89
7.69
12.14 3 o***
4.58***
Family Directors %
15.80
12.50 12.73
5.62
0.00
10.60 34.83***
37.86***
Note: Observation for dual class sample is 792 firm-year and single class sample is 792 firm-year. A grey director is defined as an outside
director who is related to the company through a transactional relationship and a busy director is a director with more than 4 board
memberships.
78
Table 3 cont'd
Panel B: Firm Characteristics
Single Class
Dual Class
Mean
Median
Std. Dev.
Mean
Median
Test for
difference
Std. Dev.
in mean
1.25 -3 02***
2.26 -2.00**
15302.38 0.78
28.12 -0.29
38.52 2.18**
9.30 0.86
18.96 -0.58
16.64 1.07
841.11 2.88***
119.54 1.79*
0.095 1.99**
308.65 3 33***
132.88 2.24**
0.068 9 21***
2.04
Tobin's Q ratio
1.24
1.62
1.87
1.47
Industry Adjusted Q
-0.41
1.73
-0.39
-0.35
-0.59
Sales ($m)
4917.45 1523.90 14597.23 4653.87 1629.00
Risk- Std Deviation %
40.27
40.88
31.10
30.79
31.46
Performance - RET %
12.61
34.87
10.80
7.45
8.35
Performance - ROA %
8.81
9.75
8.87
9.57
9.06
Financial Leverage -D/A
20.34
21.81
20.82
21.61
18.10
Growth (%)
9.98
6.67
15.61
9.37
6.46
Cash ($m)
775.81
431.08
132.80
642.28
166.00
ACash, ($m)
24.98
5.27
48.99
5.14
508.54
0.012
0.002
ACash, / Total Asset,.]
0.028
0.003
0.173
Capex ($m)
138.34
256.54
518.22
46.51
65.50
ACapex ($m)
7.44
21.73
143.35
1.36
1.78
ACapex/Market Value Equity
0.024
0.007
0.001
0.007
0.065
Note: The number of observations for each of the dual and single class samples is 792 firm-year.
Industry Adjusted Q-Mean is defined as the difference between the firm's Q ratio and the average SIC Industry Q ratio.
79
Test for
difference
in median
-4 53***
-1.98**
-0.89
-0.69
1.58
-0.04
-0.28
1.08
3 ig***
-1.03
1.73*
3.84***
0.78
9 QQ***
Table 3: Cont'd
Panel C: Executive Characteristics
Dual Class
Family
Executives
Chairman only
Chief Executive Officer & Chairman
Chief Executive Officer Only
Chief Financial Officer
Chief Operating Officer
Vice President
Other Executives
Total
115(63.5%)
201 (51.8%)
175 (40.6%)
21(3.1%)
47(12.6%)
81(5.1%)
18(6.1%)
658 (16.8%)
Non-Family
Executives
Single Class
Trttal
Family
Executives
66 (36.5%) 181
187(48.2%) 388
256 (59.4%) 431
649 (96.9%) 670
327 (87.4%) 374
1502(94.9%) 1583
275 (93.9%) 293
3262 (83.2%) 3920
33 (34.0%)
94(20.4%)
47 (13.5%)
13(1.8%)
12 (3.4%)
26(1.6*)
6(1.8%)
231(5.9%)
80
Non-Family
Executives
Total
64 (66%)
97
366 (79.6%) 460
301 (86.5%) 348
689 (98.2%) 702
340 (96.6%) 352
1599(98.4%) 1625
330 (98.2%) 336
3689(94.1%) 3920
Table 4: Regression of executive compensation on management voting leverage, firm

and governance characteristics
For a complete definition of the variables see Appendix A, Table 1A
5
Ln( Compensati on)l
1
t_x + X "!, kFirm Chara cteristics
= a ( + (j)( Mgmt Vote
k=\
+ X S ( k Governance
k=\
where Ln( Compensati on)l
'
'J'
,_,
'
, + Z Y, A Position'k,jt
k=\
+ Z
i,j,t
, is the natural log of the /"'components as well as total compensation
for all senior executives' compensation package for firm j at year /. T-statistics are adjusted for
robust standard errors and reported below the estimated coefficient. Significance level at the 1%, 5%
and 10% is indicated as *, **, and ***, respectively.
Mgmt. vote
Firm Characteristics
Performance (ROA)
Salary
Estimates
t-statistics
0.021
10.53***
-0.238
-3 80***
-0.036
Financial Leverage
(D/A)
-0.95
Size (Log of sales)
0.428
35 49***
-0.032
Risk (Std. Dev.)
-1.76*
-0.232
Growth
-6.09***
Governance Characteristics
Board Size
0.014
6.22***
Independent Directors 0.071
1.59
Busy Directors
0.069
1.90*
Grey Directors
0.198
440***
0.242
Fam. Directors
3 yg***
Institutional
-0.025
Ownership
-0.88
Bonuses
Estimates
t-statistics
0.033
6 13***
Other
Compensation
Estimates
t-statistics
0.021
2.15**
Stock
Options
Estimates
t-statistics
0.016
2.67***
Total
Compensation
Estimates
t-statistics
0.029
8.80***
0.601
2 22***
0.212
2.67***
0.848
25.82***
0.153
3 02***
0.205
1.89*
0.829
3 4i***
0.159
1.44
0.751
18.89***
-0.180
-2.57**
-0.905
-7 08***
0.760
3 59***
-0.613
-5.84***
1.003
29 39***
0.581
0.320
2.06**
-0.283
-5 51***
0.862
44.00***
0.103
2 go***
0.563
6 94***
0.026
4 23***
0.037
0.35
0.118
1.17
0.012
0.10
0.231
1.85*
0.234
2 on***
0.017
2.52**
0.207
1.69*
0.047
0.36
-0.182
-1.20
0.025
0.15
0.409
3.68***
81
Q4Q***
1.381
12.98***
0.027
4 04***
0.136
1.12
0.331
3 no***
0.124
0.86
-0.338
-2.36**
0.054
0.57
0.014
4 39***
0.383
6.15***
0.254
4.02***
0.193
2.58***
-0.251
-3 24***
0.085
1.60
Table 4 Cont'd
Executive Positions
CEO & Chairman
CEO Only
Chair Only
CFO
COO
VP
Intercept
Industry &
Year Effects
Adjusted R2
Obs.
Salary
Estimates
t-statistics
Bonuses
Estimates
t-statistics
Other
Compensation
Estimates
t-statistics
Stock
Options
Estimates
t-statistics
Total
Compensation
Estimates
t-statistics
0.613
2959***
0.421
17 79***
0.302
4.05***
-0.11
-5 92***
0.117
5 40***
-0.156
-9 00***
4.304
90 51***
1.055
17 89***
0.839
14.58***
0.731
7 39***
-0.091
-1.80*
0.272
4 39***
-0.245
-5.36***
1.807
13.00***
1.009
12 99***
0.575
7 j6***
0.778
6.48***
-0.033
-0.48
0.201
2.47**
-0.033
-0.54
-0.008
-0.04
1.353
22.40***
1.15
17 58***
0.725
6.23***
0.146
2 59***
0.473
7 32***
-0.044
-0.83
2.055
13 33***
1.008
28.05***
0.821
21 43***
0.471
7.65***
-0.036
-1.13
0.281
7 49***
-0.148
-5 13***
3.694
44.63***
Yes
Yes
Yes
Yes
Yes
0.465
7920
0.359
5585
0.223
7333
0.427
5894
0.483
7920
82
Table 5: Panel regression of executive compensation on family executives,

management voting leverage, firm and governance characteristics
For a complete definition of the variables see Appendix A, Table 1A
6
Ln(Compensation)
- a, + <P, kFamilyPosition
_i
i,j,i
'
+,MgmtVote
j,t-\
k,j
+ Z'H, kFirm Characteristicsk
tA
+ Y.&,,kGovernancek
]tA
k=\
6
T.y,ikPositionkj+e,Jit
k=\
where Ln(Compensati on)t
, is the natural log of the /* components as well as total compensation
for all senior executives' compensation package for firm j at year /. T-statistics are adjusted for
robust standard errors and reported below the estimated coefficient. Significance level at the 1%, 5%
Other
Stock
Total
Salary
Bonuses
Compensation
Compensation Options
Estimates Estimates Estimates
Estimates Estimates
t-statistics t-statistics t-statistics
t-statistics t-statistics
Fam. CEO &
0.004
0.03
0.068
0.027
0.035
-2 2i***
1.65*
Chairman
0.13
0.58
1.79*
Fam. CEO Only
0.061
0.122
0.042
0.035
0.023
3 oi***
3.65***
2.28**
0.53
0.97
0.027
Fam. Chair Only
0.217
-0.009
0.122
0.007
2.61***
1.56
-0.71
0.57
0.38
0.02
0.014
0.04
Fam. CFO
0.126
0.01
0.64
0.04
3 os***
1.29
0.67
0.053
0.027
0.029
Fam. COO
0.177
0.38
3 7i***
5 34***
2 75***
2.40**
1.52
0.001
Fam. VP
-0.048
-0.039
0.20
0.053
-1.00
0.03
-1.38
1.58
1.45
0.035
0.018
0.002
0.017
Mgmt. Voting Lev.
0.020
3 34***
4.22***
1.95*
0.25
10.07***
0.526
0.614
0.827
0.757
Performance
-0.233
-j 29***
-3 70***
3.40***
3.65***
2.01**
(ROA)
-0.285
0.254
0.151
-0.549
-0.013
Financial Leverage
-5 09***
2.88***
-5.46***
-0.33
1.39
(D/A)
0.860
0.846
0.749
0.996
Size (Log of sales)
0.426
43.81***
18.81***
28.77***
34.89***
25.75***
0.564
1.383
-0.232
0.216
-0.91
Risk (Std. Dev.)
-7 12***
13.02***
7 00***
-6.11***
2.00**
0.100
0.569
-0.032
0.155
-0.178
Growth
g
27***
2
70***
-1.74*
-2.55**
3.06***
83
Table 5 Cont'd
Bonuses
Estimates
t-statistics
Other
Compensation
Estimates
t-statistics
Stock
Options
Estimates
t-statistics
Total
Compensation
Estimates
t-statistics
0.208
3 53***
-0.028
-0.97
0.025
4.06***
0.044
0.41
0.137
1.35
0.033
0.27
0.194
1.54
0.226
2.81***
0.016
2.45**
0.209
1.71*
0.062
0.47
-0.179
-1.18
-0.072
-0.40
0.407
3.65***
0.028
5.03***
0.118
0.97
0.354
3 29***
0.113
0.78
-0.337
-2.34**
0.028
0.29
0.014
4.43***
0.373
5.98***
0.267
4.22***
0.192
2.57**
-0.266
-3 42***
0.07
1.32
0.42
16.17***
0.196
1.42
0.615
26.63***
-0.109
_5 77***
0.108
4.81***
-0.151
-8.46***
4.308
90.87***
0.789
13 03***
0.755
6.98***
1.028
16.64***
-0.091
-1.80*
0.263
4.22***
-0.242
-5 27***
1.805
12.88***
0.548
6.23***
0.725
4.30***
0.988
11.40***
-0.030
-0.45
0.179
2.18**
-0.038
-0.62
0.018
0.10
1.118
16.61***
0.733
5 64***
1.327
21.58***
0.145
2.57**
0.462
7.08***
-0.048
-0.91
2.107
13.41***
0.796
19.98***
0.440
6 53***
0.978
26.47***
-0.037
-1.18
0.273
7.22***
-0.147
-5 10***
3.736
44.46***
Yes
Yes
Yes
Yes
Yes
0.495
7920
0.392
5585
0.221
7333
0.451
5894
0.512
7920
Salary
Estimates
t-statistics
Board Size
0.014
6.05***
0.071
Independent
Directors
1.59
Busy Directors
0.075
2.05**
Grey Directors
0.198
4 4Q*H=*
Fam. Directors
Institutional
Ownership
Executive Positions
CEO & Chairman
CEO Only
Chair Only
CFO
COO
VP
Intercept
Industry &
Year Effects
Adjusted R2
Obs.
84
Table 6: Panel regression for a sub-sample of family executives

For a complete description of the variables see Appendix A, Table 1 A.
6
Ln(Compensati
on)'
, = a , + (p, k(DFam .Execu .k

k=\
x Mgmt. Vote
t_x)
'
5
+ tMgmt Vote
t_x
+ Yir\l kFirm Char a cteristics

*=1
'
6
+ Y.^,,kGovernance
*=1
,_, + Z y , kPosition
+s
A= l
where Ln( Compensati on)x , is the natural log of the /* components as well as total compensation
of each executives' compensation package for firm j at year / and DFam.Execu. is an indicator
variable equal to 1 the kth family executive in they'"1 dual class firm and zero otherwise. T-statistics
are adjusted for robust standard errors and reported below the estimated coefficients. Significance
level at the 1%, 5% and 10% is indicated as *, **, and ***, respectively.
Other
Stock
Total
Salary
Bonuses
Compensation
Compensation Options
Estimates Estimates Estimates
Estimates Estimates
t-statistics t-statistics t-statistics
DFam. CEO &Chair x 0.017
0.041
0.103
0.008
0.026
2 7Q***
2.12**
Mgmt. Vote
2.12**
1.71*
0.48
0.021
0.074
0.036
DFam. CEO only x
0.036
-0.020
2.00**
2.57**
0.89
-0.86
2.28**
Mgmt. Vote
0.021
0.048
DFam. Chair only x
0.001
0.141
-0.010
2 93***
4_]4***
0.05
4.66***
Mgmt. Vote
-0.56
0.008
DFam. CFO x Mgmt.
-0.003
0.050
-0.006
0.023
2.00**
0.51
Vote
-0.07
0.89
-0.33
0.011
0.032
DFam. COO x Mgmt.
0.022
0.009
0.132
2 07***
2 74***
7ig***
0.22
Vote
0.76
-0.054
-0.043
DFam. VP x Mgmt.
-0.016
0.073
-0.051
-0.85
-1.27
Vote
-0.86
1.44
-1.09
0.344
0.238
0.035
0.435
Mgmt. Vote
0.069
4
40***
6.24***
2.43**
1.84*
0.35
0.999
-0.112
0.749
0.841
1.701
Performance (ROA)
1.13
1.99**
2.03**
-0.46
1.15
0.108
-0.103
0.172
0.163
1.526
Financial Leverage
0.52
1.50
0.54
-0.24
(D/A)
3.56***
0.886
0.771
0.487
1.178
Size (Log of sales)
0.449
10.98**:
6.60***
3.06***
6.78***
12.09***
-0.004
1.216
-0.050
-0.036
0.022
Risk (Std. Dev.)
2
72***
-0.030
-0.26
-0.13
0.31
0.697
1.134
-0.656
0.689
-0.350
Growth
2.56**
2.80***
1.06
-2.31**
-1.6
85
Table 6 Cont'd
Other
Bonuses
Compensation
Estimates Estimates
Stock
Total
Options
Compensation
Estimates Estimates
0.033
1.90*
0.254
0.67
0.453
1.49
0.161
0.37
-0.152
-0.34
0.531
1.43
0.054
2.44**
-0.004
-0.01
0.493
1.13
-0.226
-0.46
-0.601
-1.01
-0.338
-0.80
-0.015
-0.69
0.404
0.98
-0.030
-0.08
-0.516
-1.14
0.017
0.03
-0.164
-0.36
-0.012
-1.08
0.531
2.55**
0.430
2.01**
-0.14
-0.52
-0.029
-0.11
-0.030
-0.13
0.623
1.34
0.377
0.80
0.334
0.70
-0.186
-0.31
-0.58
-1.08
-0.249
-0.51
0.215
0.29
1.457
3 08***
1.559
2 30***
0.841
1.69*
0.424
0.76
0.724
1.43
0.890
1.71*
0.426
0.54
0.571
-0.072
-0.66
-0.343
-3 39***
4.486
25.30***
0.429
2.05**
0.19
0.86
0.195
0.84
-0.112
-0.37
-0.048
-0.17
-0.748
-2.84***
2.371
c 27***
0.461
271 ***
-0.174
-0.90
0.012
0.05
-0.258
-1.32
-0.068
-0.35
3.375
11.08***
Yes
Yes
Yes
Yes
Yes
0.541
889
0.576
601
0.389
804
0.396
571
0.601
889
Salary
Estimates
t-statistics
Board Size
0.023
2 97***
Independent Directors -0.186
-1.65*
Busy Directors
-0.012
-0.11
Grey Directors
-0.156
-0.94
Family Directors
0.176
1.22
Institutional
-0.381
-3
34***
Ownership
Executive Positions
CEO & Chairman
0.277
3.00***
CEO Only
0.083
0.84
Chair Only
0.065
0.67
-0.402
CFO
-3 15***
COO
VP
Intercept
Industry &
Year Effects
Adjusted R2
Obs.
86
2 4 j # * *
Table 7: Panel regression of dual class discount on excess corporate cash holdings
Dual Class Discount
/+1
= a + (3j {Excess Cash x Mgmt.Vote)
+ pÊxcess Cash
+ (3^Financial Leverage
, + (36 Size
+ 2Mgmt. Vote
+ fi7Div. diff
+s
where dual discount in models (1) & (2) is the difference between Tobin's Q ratio of a dual class
company and the Q ratio a single class concentrated control company using propensity score
matching. The dual class discount in models (3) and (4) is the difference between Q ratio of a dual
class firm and its industry mean Q ratio.
(
A cash.
A casht t
Total assets t-\J Dual class V Total assets t-\ J Single class
Excess cash is defined as v
' "<"
^
' "wpeciass w h e r e t h e
change in cash is cash and marketable securities at the end of year t minus year t-1. Conversion right
is an indicator variable equal to 1 if superior voting shares can be converted into subordinate voting
shares and 0 otherwise. Size is the natural logarithm of market value of equity. Div. diff is an
indicator variable equal to 1 if dividend paid or payable to subordinate voting shares is greater than
dividend paid or payable to superior voting shares and zero otherwise. T-statistics are adjusted for
robust standard errors and reported below the estimated coefficients. Significance level at the 1%,
5% and 10% is indicated as *, **, and ***, respectively.
Excess Cash x Mgmt. Vote

Mgmt. Vote
Excess Cash
(1)
Estimates
t-statistics
-0.358
-3.68***
-0.134
-9 22***
1.097
3.67***
ACashf
ACash, x Casht. i f
ACash, x Financial Leverage !
Financial Leverage
Conversion Rights
Size
-0.044
-3 08***
-0.402
-5.88***
0.118
5.60***
(2)
Estimates
t-statistics
-0.414
-3.63***
-0.136
-9 35***
1.308
3 42***
-0.014
-2 90***
0.02
0.81
0.001
1.95*
-0.039
-2 79***
-0.400
-5.81***
0.122
5.77***
87
(3)
Estimates
t-statistics
-0.474
_4 4j***
-0.053
-3 81***
1.478
3 95***
-0.054
-3 02***
-0.266
-4 72***
0.204
g 44***
(4)
Estimates
t-statistics
-0.188
-2.00**
-0.052
-3 70***
0.292
0.86
-0.013
-3 11 ***
0.12
7Q5***
0.001
0.86
-0.062
-3.46***
-0.274
_4 9 ] * * *
0.193
8.01***
Table 7 Cont'd
i!)
Div. Difference
Intercept
Estimates
t-statistics
-0.152
-1.16
-0.486
-2.65***
Yes
Yes
(?)
(3)
(4)
Estimates
t-statistics
-0.144
-1.11
-0.515
-2.80***
Yes
Yes
Estimates
t-statistics
0.688
j95***
-1.178
-6.65***
Yes
Yes
Estimates
t-statistics
0.689
Y 94***
-1.083
-6.15***
Yes
Yes
0.136
792
0.159
792
Industry Effects
Year Effects
Adjusted R2
0.151
0.147
Obs.
792
792
Note: Coefficients are multiplied by 100 for presentation
88
Table 8: Panel regression of dual class discount on capital expenditure

Dual Class Discount
t+]
= a + Pj (Capex x Mgmt.Vote)
+ fi3Capex
+ ^1Mgmt. Vote
+ p^Financial Leverage
+ p 6 Size j t + p7Z)zv. diffJt
+ ey_,
where dual discount in model (1) is the difference between Tobin's Q ratio of a dual class company
and the Q ratio a single class concentrated control company using propensity score matching. The
dual class discount in model (2) is the difference between Q ratio of a dual class firm and its industry
mean Q ratio.
Capital Expenditure (Capex) is defined as
ACapex, t
ACapex, t
Market Value., _ 1 , < . , , ,
/ Dual class
Market Value, . .
, ,
' ~ ' / Single class
T-statistics are adjusted for robust standard errors and reported below the estimated coefficients.
Significance level at the 1%, 5% and 10% is indicated as *, **, and ***, respectively.
Capex x Mgmt. Vote

Mgmt. Vote
Capex !
Financial Leverage
Conversion Rights
Size
Div. Difference
Intercept
Industry Effects
Year Effects
(1)
Estimates
t-statistics
0.053
0.93
-0.086
-4.65***
-0.011
-3.86***
-0.010
-0.50
-0.463
-5 24***
0.110
2 yo***
-0.091
-0.48
-0.618
-2.33**
Yes
Yes
0.056
Adjusted R2
792
Obs.
Note: ' Coefficients are multiplied by 100 for presentation.
89
(2)
Estimates
t-statistics
-0.041
-1.08
-0.051
_3 49***
-0.010
-3 23***
-0.033
-1.62
-0.299
-5.11***
0.233
8.35***
0.715
giy***
-1.365
-7 01***
Yes
Yes
0.082
792
Table 9: Fixed effects regression of dual class discount on excess compensation and excess corporate cash holdings
5
Ln( Compensation )t
- al + ^r\t
kFirm Characteristicsk
t_x + X $ , kGovernancek
*=i
t_x + Y
*=i
Position
+s
where / is equal to different components as well as total compensation for each executives' compensation package for firmy" at year t.
Dual Class Discount
(+1
= a + $xExcess Comp.
+ ^1{Mgmt. Vote x Excess Cash)
+ $3Mgmt. Vote
+ $4Excess
Cash
+ ^Financial
Leverage
+ $6Conversion
Right
, + P 7 Size}
+ p8Dz'v. diffJt + Y,JtPositionk
+8 ,
Excess compensation y , is the residual from the above equation and it represents excess salary (column I), excess bonuses (column II), excess other
compensation (column III), excess stock options (column IV) and excess total compensation (column V). Column VI reports estimates for the
interaction term between the dummy variable for dual class executives and excess total compensation. Excess cash, dividend difference (div. diff),
financial leverage, size and conversion rights are the same as in Table 7. T-statistics are adjusted for robust standard errors and significance level at the
1%, 5% and 10% is indicated as *, **, and ***, respectively.
Panel A: Dependent variable -dual discount is the difference between Tobin's Q ratio of a dual class company and the Q ratio of a single class
concentrated control company using propensity score matching.
1^
II
in
IV
V^
VI
Est.
t-stat
Est.
t-stat
Est.
t-stat
Est
t-stat
Est
t-stat
Est.
t-stat
Excess Compensation
-0.162
-2.70*** -0.063
Mgmt. Vote x Excess
Cash
-0.351
.3.43*** _0.279
Mgmt. Vote
-0.133
-9.11*** -0.129
DFam. CEO & Chair x Excess Compensation
DFam. CEO only x Excess Compensation
DFam. Chair only x Excess Compensation
-1.61
-0.064
-2.43** -0.423
-7.65*** -0.128
90
-3.82*** 0.115
3.65***
-0.131
-3.35*** -0.106
-2.36**
-3.42*** -0.581
-8.84*** -0.160
-3.88*** -0.342
-9.00*** -0.134
-3.50*** -0.067
-8.86*** -0.364
-0.351
-0.455
0.057
-8.88***
-3.59***
-2.78***
-1.99**
0.45
Table 9 Cont'd
I
Est.
t-stat
DFam. CFO x Excess Compensation
DFam. COO x Excess Compensation
DFam. VP x Excess Compensation
T ^2***
Excess Cash
1.077
-3 21***
Financial Leverage
-0.046
Conversion Rights
-0.402
-5.83***
5 42***
Size
0.115
Div. Diff
-0.164
-1.26
Intercept
-1.74*
-0.347
Executive Positions
Yes
Industry and Year Fixed
Yes
Effects
Adjusted R2
0.166
Obs.
3960
Est.
0.890
-0.044
-0.605
0.148
-0.254
-0.353
II
t-stat
2.57**
-3 03***
-7 12***
5.72**
-1.59
-1.49
Yes
in
Est.
1.503
-0.047
-0.375
0.118
-0.132
-0.320
t-stat
5 12***
-3 70***
5 29***
5.53***
-1.03
-1.67*
Yes
rv
Est.
t-stat
1.562
-0.070
-0.573
0.098
0.277
-0.112
3.55***
-2.81**
-6.97**
3.52**
2.37**
-0.43
Yes
V
t-stat
Est.
1.041
-0.050
-0.405
0.137
-0.178
-0.441
3 4i***
-3 23***
-5.84***
5 04***
-1.30
-2.19**
Yes
Est.
0.285
0.141
0.050
1.110
-0.052
-0.409
0.127
-0.135
-0.622
VI
t-stat
0.74
0.67
0.32
3 53***
-3.49***
_5 95***
5 96***
-1.04
-3 22***
Yes
Yes
Yes
Yes
Yes
Yes
i0.113
i0.161
I0.106
2860
3731
2809
0.191
3960
0.221
3960
91
Table 9: Cont'd
Panel B: Dependent variable - dual discount is the difference between Tobin's Q ratio of a dual class company and its industry mean Q ratio.
I
Est.
-0.046
t-stat
-1.82*
Excess Compensation
Mgmt. Vote x Excess
-3 35***
Cash
-0.378
Mgmt. Vote
-0.051
-3.60***
DFam. CEO & Chair x Excess Compensation
DFam. CEO only x Excess Compensation
DFam. Chair only x Excess Compensation
DFam. CFO x Excess Compensat ion
DFam. COO x Excess Compensation
DFam. VP x Excess Compensation
2 74***
1.102
Excess Cash
-0.058
Financial Leverage
-3 20***
-4 92***
Conversion Rights
-0.276
Size
0.199
8.21***
1 5***
Div. diff
0.681
Intercept
-1.149
-6.23***
Yes
Executive Positions
Industry and Year
Yes
Fixed Effects
0.093
Adjusted R2
3960
Obs.
II
Est.
t-stat
-0.013 -0.41
-0.546
-0.056
1.670
-0.059
-0.331
0.212
0.698
-1.259
in
rv
Est.
-0.051
t-stat
Est.
_3 19*** 0.042
t-stat
1.63
_4.64*** -0.433
-3 57*** -0.025
_4 5i*** -0.185
-1.90*
-0.055
-0.455
-1.69*
_3 34*** -0.047
4 IT***
5 35***
1.754
-3.05*** -0.042 -2.55**
_4 7i*** -0.284 -5.16***
7 QQ***
7 15***
0.184
7 gi***
6.41*** 0.677
-6.12*** -1.303 -7 24***
Yes
Yes
0.622
-0.006
-0.293
0.213
0.600
-1.030
1.48
-0.19
-4.42***
7 26***
4 g3***
-4.34***
Yes
Est.
-0.204
1.380
-0.055
-0.244
0.220
0.643
-1.342
V
VI
t-stat
Est.
t-stat
c 7i*** -0.170 -4.61***
_4 5j*** -0.030 -4.34***
-3 35*** -0.154 -4 49***
-0.383 -3 07***
-0.207 -1.93*
0.27
0.049
-0.045 -0.17
-0.212 -0.56
0.213
1.31
3 gg*** 1.359
3 9i***
-3.06*** -0.055 -3 09***
-4.40*** -0.250 -4.38***
9 in***
9 08*** 0.221
7 4i*** 0.650
7 30***
-7.26*** -1.427 -8.02***
Yes
Yes
Yes
Yes
Yes
Yes
Yes
0.105
2860
(3.085
(3.101
3731
2809
0.105
3960
0.109
3960
92
Table 10: Pair-wise test of separation of voting and cash flow rights on executive compensation: Fixed effects and 2SLS
Note: This table presents results of the fixed effects pair-wise regressions using the Two Stage-Least Square approach. The results reported in column I
are the second stage estimates for equation (10) using the predicted management voting leverage estimated using equation (9). Column II reports the
results of estimating equation (5) which examines executive compensation of family members using the predicted management voting leverage from
equation 9. Column III reports the results of re-estimating equation (6) using predicted management voting leverage for a sub-sample of family
executives only. For a complete description of the variables, see Table 1. Significance level at the 1%, 5% and 10% is indicated as *, **, and ***,
respectively.
ra
Dependent Variable: Ln (Total Compensation)
Mgmt. Voting Leverage
Estimates
0.387
t-statistics
21 94***
Estimates
0.379
3 Qg***
0.26
38.62***
2.44**
6.88***
4.60***
6.61***
4 49***
3 Qg***
0.022
0.035
0.024
0.036
0.023
0.004
0.383
0.019
0.764
0.086
0.534
0.013
0.402
0.278
0.246
Fam. CEO & Chairman

Fam. CEO Only
Fam. Chair Only
Fam. CFO
Fam. COO
Fam. VP
Performance
Financial Leverage (D/A)
Size
Risk (Std. Dev.)
Growth
Board Size
Independent Directors
Busy
Grey
0.369
0.014
0.769
0.085
0.535
0.014
0.395
0.271
0.222
t-statistics
21.70*** Mgmt. Voting Leverage
DFam. CEO & Chair x Mgmt.
Vote
2.13**
3 28***
DFam. CEO only x Mgmt. Vote
DFam. Chair only x Mgmt. Vote
2.36**
2 03***
DFam. CFO x Mgmt. Vote
4g3***
DFam. COO x Mgmt. Vote
DFam. VP x Mgmt. Vote
0.19
3 19***
Performance
0.33
38.43*** Size
6.89***
Risk (Std. Dev.)
Growth
2.49**
4 34***
Board Size
6 74***
Independent directors
4.62***
Busy
3 4]***
Grey
93
Estimates
0.344
tstatistics
6.20***
0.031
0.030
0.047
0.008
0.030
-0.060
1.107
0.088
0.888
0.027
-0.670
-0.012
0.493
0.436
-0.173
2.32**
1.85*
3 99***
0.48
3.68***
-1.28
2.22**
0.42
10.96**
2 47***
-0.18
-1.07
2.37**
2.05**
-0.64
Table 10 Cont'd
I
Family Directors
Inst. Investors
Intercept
Executive Positions
Industry and Year Effects
Adjusted R2
Obs.
Estimates
-0.230
0.083
3.207
t-statistics
-3 10***
1.62
3819***
Yes
Yes
0.518
7920
in
II
Estimates t-statistics
-3 81***
-0.283
1.64
0.082
3.242
38.50***
Yes
Yes
0.529
7920
94
Family Directors
Inst. Investors
Intercept
Executive Positions
Adjusted R2
Obs
-0.098
-0.36
-0.060
-0.25
11 29***
3.423
Yes
Yes
0.601
881
Table 11: Total Compensation based on the Heckman two stage treatment effects model
This table reports the second stage coefficients of estimates from Heckman two stage treatment effect models. In the first stage, we run the probit
selection model where the dependent variable is an indicator variable that equals to 1 if a firm has more than one class of common shares and 0
otherwise. I include Lambda (inverse Mills ratio) in the second stage with control valuables. The dependent variable in the second stage is the Ln(total
compensation). Model (1) is similar to the last column in Table 4, Model (2) includes indicator variables for various executive positions held by family
members and Model (3) examines a sub-sample of only family executives. For a complete description of the variables see Appendix A, Table 1A. Tstatistics are adjusted for robust standard errors and significance level at the 1%, 5% and 10% is indicated as *, **, and ***, respectively.
Model (1)
Model (2)
Model (3)
Estimates t-statistics Estimates t-statistics
3 42***
8.82***
0.034
0.030
0.017
6.17***
3
]3***
2.04**
Fam. CEO & Chairman
0.034
DFam. CEO & Chair x Mgmt. Vote 0.025
3 oi***
Fam. CEO Only
0.035
DFam. CEO only x Mgmt. Vote
0.037
2.29**
2 2***
412***
Fam. Chair Only
DFam. Chair only x Mgmt. Vote
0.048
0.029
306***
Fam. CFO
DFam. CFO x Mgmt. Vote
0.008
0.50
0.040
Fam. COO
5.28***
0.032
0.028
DFam. COO x Mgmt. Vote
3.84***
Fam. VP
0.04
0.001
DFam. VP x Mgmt. Vote
-0.050
-1.26
Performance
0.80
0.76
Performance
1.046
2.11**
0.100
0.095
_5 3i***
.4 89***
-0.276
Financial Leverage
0.106
0.51
-0.276
43
79***
Size
0.862
43.58***
Size
0.884
10.84***
0.859
2 74***
7 14***
Risk (Std. Dev.)
-0.007
-0.04
0.100
0.579
Risk (Std. Dev.)
7 09***
Growth
2.66***
Growth
0.700
2.56**
0.579
0.097
4 49***
4 5i***
Board Size
0.014
0.014
Board Size
-0.012
-1.05
5 gj***
5.75***
0.558
2.41**
0.391
0.387
3 9g***
419***
Busy
0.254
Busy
0.438
2.02**
0.268
Grey
0.203
2.66***
0.202
2.64***
Grey
-0.159
-0.59
Family Directors
-0.202
-1.85*
-2.14**
-0.233
Family Directors
-0.010
-0.03
Inst. Investors
1.26
0.069
1.01
Inst. Investors
-0.031
0.055
-0.13
95
Table 11 Cont'd
Lambda (Inverse Mills Ratio)

Intercept
Executive Positions
Adjusted R2
Obs.
Model (1)
0.014
0.52
3.668
43.05***
Yes
Yes
0.456
7920
Model (2)
0.008
0.32
3.713
42.84***
Yes
Yes
0.484
7920
96
Model (3)
Lambda (Inverse Mills Ratio)
Intercept
Executive Positions
Adjusted R2
Obs.
Estimates
0.002
3.379
t-statistics
0.01
10.96***
Yes
Yes
0.583
881
Table 12: Panel regression of total compensation on management voting leverage, age
and equity ownership
The dependent variable is the natural logarithm of total compensation. Share ownership is the
percentage of equity owned by an executive. For a complete definition of the variables see Appendix
A, Table 1A. T-statistics are adjusted for robust standard errors and significance level at the 1%, 5%
Estimates t-statistics Estimates t-statistics Estimates t-statistics
Mgmt. Vote
0.034 6.98***
0.043 7.05***
0.048 5.56***
Executive Age
-0.001 -0.73
0.001 0.56
Equity Ownership
-0.741 -2.08**
-0.636 -1.96**
Performance (ROA)
0.454 2 15***
0.23 2.14**
0.327 2.39**
Financial Leverage
(D/A)
0.278 4.08***
0.282 1.88*
0.438 4 IT***
Size (Log of sales)
0.780 31.15***
0.758 19 95***
0.726 28.28***
Risk (Std. Dev.)
0.131 2.06**
0.146 1.96**
0.246 1.99**
Growth
0.851 11 29***
0.615 6.12***
0.676 4 go***
0.012 2.45**
Board Size
0.015 4.46***
0.019 4 78***
0.632 6.92***
0.178 1.72*
0.471 2 io***
0.421 4 0 3 * * *
0.472 2 "7Q***
Busy Directors
0.626 g Q1 ***
Grey Directors
0.461 2 07***
0.278 1.88*
0.198 2.03**
-0.254 -1.68*
Fam. Directors
-0.261 -2.34**
-0.533 -2.41**
-0.054 -0.63
-0.120 -1.26
-0.309 -2.14**
Institutional Ownership
Executive Positions
CEO & Chairman
0.854 15.44***
1.039 11.67***
1.171 18 91***
CEO Only
0.711 12 94***
0.828 o 67***
0.903 15.56***
0.474 4.80***
Chair Only
0.369 5.30***
0.645 8.86***
0.024 0.26
-0.094 -1.59
0.036 0.67
CFO
r Qc***
0.270 2 03***
0.170 2.80***
0.359
COO
VP
-0.202 -3 80***
-0.013 -0.27
0.009 0.11
4.003 15.80***
Intercept
3.990 26.46***
4.083 30.52***
Yes
Yes
Yes
Industry & Year Effects
0.483
Adjusted R2
0.469
0.475
3332
3889
3448
Obs.
97
Figure 1: Median total compensation of the top three highest paid executives over the period 2001-2007
4000i
35003000-
Dual Class Firms
Single Class Firms
2500\
$000 2000-|
15001000500-
0^
2001
2002
2003
B Dual Class
2004
2005
D Single Class
98
2006
2007
Figure 2: Average discount of dual class firms with both classes of shares trading over the period 1998-2007
4 --
-4
-5.10
v yuA
Mean
VT-cnr
I!
-8
-12 1998
1999
2000
2001
2002
2003
Year
2004
2005
2006
2007
Note: The discount is calculated using monthly data. The mean discount over the period 1998-2007 is -5.10% and statistically significant
from zero. This is consistent with the findings of Zingales (1995). The definition for share discount = ^ 51,
" ' ' ' , where r is the ratio of
votes per restricted voting shares divided by votes per superior voting shares and PRV is the price of restricted voting shares and PSV is the
price of superior voting shares. This is a modified version of Zingales (1995) voting premium.
99
8.0
Appendix A:
Table 1A: Variable description and measurement

Variable
Label
Measurement
Salary
Salary
The dollar value of the base salary earned by the named

executive officer.
Annual Bonuses
Bonuses
The dollar value of bonuses earned by the named

executive officer.
Stock Option Grants
Stock
Options
Executive's stock options compensation valued using

the Black-Scholes model.
Other compensation
Other
Other compensation received by the executive including

perquisites and other personal benefits, termination or
change-in-control payments, contributions to defined
contribution plans, life insurance premiums, gross-ups
and other tax reimbursements, discounted share
purchases, value of restricted stock grants and long term
incentive payouts.
Total compensation
Total
Sum of salary, bonuses, stock options and other

compensation (Execucomp- TDC1).
Compensation
Management Voting
Leverage
Mgmt. vote
The percentage of total votes controlled by the

management and directors divided by the percentage of
total equity stake owned by management and directors
as a group.
Board size
Boardsize
Number of directors on the board.
Independent directors
Independent
directors
Number of unrelated outside directors divided by total

number of directors.
Grey outside directors
Grey
Number of outside directors who are related to the

company through transactional relationships divided by
total number of directors.
Busy directors
Busy
Number of directors with more than 4 board

memberships divided by total number of directors.
Family Directors
Fam. Dir
The percentage of family members on the board of

directors.
Institutional
Ownership
Inst, investors
Percentage of shares owned or controlled by

institutional investors.
Performance
Annual return on assets as measured by EBIT/total

assets at fiscal year end.
Firm performance
100
Table 1A Cont'd
Variable
Label
Measurement
Firm size
Financial
Leverage
Size
Firm risk
Risk
Growth
Growth
Total short-term and long-term debt divided by total

assets.
Natural logarithm of firm's sales.
Standard deviation of firm's annual stock returns over the
previous 5 years.
Geometric mean growth in total assets over the previous
five year period.
Cash and marketable securities.
Firms capital expenditure spending.
Financial leverage
Cash
Cash
Capital expenditure
Capex
Executive Characteristics
Chief Executive
CEOChair
Officer & Chairman
Chairman Only
Chair only
Chief Executive
Officer Only
CEO Only
Chief Financial Officer
CFO
Chief Operating
Officer
COO
Vice President
VP
Other Executives
Other
Family CEO &

Chairman
Fam.
CEOChair
Family Chairman Only
Fam. Chair
only
Fam. CEO
Family Chief
Executive Officer Only Only
Family Chief Financial
Fam. CFO
Officer
Family Chief
Fam. COO
Operating Officer
Family Vice President
Fam. VP
Other Family
Executives
Fam. Other
Indicator variable =1 if the executive is the CEO and is

also the Chairman of the board and 0 otherwise
Binary variable equal to 1 if the individual is the
Chairman of the board and 0 otherwise.
Indicator variable =1 if the executive is the CEO only
and 0 otherwise
Indicator variable =1 if the executive is the CFO and 0
otherwise
Indicator variable=l if the executive is the COO and 0
otherwise
Indicator variable =1 if the executive is the VP and 0
otherwise
Binary variable =1 if the executive in not CEO,
Chairman, CFO COO or VP and zero otherwise. For
example, President of a subsidiary is classified as an
other executive.
Indicator variable =1 if the executive is the CEO is also
the Chairman of the board and a family member and 0
otherwise
Binary variable equal to 1 if the individual is the
Chairman of the board and a family member and 0
otherwise
Indicator variable equal to 1 if the CEO is the founder or
a related family member and 0 otherwise
Indicator variable equal to 1 if the CFO is the founder or
a related family member and 0 otherwise.
Indicator variable equal to 1 if the COO is a family
member and 0 otherwise.
Indicator variable equal to 1 if the VP is a family member
and 0 otherwise.
Indicator variable equal to 1 if the executive is a family
member and classified as an other executive and zero
otherwise.
101
Chapter 3
Ownership Structure, Agency Problems and Dividend Policy
1.0
Introduction
This research investigates the effects of different ownership structures on corporate
distribution of wealth when companies are closely controlled. The sample of companies
studied is drawn from the S&P 1500 and the aim is to examine whether dividends are higher
in firms with dual class ownership structure relative to firms with a single class concentrated
ownership structure.40 There are generally two agency issues associated with concentrated
ownership and control. The separation of ownership and control gives rise to classical
principal-agent problems between managers and shareholders. This type of agency problem
can be mitigated through an increase in ownership and monitoring by large shareholders
with a significant equity stake in the company (Jensen and Meckling, 1976). However this
in turn, gives rise to a different type of agency problem. Concentrated ownership and control
in both single class and dual class companies can result in conflicts between large majority
and minority shareholders.41 The agency costs associated with dual class ownership
structure are higher than those of single class concentrated control firms due to the
additional monitoring costs associated with two groups of non-controlling shareholders:
Following LaPorta et al., (1999) and Claessen et al., (2000), a single class firm with concentrated ownership
is defined as a firm with ownership of 15% of the shares outstanding by an individual, a family or an institution.
41
Minority shareholders are defined as a group of dispersed shareholders who each hold a small proportion of
votes in the firm.
102
minority shareholders who hold superior voting shares and minority shareholders who hold
restricted voting shares.
Agency costs imposed by controlling shareholders with large voting rights and small
equity interest (dual class firms) can be larger than those imposed by controlling
shareholders who hold a majority of the cash flow rights (single class with concentrated
ownership).4 As the size of cash flow rights decrease, the agency costs increase however,
not linearly, but rather at a sharply increasing rate (Bebchuk et al., 2000). In dual class
firms, the controlling shareholder receives only a fraction of the corporate distribution, but
extracts the full private benefits from assets left in the firm (Bebchuk et al., 2000).
Therefore, there is a desire to institute lower distribution policies in order to retain assets
which are extracted as private benefits. In comparison, controlling shareholders of single
class firms will receive a larger fraction of the corporate distribution to shareholders because
of their significant equity ownership. Hence, they have an incentive to distribute more cash
flow than their dual class counterparts.
Dual class ownership structure can also lead to weaker alignment of interests among
shareholders compared to single class companies with concentrated ownership due to the
divergence of voting and cash flow rights. This can lead to potential expropriation of
minority shareholders (Grossman and Hart, 1988). Concentrated control in the hands of a
single shareholder gives him/her the opportunity to extract private benefits both in dual class
and single class companies at the expense of minority shareholders. However, this may be
more pronounced in dual class firms where large shareholders control a significant
A controlling shareholder is defined as an owner with a certain percentage of total votes which normally
allows them to have de facto control due to the difficulty of organizing dispersed shareholders. Hence, I focus
on the agency problems of concentrated ownership and control.
103
proportion of the voting rights relative to the small proportion of cash flow rights they own
(Correia da Silva et al., 2004).
The main objective of this research is to examine the agency problems associated
with concentrated ownership and control. In particular, I examine dividend policy of dual
class companies compared to single class companies with concentrated control. I attempt to
provide evidence in support of one of the three competing explanations for dividend policy
in a sample of firms with concentrated ownership and control. This research intends to test
which of the three hypotheses hold: the rent extraction/private benefits of control
explanation, the family legacy explanation and the managerial reputation explanation for
corporate distribution. The rent extraction/private benefit hypothesis predicts that
executives and controlling shareholders in dual class firms will set a low dividend payout
policy in order to retain and use firms' resources for their own benefits. The separation of
voting from cash flow rights allows a controlling shareholder to extract private benefits
without facing the proportionate cash flow consequences that a similar controlling
shareholder with significant equity ownership in a single class firm would face.
Alternatively, in firms with family control, the family reputation, identity and wealth are
attached to the firm or a group of related companies. Therefore, in order to ensure
intergenerational transfer of wealth, control and the family legacy to their heirs,
controlling shareholders may set a low dividend payout policy. The resources retained
within the firm are then used to grow the wealth in the firm.
The managerial reputation explanation for dividend policy in dual class firms
states that investors know that managers of dual class firms may expropriate resources
from the firm and as a result, investors will discount the value of dual class firms. In
104
order to entice investors to hold restricted voting shares and alleviate concerns about
expropriation, managers will set a high payout policy. Hence, managers subject the firm
to raise capital more frequently and therefore, subject the firm to scrutiny from
investment banks and the capital markets in general (Rozeff, 1982 and Easterbrook,
1984). Hence, managers establish a reputation of limited expropriation.
Using a propensity score matched sample of dual and single class concentrated
control firms from the S&P 1500 group of companies, the findings are consistent with the
extraction of private benefits hypothesis. Dual class firms, on average, tend to retain more
wealth within the firm compared to single class companies. I find that cash dividend and
total distribution scaled by market capitalization (dividend yield), earnings (dividend
payout) and operating cash flow are negatively related to the divergence of voting and
cash flow rights. The evidence indicates that the greater the excess voting rights relative
to cash flow rights, the more wealth is retained within the company which can be used to
pursue private benefits or grow the firm for the future generation of family members. In
order to separate the two explanations, I identify firms with controlling shareholdersmanagers and examine excess compensation of these executives. I find evidence
consistent with the extraction of private benefits hypothesis. Excess total compensation
paid to controlling shareholders-managers is negatively associated with cash dividend
and total distribution. This implies that controlling shareholders-managers retain
corporate wealth which is then extracted as excess compensation.
105
2.0
Literature Review
According to Miller and Modigliani (1961), in a perfect world where there are no
taxes, information asymmetry or agency costs, dividend policy does not matter. However, in
most countries there are taxes on dividends as well as capital gains. In most countries,
capital gains are taxed at a lower rate than dividends. Therefore, investors should prefer
capital gains over dividends and hence, companies should not be motivated to pay dividends
on this basis. However, firms continue to initiate and pay dividends. This behavior by firms
has lead to the "dividend puzzle" (Black, 1976). As a result, several theoretical and
empirical studies have attempted to explain why some firms pay dividends and why others
choose not to pay dividends. Numerous studies have investigated this issue in the context of
taxation and argue in favour of a tax clientele explanation for dividend payments including:
Miller and Modigliani (1961), Brennan (1970), Elton and Gruber (1970), Lewellen et al.
(1978), Litzenberger and Ramaswamy (1979), Porterba and Summer (1984) and Masulis
and Trueman (1988).
A second group of studies try to explain dividend policy using
information asymmetry arguments including: Fama et al. (1969), Ross (1977), Bhattacharya
(1979), Kalay (1980), Aharony and Swary (1980), Miller and Rock (1989), Asquith and
Mullins (1983) and DeAngelo and DeAngelo (1990). Another group of studies provide a
behavioural explanation for dividend policy. These include: Shefrin and Statman (1984) and
Frankfurter and Lane (1984).44
For example, Black and Scholes (1974) did not find any support for the tax effects on dividend policy
whereas Litzenberger and Ramaswamy (1979) provided evidence relating dividends to pretax returns. Capital
gains are more highly valued than cash dividends (Porterba and Summers, 1984) because of differential
taxation of the two equity returns.
44
For a comprehensive review of the literature on dividend policy see Frankfurter and Wood (2000).
106
In addition, recent studies investigate the agency explanation for dividend policy (La
Porta et al., 2000) while others examine dividend policy in a concentrated ownership setting
(Faccio et al, 2001; Gugler and Yurtoglu, 2003; Chen et al., 2005 and Mancinelli and
Ozkan, 2006). In this paper, I used dividend policy to examine agency problems and the
extraction of private benefits in two different ownership structures. I propose three
explanations, managerial reputation explanation, private benefits and family legacy, for
dividend policy in firms which are controlled by votes (dual class firms) compared to
companies which are controlled by significant equity ownership (single class firms with
concentrated ownership).
2.1
Agency Theory, Private Benefits and Dividend Policy
2.1.1
Private Benefits and Family Legacy Explanation for Dividend Policy
Outside the U.S., the Berle-Means (1932) notion of separation of ownership and
control is an exception and not the norm (La Porta et al., 1999). In fact, concentrated control
is very common in most countries, particularly in countries with weak investor protection.
Control can occur in the form of dual class ownership structures, pyramids and crossholdings. By having significant voting power, controlling shareholders have the ability to
influence the firm's decisions such as payout policy.
Testing the rent extraction property of dividends, Gugler and Yurtoglu (2003)
examine the conflict between large and small outside shareholders using high ownership
concentration in Germany. They provide evidence that the voting rights of the largest
shareholder significantly reduce payout ratios. Also, the divergence between cash flow and
107
voting rights significantly lowers payout ratio.
Similarly, Mancinelli and Ozkan (2006)
find that as the voting rights of the largest shareholder increase, firms make lower dividend
payments.46 They argue that the results support the argument that a higher level of
ownership concentration is associated with a higher probability of expropriation of outside
shareholders. To the extent that there are private benefits to the largest shareholder for
holding larger amounts of cash, lower payouts will increase the ability of the larger
shareholders to expropriate the outside minority shareholders. Therefore, when large owners
gain nearly full control of the firm, they prefer to generate private benefits of control that are
not shared by minority shareholders (Shleifer and Vishny, 1997).
Using a sample of Canadian firms, Eckbo and Verma (1994) develop and test a
model where managerial benefits from free cash flow, heterogeneous personal taxes and
information asymmetries give rise to internal shareholder conflicts over the dividend
decisions. The consensus dividend hypothesis implies that actual cash dividend distribution
will vary with relative voting power of shareholder groups having different preferences for
dividends.47 The model predicts that the magnitude of cash dividend increases with
corporate or institutional share ownership whereas it decreases with voting power of
manager-owners and it is almost zero when managers-owners have absolute voting control.
The empirical evidence supports the above predictions.
The deviation of voting and cash flow rights occurs through pyramids and cross-holdings. The voting rights
of the second largest shareholder have a positive influence on payout. This implies that the second largest
shareholder acts as a monitor of the largest shareholder. However, Mancinelli and Ozkan (2006) did not find
any relationship between the voting rights of the second largest shareholder and dividend payout.
46
They studied the relationship between dividend policy and ownership structure of 139 Italian-listed closelyheld companies.
47
According to the dividend consensus hypothesis, actual dividend polices represent a compromise solution
where the interests of various heterogeneous shareholder groups are represented by the groups' voting power.
The interests of the manager-owner are pitched against the interests of the large corporate/institutional
shareholders.
108
These studies interpret lower dividend payout ratios as evidence of rent extraction or
expropriation of minority shareholders. However, low dividend payout ratio can also be
consistent with the idea that controlling shareholders retain wealth in order to grow the firms
for intergenerational transfer of wealth to heirs or other family members. In addition, if
ownership structures like dual class are costly, why do investors voluntarily become
minority shareholders of dual class firms, especially in countries with weak legal protection
for minority shareholders? Initially, investors may have been induced by higher dividends to
invest in these firms. Investors may be compensated for holding restricted voting shares
with higher dividends relative to dividends paid to superior voting shareholders. Also,
capital gains are another potential reason why investors may hold restricted voting shares in
dual class companies. In addition, legal protection in the U.S. may be stringent enough to
limit expropriation or extraction of private benefits and hence, investors are more willing to
hold restricted voting shares in this market.
Examining dividend policy in family firms, Chen et al., (2005) provide evidence of a
negative relationship between family ownership (up to 10%) and dividend policy and a
positive relationship between ownership (10%-35%) and dividend yield for small
capitalization firms in Hong Kong.49 They interpret their findings as an indication that
controlling shareholders are using dividends as a way to extract resources because dividends
The evidence that dual class share structure is detrimental to minority shareholders is somewhat mixed. For
example, Partch (1987) compares managerial ownership before and after the creation of a class of limited
voting common stock for 44 publicly traded firms between 1962 and 1984, and examines whether the event
affects the wealth of current shareholders. He argues that there is no evidence that current shareholders are
harmed by the creation of limited voting common stock. Similar evidence is provided by Ang and Megginson
(1989) and Cornett and Vetsuypens (1989). On the other hand, Jarrell and Poulsen (1988) find significant,
negative abnormal stock price returns at the announcement of the dual-class recapitalization. In the case of
Canada, Jog and Riding (1986) provide similar evidence.
49
Chen et al. (2005) examine ownership concentration as the fraction of total company shares outstanding held
by the controlling family. They investigate the relationship between ownership and dividend policy for a group
of family-controlled firms in Hong Kong. They did not examine how the divergence of voting and cash flow
rights affect dividend policy.
109
make up a disproportionately large part of the income they derive.
They also argue that
small firms are subject to less scrutiny by investors and these firms may be using dividend
policy as a means to extract resources. Alternatively, their results are consistent with the
argument that outside investors anticipate potential expropriation and therefore, demand
higher payouts from firms with potentially the largest agency conflict. I can distinguish
between these alternative explanations by examining dual class companies compared to
single class concentrated control companies. Since controlling shareholders in dual class
firms attain control using votes and not cash flow rights, evidence of higher dividend
payouts would be consistent with the argument that outside investors anticipate potential
expropriation and hence, demand higher payouts.
2.1.2 Managerial Reputation Explanation for Dividend Policy
According to La Porta et al. (2000), the agency approach does not rely on the
assumptions of Miller and Modigliani (1961) when explaining dividend policies. First, the
investment policy of firms cannot be viewed as independent from the firm's dividend
policy. Payouts can reduce cash flow available to invest in poor NPV projects. Second, the
allocation of profits to all shareholders on a pro rata basis cannot be taken for granted. It
does not allow for the possible diversion of resources by insiders at the expense of minority
shareholders. Therefore, dividend payments can be seen as a mechanism to reduce agency
costs. Dividend payments help to alleviate agency conflicts between managers and
shareholders because paying dividends and subsequently raising funds in the capital markets
50
In fact, an executive in position of the CEO and Chairman received 14 times the dividend income relative to
their cash salary. Also, the average CEO who was not the Chairman received 4 times the dividend income
relative to their cash salary.
110
serve as a disciplinary mechanism (Rozeff, 1982 and Easterbrook, 1984)5I By raising

capital, firms are subjected to scrutiny by investment banks, regulators and the capital
market as a whole. Born and Rimbey (1993) and Moh'd, Perry and Rimbey (1995) provide
evidence to support Rozeff (1982) and Easterbrook (1984) hypotheses. Also, Jensen (1986)
argues that higher dividend payments reduce agency costs of "free cash flow" by preventing
managers from using excess cash to undertake low return projects or "pet" projects which
benefit managers rather than shareholders.
Examining firms from 33 countries, La Porta et al., (2000) test two competing
agency explanations of dividend policy, the outcome model versus the substitution model of
dividends explanation.
They provide evidence to support the outcome agency model of
dividends. In particular, they show that firms in common law countries have higher
payouts.
They argue that investors in countries with good legal protection use their legal
power to extract dividends from firms especially when investment opportunities are poor.
However, it is not clear whether the outcome or substitution model of dividend policy is
dominant in firms with ownership structures that allow for the expropriation of minority
shareholders. Shareholders can use their legal power to force firms to pay dividends thereby
51
Rozeff (1982) argues that dividend payments are part of the firm's optimal monitoring/bonding package and
they serve to reduce agency costs This implies that firms will choose a level of dividend to minimize agency
and transactions costs
52
The first explanation views dividend policy as an outcome of legal protection Shareholders use their legal
power to force firms to pay dividends thereby disgorging any excess cash This can be achieved by voting for
directors who offer better dividend policies, selling shares to potential hostile raiders or by suing companies that
spend too lavishly on activities that benefit insiders only The second explanation argues that dividends can act
as a substitute for legal protection According to the substitute model, insiders interested in issuing equity in the
future pay dividends to establish a reputation for decent treatment of minority shareholders Therefore,
dividends play an important role in establishing a reputation of limited expropriation La Porta et al, (2000)
provide several examples where shareholders have successfully forced firms to pay dividends such as Chrysler
and Velcro Industries
53
Countries with a common law origin tend to have better legal protection of minority shareholders
111
disgorging any excess cash. This, in theory, can effectively limit the extraction of private
benefits.
Furthermore, Faccio et al. (2001), using cross-country data from Europe and Asia,
show that investors appear to be more conscious of expropriation within tightly controlled
pyramidal groups.54 To offset investor's concerns, higher dividends are paid by corporations
affiliated with such groups especially those exhibiting a wider discrepancy between
ownership and control. This is consistent with the managerial reputation explanation for
dividend policy in firms with concentrated control. On the other hand, they provide evidence
that investors seem to be less conscious of expropriation in corporations that are loosely
affiliated within groups in a pyramid structure (control links exceed 10% but not 20%).
In a corporation with low cash flow to control rights, dividend payments are
expected to be low since controlling shareholders seek to retain corporate resources (Faccio
et al., 2001). However, a rational investor, perceiving risk of expropriation, will attach a
lower value to the shares of these corporations and will be less willing to supply capital.
Dividend policy can address this concern. Firms with low ownership rights relative to
control rights can refrain from expropriation by committing to a high dividend policy and
hence, reduce cash flow that may be potentially expropriated.
2.2
Why study dual class share structures of firms in the U.S.?
Dual class structures are more complex compared to single class firms in the U.S. It
is also a more convenient way of transferring corporate control to heirs of controlling
54
Control occurs mainly through group affiliation in the form of pyramids, cross-holdings and reciprocal
holdings. In their sample, using the 10% ownership cut off, 56% of the European firms are controlled by
families compared to 45% in Asia. Also, 15% has no controlling shareholder in Europe whereas 20% are
widely held in Asia.
112
shareholders with a small proportion of wealth invested compared to single class companies
with concentrated ownership. In addition, the protection of minority shareholder rights in the
U.S. is much greater than several European and Asian countries where concentrated control
is more prevalent (La Porta et al., 1998). Hence, expropriation or rent extraction by the
controlling shareholder may be more difficult in the U.S. Therefore, the costs of
expropriation may be much higher in the U.S. and hence, using dividend policy to
expropriate shareholder wealth may not be as common. In addition, the SEC disclosure
requirements ensure that information relating to voting rights, accounting information and
executive compensation is properly disclosed and easily accessible to investors. This makes
controlling mechanisms, such as multiple voting share structures, more visible compared to
pyramid structures with several complex layers which are used to the control company.
3.0
Hypotheses, Methodology and Data
3.1.0
Hypotheses
Several studies argue that dual class ownership structure fosters managerial
entrenchment and that controlling shareholders have incentives and opportunities to extract
private benefits of control. Dual class share structure allows for separation of voting and
cash flow rights. This separation enables the controlling shareholder to make decisions that
provide them with private benefits without facing the proportionate cash flow consequences
that they would in a single class firm. The divergence of cash flow and voting rights leads to
lower accountability which is consistent with entrenchment (Harris and Raviv, 1988).
113
Therefore, controlling shareholders may set low dividend payouts and extract resources in
other ways such as excess compensation. This argument leads to the hypothesis below.
HI:
Dividend payout for dual class firms with a high degree of concentrated
control will have a lower payout policy compared to single class firms.
Alternatively, dual class firms serve to retain the "family legacy". Accordingly, dual
class controlling shareholders may retain resources in order to grow the company since the
wealth of several generations is tied up in the company. As a result, founding families
represent a special class of large shareholders that may potentially have unique incentive
structures and a strong voice (Anderson et al., 2003). Also, since dual class firms tend to be
family-controlled firms, family integrity and strong feelings of identity may result in family
managers seeking to achieve the firm's goal over their individual goals (Davis, Schoorman
and Donaldson, 1997 and Corbetta and Salvato, 2004). The firm may be passed on to the
family's next generation and therefore, it is important to pursue long-term oriented business
strategies such as, investing in R&D. Firm survival is important since a founding family
may view the firm as an asset to bequest to family members and their dependents rather than
consuming wealth (Casson, 1999; and Anderson et al., 2003). As a result, controlling
shareholders in dual class firms are more likely to maximize firm value and therefore, are
less likely to extract wealth from the firm for their own benefits. Therefore, firms that are
family controlled may retain resources in order to grow the firm by undertaking long-term
value enhancing projects that are beneficial to several generations of family members.
A competing explanation is the managerial reputation explanation. Investors,
knowing that controlling shareholders have the ability to extract private benefits, will tend to
discount the value of dual class firms relative to the value of single class firms with
concentrated ownership. Therefore, in order to entice investors to hold restricted voting
114
shares and alleviate investors' concern about expropriation, controlling shareholders can
commit to a high dividend payout policy. By paying higher dividends, managers commit the
firm to raise capital more frequently and hence, the firm is subjected to increased scrutiny by
investment professionals, investors and the capital market (Rozeff, 1982 and Easterbrook,
1984). Hence, dual class firms are more likely to pay dividends than single class firms and
payout larger dividends (Francis et al., 2005).
To distinguish between the private benefits hypothesis and the family legacy
explanation for low dividend payout, dual class firms will be sorted into different groups.
Firms with controlling shareholders as executives will be classified into one category and
those with controlling shareholders who are not executives of the firm but sit on the board of
directors, will be classified into another group. This distinction between the two groups of
controlling shareholders allow us to separate controlling shareholders who may potentially
extract private benefits from those who may find it difficult to extract private benefits in the
form of excess compensation because they are not executives. Using an indicator variable to
identify firms with executive-controlling shareholders, I am able to test the family legacy
and rent extraction hypotheses. Excess executive compensation is one potential channel
through which controlling shareholders and managers can extract private benefits (Masulis
et al., 2009). If controlling shareholders-executives set a low payout policy in order to
extract private benefits, I expect executive compensation in those firms to be greater than
those in dual class firms with non-executive controlling shareholders.
H2:
Dual class firms with controlling shareholders as executives will have a
higher level of executive compensation relative to dual class and single class firms
where the controlling shareholder is not apart of the executive team.
115
3.2.0
Methodology
I will examine the relationship between concentrated control and dividend policy
after controlling for relevant cross-sectional factors using several econometric techniques.
These include: (i) panel regression controlling industry and year fixed effects and (ii) Tobit
regression which is suitable as an econometric technique since dividend distribution is
censored from below at zero. The regression equation is specified as follows:
Payout ratio lt =fi0+\i]Mgmt.Vote
lt_\+y'X
+ slt
(1)
The dependent variable in equation (1) is corporate distribution to common

shareholders over the fiscal year t. Based on prior literature, the vector X comprises of firm
characteristics which have been proven to influence dividend payout ratio. These variables
are size (log of sales), financial leverage (total debt-to-total asset), growth (five year
geometric mean growth in total assets), capital rationing dummy, performance (return on
assets), and institutional ownership (percent of equity held by institutional investors) and ,,
is the error term with the usual white noise assumptions.
I measure corporate distribution to common shareholders as follows: cash
distribution is defined as cash dividends declared and paid to common shareholders
including special cash dividends and total distribution is defined as:
Total Distribution = cash dividends + share repurchases
(2)
Stock repurchase is calculated as common and preferred stock repurchases adjusted

for any decreases in preferred stocks (Stephens and Weisbach, 1998 and Banyi et al.,
116
2008).55 Both measures of corporate distribution are scaled by operating cash flow, after
taxed earnings (dividend payout) and market capitalization (dividend yield). The primary
reason for using several ratios is to insulate our overall findings from potential biases that
may arise due to accounting flexibility or manipulations.
To test whether divergence of voting and cash flow rights influence corporate
distribution, I construct a voting leverage ratio as follows: Management voting leverage
,,
% of total votes controlled by management and directors T_ ,

rr
(Mgmt Vote) =
. If the private
% of total equity held by management and directors
benefits/rent extraction hypothesis is correct, then I expect management voting leverage to
be negatively related to dividend payout. However a negative relationship is also consistent
with the preservation of family wealth for future generations. If the rational expectation
hypothesis is correct, then management voting leverage will be positively related to payout.
In terms of the control variables, debt can be used to curb agency costs, and expropriation,
by limiting the amount of resources that can be paid as dividends (Jensen, 1986 and Francis
et al., 2005). Therefore, financial leverage (debt-to-asset) is expected to be negatively related
to dividend payout. In addition, institutional investors can act as a monitor and hence, reduce
the need for high payout as a disciplinary mechanism. However, institutional investors can
also force managers to pay out more of the corporate resources as dividends and raise capital
in the external capital markets. Therefore, the effects of institutional investors on dividend
payout can be positive or negative.
55
Banyi et al (2008) find that Compustat purchase of common and preferred stocks (item #115) minus changes
in the value of preferred stocks is the best estimate of actual repurchases This measure is usually not offset by
either concurrent sales of stocks through equity offering or exercise of employee stock options unlike the CRSP
measure of monthly decline m shares outstanding (alternative approach to estimate share repurchases)
117
Following Farinha (2003), I define growth as the geometric mean growth in total
assets over the previous five year period.56 It is expected to be negatively related to dividend
payout. In addition, firms facing difficulties in raising capital in the external market may
limit their payouts. Therefore, I have to control for the effects of this potential capital
rationing behavior by firms on dividend payout. As a result, I construct a capital rationing
dummy variable. Following Faccio et al. (2001), I compute the average increase in capital
stock (excludes reserves and retained earnings) plus financial debt as a ratio of sales. I then
take a five year average of this ratio in order to smooth capital rationing which may be due
to transitory factors such as, the business cycle. The capital rationing dummy variable is
equal to 1 if the ratio is below the sample median and the company's growth rate is above
the sample median, otherwise it is set equal to 0. It is expected to be negatively related to
dividend payout.
Business risk or fixed operating costs may affect corporate distribution. As a result,
firms will retain more of their earnings since retained earning is the least costly method of
financing. An increase in risk profile of the firm can result in an increase in financing costs.
Therefore, risk (beta) is expected to be negatively associated with dividend payout. In
addition, larger and more profitable firms are more likely to pay higher dividends and thus
retain less cash within the firm. Hence, profitability (ROA) and size (log of sales) are
expected to be positively related to dividend payout.
In order to distinguish between the extraction of private benefits hypothesis and the
family legacy hypotheses, I estimate equation (3). Following Zingales (1995) and Masulis et
al. (2009), I estimate equation (3) using firm characteristics and governance variables that
56
1 replace asset growth with growth rate in sales as a measure for the firm's growth rate and the results are
similar.
118
have been proven to explain CEO executive compensation and extract the residuals as a
measure of excess CEO total compensation. I then use the excess compensation from
equation (3) to test the extraction of private benefits hypothesis by estimating equation (4)
below:
Ln(Compensation)
, = a + r\kFirm Characteristicsk
k=\
,_, + 8kGovernancek
, t_\
k=\
+ e
(3)
where Ln(Compensation) is the natural logarithm of CEO total compensation for firm j at
year t. Firm characteristics include: size, profitability, growth, risk and financial leverage.
The governance variables are as follows: board size, percentage of independent, busy and
grey directors, percentage of institutional ownership, CEO tenure, CEO-Chairman duality
and percentage of family members on the board of directors.57
To test the second hypothesis that controlling shareholder-executives of dual class
firms extract excess compensation compared to single class closely-held firms, the
following Tobit regression specification is estimated:
Payout Ratio, t = P 0 + p, (EC, t x ExcessCompl t x Mgmt. Votel r_,) + P2 (EC,, x Mgmt. Vote,,_,)
+ p 3 (ExcessComplt x Mgmt. Vote,t_x) + p 4 (EC x ExcessComp), t
+ P5Mgmt. Vote,,_, +fi6Family Director, t +fi7EClt +p 8 Excess Comp.It
+ y ' x + 8l,/
(4)
where EC is a binary variable equal to 1 if the controlling shareholder is the CEO or

President of the firm and 0 otherwise, and EC x ExcessCompx Mgmt Vote is an interaction term
57
CEO-Chairman duality is a dummy variable equal to 1 if the CEO is also the chairman and zero otherwise.
119
which is used to test the second hypothesis. It is expected to be inversely related to dividend
payouts. Hence, if controlling shareholders-executives are extracting resources (excess
compensation) from the firm which they controlled with votes, I expect the dividend payout
ratio to be lower than in firms where the controlling shareholder is not an executive. In
addition, family director which is the percentage of family members on the board of
directors is included as a control variable. It is expected to be negative and significant if the
firm is retaining wealth to grow the firm and preserve wealth for future generations rather
than extracting resources (family legacy).
I include management voting leverage and
excess total compensation so that our interaction term does not merely pick up the effects of
excess voting rights as well as the influence of excess total compensation on dividend
payout. The dividend policy will alternatively be defined as dividend-to-market
capitalization (dividend yield), dividend-to-earnings (dividend payout) and dividend-to-cash
flow ratio. The vector, X contains a set of control variables as defined in equation (1).
3.3.0
Data
The dataset examined in this research is constructed using a variety of sources.

To construct a sample of U.S. dual class firms, I retrieve a list of firms with dual class share
structure from Corporate Library for 2005-2007. Corporate Library identifies all firms with
dual class common shares as a takeover defense mechanism during this period. I extend this
list of dual class firms to 2001 using a list of dual class IPOs and a list of dual class firms
58
1 replace family director with a binary variable equal to 1 if the firm is a family firm and zero otherwise. The
results are qualitatively the same.
120
used in Gompers et al. (2010). The list of 1,910 dual class firms over the period of 20012007 is merged with Execucomp database to determine whether compensation data is
available for these firms. Execucomp database contains executive compensation data for the
top executives representing the S&P 1500 group of companies. For each dual class company
with compensation data, I retrieve proxy statements from the Securities and Exchange
Commission (SEC) website and check the proxy statement for each firm in the sample to
ensure that they are in fact, dual class companies. Next, using proxy statements, I collect
voting rights per share and the number of superior voting and restricted voting shares owned
by the largest shareholders and management and directors as a group.
information such as cash dividends, purchase of common and preferred shares, total assets,
sales, long-term debt, common equity and operating income. In addition, I collected several
governance variables and equity ownership data from Corporate Library and Execucomp.
These include the number of directors, outside related directors and unrelated directors. I use
proxy statements, the firm website and internet search engines such as Lexus Nexus and
Google to identify family executives and family directors. I then calculate the percentage of
family members who are directors of the board. In order to complete the set of control
variables I collect stock return data from CRSP. I obtain monthly returns to estimate beta
and annual returns to compute standard deviation. Finally, I collect CEO total compensation
(TDC1) from Execucomp.
59
121

The list of dual class companies is matched with a list of single class firms with
concentrated ownership using propensity score matching. Propensity score matching
methods were developed in Rosenbaum and Rubin (1983), Heckman and Robb (1986) and
Heckman et al. (1998). Propensity score matching has become a popular matching technique
applied to studies of the financial markets (see Hillion and Vermaelen, 2004 and Villalonga,
2004) which allows for matching using a larger number of characteristics and hence,
reducing or eliminating potential bias. The greater the overlap in all characteristics of the
treated and control groups, the more comparable the groups are and the smaller the bias
(Heckman et al., 1997 and Heckman et al, 1998).
directors, grey directors, institutional ownership, company age, R&D-to-sales, capex-to-total
asset and family firms. This matching exercise results in a final sample of 792 dual class
firm-years over the period of 2001-2007. This represents an average of 113 dual class firms.
The final matched sample is made up of 1,584 firm-year observations.
122
4.0
Results
4.1.0 Descriptive Statistics
Table 1 summarizes the data used in the regression. The last column in Table 1
shows the test statistics for the difference in mean and median for the two samples. In the
dual class sample, the largest shareholder owns an average of 22.5% of the equity stake
compared to 57.8% of the voting rights. Management and directors as a group, control 58.3
% of the total votes compared to 24.9% of the equity stake. The disparity between voting
and cash flow rights is at the heart of the agency problems associated with dual class
ownership structure.
Table 1 shows that dual class firms pay significantly lower cash dividends and total
distribution as a ratio of market capitalization (dividend yield) or earnings (dividend payout)
compared to single class companies. For example, dual class firms pay 20.2% of their
earnings as cash dividends compared to single class companies with concentrated control
which pay 24.9% of their earnings to common shareholders. In terms of total distribution,
dual class companies seem to repurchase fewer shares compared to single class closely-held
firms with concentrated control. One possible explanation is that controlling shareholders of
dual class companies retain resources in order to extract private benefits.
One possible reason for holding restricted voting shares is the higher dividends paid
to those shareholders relative to superior voting shareholders. In the sample of dual class
firms, 13.5%) of these firms pay preferential dividends to shareholders of restricted voting
shares. This is a form of compensation for the lack of voting rights of restricted voting
123
shares. Typically, restricted voting shareholders receive at least 10% more in dividends paid
than superior voting shareholders.60
As for specific characteristics, Table 1 shows that the dual class sample and the
matching single class concentrated ownership sample are very similar in terms of size, risk
(beta), performance (ROA), financial leverage (D/A) and growth. The tests for mean
(median) difference are insignificant for these variables. The mean (median) size of dual
class firms in the sample is $4.9 billion ($1.5 billion) compared to $4.7 billion ($1.6 billion)
for single class companies. Similarly, the mean (median) return on assets for the dual class
sample is 9.68% (8.87%) compared to 9.57% (9.06%) for matching single class companies.
Based on the summary statistics, it appears that the matching procedure is reasonable since
the two samples are similar on several other dimensions such as risk, profitability, financial
leverage and growth.
Superior voting shares of dual class firms are often held by insiders and in most
cases superior voting shares are not traded. Hence, ownership of restricted voting shares by
institutions limits their voting power. Therefore, if voting rights are an important
determinant of ownership, then institutional investors are less likely to hold a significant
proportion of equity in dual class companies. In fact, institutional ownership is lower in dual
class companies with the mean holdings of 16.9% compared to mean ownership of 21.5% in
single class firms. The difference in mean (median) is negative and significant at the 1%
level.
Preferential dividends paid to restricted voting shareholders can range from 4% to 100% more than dividends
paid to superior voting shareholders. For example, Haverty Furniture in its proxy statement states that common
stock has a preferential dividend rate of at least 105% of the dividend paid on class A common stock (superior
voting shares). However, the actual difference in dividend paid is 8% for fiscal year 2007. Claire Stores Inc, in
fiscal year 2006, paid a total of $0.40 per share on our common stock and $0.20 per share on our Class A
common stock (superior voting shares).
124
Dual class firms are often family firms (83.2%) with several family members
serving as officers and directors. In fact, 15.8% of all directors in dual class firms are family
members compared to 5.6%) in similar single class companies. Since dual firms are
frequently managed by family members, one possible explanation for the high retention rate
in dual class firms is the family legacy explanation.
4.2.0 Regression Analysis
Table 2 presents the results of fixed effects and Tobit regression estimation for each
of the three measures of cash dividend on management voting leverage and several firm
specific factors. Management voting leverage is negatively related to cash dividend ratio in
all the regression specification. The coefficients are negative and significant at either the 1%
or 5% level except for cash dividend as a ratio operating cash flow which is significant at the
10% level in the Tobit specification. This finding is consistent with prior studies such as
Gugler and Yurtoglu (2003) and Mancinelli and Ozkan (2006).61 The results are consistent
with the extraction of private benefits hypothesis. The greater the divergence of voting and
cash flow rights, the lower the dividend payout. Controlling shareholders are retaining
wealth which can be used to pursue private benefits. However, the results provided in Table
2 are also consistent with the family legacy explanation. A large proportion of dual class
firms are family controlled and therefore, it is possible that controlling shareholders are
retaining wealth in order to preserve and grow the firm for future generations of family
members.
61
These studies use the voting rights of the largest shareholder as a measure of concentrated control. Deviation
of voting and cash flow rights predominantly occur through the use of pyramid structures or cross-holdings.
125
In terms of the firm specific factors, the sign of the coefficients are consistent with
prior studies. Several of the firm specific variables are significant across the various
measures for payout ratios. The capital rationing variable is negative and significant in all
three measures of cash dividend ratio. The evidence is consistent with previous studies, such
as Faccio et al. (2001). This implies that firms facing difficulties raising capital will tend to
retain more resources by limiting their payout. Similarly, I find that firms with higher
growth rates will pay lower dividends. This finding is consistent with the results provided by
prior studies, such as Farinha (2003) and Mancinelli and Ozkan (2006). Since internal
financing is the least costly form of financing, firms with growth opportunities are likely to
retain earnings in order to exploit future growth potential. In terms of risk, firms with higher
risk (beta) tend to pay lower dividends. This is consistent with conventional wisdom that
risk increases borrowing costs and therefore, firms tend to retain more earnings. The
negative and significant coefficient for the risk variable corroborates previous studies, such
as Farinha (2003) and Bhatacharyya et al. (2008). Other control variables including size (log
of sales), performance (ROA) and institutional ownership are significant.
In Table 3, I present regression results for total distribution on management voting
leverage and a set of firm specific control variables. In both the Tobit and Fixed effects
estimation, management voting leverage is negative and significant for all three measures of
total distribution. The results are also economically significant. Using estimate from the
Tobit regression, a one standard deviation change in management voting leverage resulted in
a 5.81% decrease in total distribution as a proportion of earnings. The evidence for total
distribution presented in Table 3 is consistent with both the extraction of private benefits and
126
the family legacy hypotheses. In addition, firm specific characteristics generally have the
expected sign.
In order to distinguish between the extraction of private benefits and family legacy
hypotheses, I first identify all firms with controlling shareholders as the CEO or President. If
the extraction of private benefits hypothesis is correct then I expect controlling shareholdersCEOs to extract private benefits in the form of excess executive compensation. In the
sample, 59.9% of dual class firms have a controlling shareholder-CEO. In comparison,
19.4% of single class firms have a controlling shareholder-CEO.
Panel A of Table 4 presents summary statistics of cash dividend and total
distribution (scaled by market capitalization, earnings and cash flow) for a combined sample
of dual and single class firms. It breaks the sample into firms with a controlling shareholderCEO versus those without a controlling shareholder-CEO. For both the cash dividend and
total distribution, firms with a controlling shareholder-CEO tend to pay out less cash than
firms without a controlling shareholder-CEO. The tests for difference in mean and median
are negative and significant. This implies that firms with a controlling shareholder-CEO
retain more wealth within the firm which can be used either to extract private benefits or to
grow the firm for future generations. To evaluate the latter explanation, we compare the
capital expenditure-to-sales or capital expenditure-to-asset ratios as proxies for growth
opportunities. Firms with controlling shareholder-CEOs invest less in capital expenditure as
a proportion of sales or total asset compared to firms without controlling shareholder-CEOs.
Thus, the family legacy explanation of low dividend payout is not supported.
Panels B and C of Table 4 decompose the sample further and show that the lowest
shareholder payouts are in dual class firms with controlling shareholder-CEOs. Panel B
127
shows that dual class firms with a controlling shareholder-CEO have lower payout ratios
than dual class firms without a controlling shareholder-CEO. Panel C shows that dual class
firms with a controlling shareholder-CEO have lower cash distributions than single class
firms with a controlling shareholder-CEO. Thus, controlling shareholder-CEOs in dual class
firms tend to distribute the least cash to their outside shareholders. As discussed below, we
find a link between this low payout and excess executive compensation of the dual class
controlling shareholder-CEOs.
Table 5 presents the Tobit regression results for cash dividends. The interaction
term, EC x ExcessComp x Mgmt. Vote, is negative and significant for all of the various
measures of cash dividend ratios. This evidence is consistent with the prediction of the
extraction of private benefits of control hypothesis. The results imply that the higher the
excess compensation in firms with controlling shareholders-executives and with greater
divergence of voting and cash flow rights the less cash dividends are distributed. These
executives institute lower payout policies in order to retain wealth to pursue private benefits
in the form of excess total compensation. Management voting leverage is negative and
significant for two of the three of the cash payout ratio measures (dividend yield and
dividend payout ratio). In addition, all of the firm specific variables generally have the
expected sign based on previous studies such as Faccio et al. (2001), Farinha (2003) and
Mancinelli and Ozkan (2006). For example, growth is negative and significant at the 1%
level for all three measures of cash payout ratio. Table 6 reports the results for total
distribution. The coefficient of the interaction term between controlling shareholdersexecutives and excess CEO total compensation is negative and significant for total
128
distribution-to-market capitalization and total distribution-to-earnings. This provides

additional support for the extraction of private benefits of control explanation.
To provide further evidence of the extraction of private benefits, I analyze a subsample of family firms only. If the family legacy hypothesis is correct, then controlling
shareholders-family executives are not expected to extract excess compensation in family
firms. However, if the extraction of private benefits hypothesis is correct, then controlling
shareholders-executives are expected to extract excess compensation and therefore distribute
less cash dividends or repurchase less shares. The results are presented in Table 7. Excess
compensation paid to controlling shareholders-executives is negative and significant for
cash dividends and total distribution. This implies that the higher the excess compensation
paid to family members who are executives, the lower the corporate distribution. This
finding is consistent with the prediction of the extraction of private benefits hypothesis for
dividend policy. Finally, I examine a sub-sample of dual-class family firms only. The
results presented in Table 8 are consistent with the extraction of private benefits explanation.
4.3.0
Robustness
Following Faccio et al. (2001), I computed industry adjusted cash dividends and
total distribution ratios. I first computed for each SIC industry, the median cash dividend or
total distribution scaled by market capitalization, earnings or cash flow. Then the
corporation's cash dividend and total distribution as a ratio of market capitalization, earning
or cash flow is the difference between the corporation's dividend or total distribution ratio
and the industry median ratio. In Table 9, I report the results using industry adjusted cash
dividend and total distribution ratios. The results are similar to those presented in Tables 2
129
and 3. As a further robustness check, I excluded all firms in the financial and utilities
industries.62 The results are similar to those presented in Tables 2 and 3.
4.4.0 Simultaneity and Two-Stage Least Square Estimation
The identification assumption is crucial to the causal interpretation of the findings
above which is that variation in the ownership structure and management voting leverage is
unrelated to unobserved factors which influence dividend policy. Eckbo and Verma (1994)
examine the possibility that the firm's dividend policy and ownership are determined
simultaneously. They find that dividend yield decreases with the voting power of ownersmanagers and it increases with the voting power of institutional shareholders but the
opposite is not true. They argue that the direction of the causality appears to run from voting
power to dividend policy. However, it is possible that some unobserved factors correlate
with changes in control structures that affect dividend policy. Hence, the regression
estimates are likely to be biased and inconsistent due to simultaneity. In order to address the
potential simultaneous processes determining dividend policy and ownership structure or
omitted variable bias, a two stage least square estimation technique is utilized. The first
stage is estimated using a panel regression.
Mgmt.Vote J t = a + \\S^Beta + \\)2Financial Leveraget + î^Qt "*" V\Namet + \\J5Mediat

+\\J 6 Sizet + \\i 7 payout + Sjt
(5)
where beta is estimated using the previous 5 years of monthly returns, financial
leverage(debt/total asset) and Q is the Tobin's Q ratio. Name is a binary variable equal to
Financial and utilities firms that belong to SIC 6000-6999 and 4900-4999 were excluded from the sample.
130
1 if the firm name is the same as an individual's name. Media is a dummy variable equal to
1 if the firm is in the media industry.64 Size is defined as the natural logarithm of sales.
Using the predicted management voting leverage from equation (5), I estimate
equation (6) using Tobit regression technique because cash dividend and total distribution
ratios are censored from below at zero.
Payout
ratiolt
= P 0 + $xPred.Mgmt.
Vote,,_, + y ' X + e,,
(6)
where Pred.Mgmt.Vote is the predicted management voting leverage from equation (5)
above and the vector X contains several control variables such as size, financial leverage,
growth, capital rationing dummy, performance and institutional ownership. The results are
reported in Table 10. The estimated coefficients for predicted management voting leverage
and control variables are similar to those presented in Tables 2 and 3.
5.0
It is argued that concentrated ownership and control have a significant impact on

agency costs. Ownership concentration can lead to alignment of interest of managers and
shareholders. Hence, there is less need for monitoring by shareholders and therefore, lower
agency costs. However, ownership concentration can also lead to different types of agency
problems such as, conflict of interest between controlling shareholders and minority
shareholders. In the case of dual class firms, the resulting monitoring costs may be greater
63
I used leverage, size, beta and Tobin's Q as determinants of management voting leverage following Lins
(2003) and Smith et al., (2009).
64
A firm is defined as a media firm if it belongs to SIC codes 2710-11, 2720-21, 2730-31, 4830, 4832-33,
4840-41, 7810, 7812 and 7820. Gompers et al., (2010) show that name and media variables are important
predictors of dual class ownership structure. Therefore, I include a name and media variable in estimating
equation 15.
131
than single class firms with concentrated control since there are two groups of minority
shareholders (minority shareholders who hold superior voting shares and minority
shareholders who hold restricted voting shares). In addition, a greater extraction of private
benefits is possible compared to single class companies with concentrated ownership due to
the divergence of voting and cash flow rights associated with dual class ownership structure.
I analyze a sample of dual class and single class S&P 1500 firms during the period
2001-2007 and provide evidence in support of one of the three explanations of dividend
policy in a concentrated control setting. First, the managerial reputation explanation states
that in order to alleviate concerns related to expropriation, dual class firms set a high payout
policy. By paying out higher dividends, managers commit the firm to raise capital more
frequently and hence, the firm is subjected to increased scrutiny by investment banks and
the capital markets in general (Rozeff, 1982 and Easterbrook, 1984). Second, the extraction
of private benefits hypothesis predicts that dual class ownership may foster managerial
entrenchment. The separation of voting and cash flow rights enables controlling
shareholders to undertake decisions which provide them with private benefits without facing
the proportionate cash flow consequences that they would have in a single class company
with concentrated control. Therefore, they set lower payout policies because they can extract
resources in other ways such as excess compensation. Third, the family legacy explanation
states that, since a large fraction of dual class firms are family controlled, controlling
shareholders may set a lower dividend policy in order to retain resources which are used to
grow the company for future generations of family members.
The empirical results show that dual class ownership structure is negatively
associated with dividend policy. The test of mean (median) difference indicates that dual
132
class firms pay less cash dividend and total distribution using various measures of cash
dividend and total distribution ratios (cash dividend and total distribution scaled market
capitalization, earnings and cash flows). This implies that the type of concentrated control
matters. The regression tests (Tobit estimation and panel regressions with industry and year
fixed effects) confirm my univariate findings. The greater the divergence between voting
and cash flow rights, the lower the cash dividend and total distribution. This is consistent
with the extraction of private benefits and the agency problems associated with dual class
share structure. It is also consistent with the family legacy hypothesis. Therefore, I examine
excess total compensation of controlling shareholders-executives in order to distinguish
between the two explanations for lower dividend policy and the results are consistent with
the extraction of private benefits hypothesis. Further, excess compensation paid to a
controlling shareholder-CEO in a firm with a larger divergence of voting rights relative to
cash flow rights is negatively associated with dividend policy.
133
6.0
References
Aharony, J., & Swary, I. (1980), Quarterly dividend and earnings announcements and
stockholders' returns: An empirical analysis, Journal of Finance, 35, 1-12
Anderson, R.C., Mansi, S.A., and Reeb, D.M., (2003), Founding family ownership and the
agency cost of debt, Journal of Financial Economics, 68, 263-285
Ang, J.S., and Megginson, W.L., (1989), Restricted voting shares, ownership structure and
market value of dual-class firms, Journal of Financial Research, 12, 301-318
Asquith, P. & Mullins, D. (1983), The impact of initiating dividend payments on
shareholders' wealth, Journal of Business, 56, 77-95
Baker, H.K., Powell, G., & Veit, E. T. (2002), Revisiting the dividend puzzle. Do all the
pieces now fit? Review of Financial Economics, 11, 241-262
Banyi, M. L., Dyl, E. D., and Kahle, K. M., (2008), Errors in estimating share repurchase,
Journal of Corporate Finance, 14, 460-474
Bebchuk L.A., Kraakman R., and Triantis G.G. (2000), Stock pyramids, cross-ownership
and dual-class equity: The mechanisms and agency costs of separating control from
cash flow rights. In Randall K. Morck (ed), Concentrated Corporate Ownership.
Chicago: The University of Chicago Press.
Berle, A. and Means, G. (1932), The modern corporation and private property, New York:
Macmillan.
Bhattacharya, S. (1979), Imperfect information, dividend policy and "the bird in hand"
fallacy, Bell Journal of Economics, 10, 259-270
Black, F., (1976), Dividend puzzle, Journal of Portfolio Management, 2, 5-8
Black, F., and Scholes, M., (1974), The effects of dividend yield and dividend policy on
common stock prices and returns, Journal of Financial Economics, 1, 1-22
Born, J. & Rimbey, J. (1993), A test of the Easterbrook hypothesis regarding dividend
payments and agency costs, Journal of Financial Research, 46, 251-60
Brennan, M. (1970), Taxes, market valuation and financial policy. National Tax Journal, 23,
417-429
Casson, M., (1999), The economics of the family firms, Scandinavian Economic History
Review, 47, 10-23
134
Chen, Z., Cheung, Y., Stouraitis, A. and Wong, A. (2005), Ownership concentration, firm
performance and dividend policy in Hong Kong, Pacific-Basin Finance Journal, 13,
431-449.
Cheung, Y., Stouraitis, A., and Wong, A.W.S. (2005), Ownership concentration and
executive compensation in closely held firms: Evidence from Hong Kong, Journal of
Empirical Finance, 12, 511-532
Cho, M., (1998), Ownership structure, investments and the corporate value: An empirical
analysis, Journal of Financial Economics, 47, 103-121
Constantinides, G. M , (1983), Capital market equilibrium
Econometrica, 51,611-36
with personal tax,
Corbetta, G., and Salvato, C , (2004), Self-serving or self-actualizing? Models of man and
agency costs in different types of family firms: A commentary on comparing the
agency costs of family and non-family: conceptual issues and exploratory evidence,
Entrepreneurship Theory and Practice, 28, 335-354
Cornett, M.M., and Vetsuypens, M.R., (1989), Voting rights and shareholder wealth: The
issuance of limited voting common stock, Managerial and Decision Economics, 10,
175-188
Correia da Silva, L., Goergen M., & Renneboog, L., (2004), Dividend policy and corporate
governance, New York: Oxford University Press Inc
Davis, J.H., Schoorman, F.D., and Donaldson, L. (1997), Toward a stewardship theory of
management, Academy of Management Review, 22, 20-47
DeAngelo H., DeAngelo, L., and Skinner, D.J., (1996), Reversal of fortune: Dividend
signaling and the disappearance of sustained earnings growth, Journal of Financial
DeAngelo, H., & DeAngelo, L., (1990), Dividend policy and financial distress: An
empirical investigation of troubled NYSE firms, Journal of Finance, 45, 1415-31
DeAngelo, H., and DeAngelo, L., (1985), Managerial ownership of voting rights: A study of
public corporations with dual-classes of common stock, Journal of Financial
Demsetz, H., (1983), The structure of the ownership and the theory of the firm, Journal of
Law and Economics, 26, 375-390
Demsetz, H., and Villalonga, B., (2001), Ownership structure and corporate performance,
135
Easterbrook, F., (1984), Two agency cost explanations of dividends, American Economic
Review, 74, 650-659
Eckbo, B.E. and Verma S., (1994), Managerial Shareownership, voting power and cash
dividend policy, Journal of Corporate Finance, 1, 33-62
Elton, E.J., & Gruber, M.J. (1970). Marginal stockholder tax rates and clientele effects,
Review of Economics and Statistics, 52, 68-74
Faccio, M., Lang, L.H., and Young L., (2001), Dividend and expropriation, American
Economic Review, 91, 54-78
Fama, E. F., Fisher, L., Jensen, M. C, and Roll R. W., (1969), The adjustments of stock
prices to new information, International Economic Review, 10, 1-21
Farinha, J., (2003), Dividend policy, corporate governance and the managerial entrenchment
hypothesis: An empirical Analysis, Journal of Business Finance and, Accounting, 30,
1173-1209
Francis, J., Schipper K. and Vincent L., (2005), Earnings and dividend informativeness
when cash flow rights are separated from voting rights, Journal of Accounting and
Frankfurter, G. M., & Wood Jr. B.G., (2002), Dividend policy theories and their empirical
tests, International Review of Financial Analysis, 11, 111-138
Frankfuter, G. M., and Lane, R.W., (1984), The rationality of dividends, International
Review of Financial Analysis, 1, 115-130
Grossman, S. and Hart, O., (1988), One share-one vote and the market for corporate control,
Gugler, K., and Yurtoglu B. B., (2003), Corporate governance and dividend pay-out policy
in Germany, European Economic Review, Al, 731-758
Harris, R., and Raviv, A., (1988), Corporate control contests and capital structure, Journal of
Financial Economics, 20, 203-235
Hermalin, B.E., and Weisbach, M. S., (1991), The effects of board composition and direct
incentives on firm performance, Financial Management, 21, 101-112
Holderness, C , and Sheehan, D.P., (1985), Raiders or saviors? The evidence on six
controversial investors, Journal of Financial Economics, 14, 555-579
Jarrell, G.A., and Poulsen, A., (1988), Dual-class recapitalization as antitakover
mechanisms: The recent evidence, Journal of Financial Economics, 20,129-152
136
Jensen, M. C, (1986), Agency cost of free cash flow, corporate finance and takeovers,
Jensen, M. C, and Meckling, W. H., (1976), Theory of the firm: Managerial behavior,
agency costs and ownership structure, Journal of Financial Economics, 3, 305-360
Jog, V. M., and Riding, A. L., (1986), Price effect of dual-class shares, Financial Analyst
Journal, 86, 58-66
Kalay, A., (1980), Signaling, information content and the reluctance to cut dividends,
Journal of Financial and Quantitative Analysis, 15, 855-67
Kowalewski, O., Stetsyuk, I., and Talavera, O., (2008), Does corporate governance
determine dividend payouts in Poland? Post-Communist Economies, 20, 203-218
Krivogorsky, V., (2006), Ownership, board structure, and performance in continental
Europe, The International Journal of Accounting, 41, 176-197
La Porta, R., Lopez-de-Silanes, F., Shleifer, A., and Vishny, R., (1998), Law and finance,
Journal of Political Economy, 106, 1113-55
La Porta, R., Lopez-de-Silanes, F., and Shleifer, A., (1999), Corporate ownership around the
world, Journal of Finance, 54, 471-517
La Porta, R., Lopez-de-Silanes, F., Shleifer, A., and Vishny, R., (2000), Agency problems
and dividend policies around the World, Journal of Finance, 55,1-33.
Lewellen, W., Stanley, K., Lease, R., & Schlarbaum, G. (1978). Some direct evidence of the
dividend clientele phenomenon, Journal of Finance, 33, 1385-1399
Litzenberger, R., and Ramaswamy, K., (1979), The effects of taxes and dividends on capital
asset prices: Theory and empirical evidence, Journal of Financial Economics, 7, 16395
Long, J., (1978), The market valuation of cash dividends: A case to consider, Journal of
Mancinelli, L., and Ozkan, A., (2006), Ownership structure and dividend policy: Evidence
from Italian firms, The European Journal of Finance, 12, 265-282
Masulis, R.W., and Trueman, B., (1988), Corporate investment and dividend decisions
under differential personal taxation, Journal of Financial and Quantitative Analysis,
23, 369-386
Masulis, R.W., Wang, C , and Xie, F., (2009), Agency problems at dual class companies,
137
Maury, B., (2006), Family ownership and firm performance: Empirical evidence from
Western European corporations, Journal of Corporate Finance, 12, 321-341
McConaughy, D. L., Walker, M. C , Henderson, G. V., and Mishra, C.S., (1998), Founding
family controlled firms: Efficiency and value, Review of Financial Economics, 7, 1-19
Mikkelson, W. H., and Partch, M. M., (1997), The decline of takeovers and disciplinary
managerial turnover, Journal of Financial Economics, 44, 205-228
Miller, M., and Modigliani, F., (1961), Dividend policy, growth and the valuation of shares,
Journal of Business, 34, 411- 433
Miller, M., and Rock, K., (1985), Dividend policy under asymmetric information, Journal of
Finance, 40, 1031-1051
Miller, M., and Scholes, M., (1978), Dividend and taxes, Journal of Financial Economics, 6,
333-364
Miller, M., (1986), Behavioral rationality in finance: The case of dividends, Journal of
Business, 59, S451-S468
Moh'd, M., Perry, L., and Rimbey, J. (1995), An investigation of the dynamic relationship
between agency theory and dividend policy, Financial Review, 30, 367-85
Morck, R., Shleifer, A., and Vishny, R., (1988), Management ownership and market
valuation: An empirical analysis, Journal of Financial Economics, 20, 293-315
Nenova, T., (2003), The value of corporate voting rights and controls: A cross-country
analysis, Journal of Financial Economics, 68,325-357
Partch, M., (1987), The creation of a class of limited voting common stock and shareholder
wealth, Journal of Financial Economics, 18,313-339.
Perez-Gonzalez, F., (2006), Inherited control and firm performance, America Economic
Review, 96, 1559-1588
Porterba, J., and Summers, L., (1984), New Evidence that taxes affect the valuation of
dividends, Journal of Finance, 39, 1379-415
Ross, S. A., (1977), The determinants of financial structure: The incentive-signalling
approach, Bell Journal of Economics, 8, 23-40
Rozeff, M. S., (1982), Growth, beta and agency costs as determinants of dividend payout
ratios, Journal of Financial Research, 5, 249-259
138
Shefrin, H. M., and Statman, M., (1984), Explaining investor preference for cash dividends,
Shleifer, A., and Vishny, R., (1997), A survey of corporate governance, Journal of Finance,
52, 737-784
Short, H., Zhang, H., and Keasey, K., (2002), The link between dividend policy and
institutional ownership, Journal of Corporate Finance, 8, 105-122
Stephens, C. P., and Weisbach, M. S., (1998), Actual share reacquisitions in open-market
repurchase programs, Journal of Finance, 53, 313-333
Stulz, R., (1988), Managerial control of voting right: Financing polices and the market for
corporate control, Journal of Financial Economics, 20, 25-54
Villalonga, B., and Amit, R., (2006), How do family ownership, control and management
affect firm value?, Journal of Financial Economics, 80, 385-417
Waller Jr., W.T., (1989), The concept of habit in economic analysis, Journal of Economic
Issue, 22, 113-126
Woolridge, J., (1983), Dividend changes and security prices, Journal of Finance, 38, 16071615
Zeckhouser R.J., and Pound, J., (1990), Are large shareholders effective monitors? An
investigation of share ownership and corporate performance. In R. Glenn Hubbard
(ed), Asymmetric information, corporate finance and investment. Chicago: The
University of Chicago Press.
Zingales, L., (1995). What determines the value of corporate votes? The Quarterly Journal
of Economics, 110, 1047-1073
139

Cash flow rights is defined as the percentage of equity owned by the largest shareholder or management and directors. Voting rights is the percentage of
votes held by the largest shareholder or management and directors. Management voting leverage (Mgmt. Vote) is the percentage of total votes
controlled by management and directors divided by the percentage of total equity held by management and directors, size (natural logarithm of sales),
financial leverage (total debt divided by total assets), performance (ROA=EBIT divided by total assets), risk (beta is estimated using the CRSP equally
weighted index and the previous five year monthly stock returns ), growth is the geometric mean growth in total assets over the previous five year
period, institutional ownership (percentage of shares held by institutional investors) and family director is the percentage of family members on the
board of directors. Cash dividend is cash dividends plus special cash dividends paid to all common shareholders. Total distribution is defined as cash
dividends plus share repurchases. Significance level at the 1%, 5% and 10% is indicated as *, **, and ***, respectively. Test for difference in mean is
the t-test and test for difference in median is the Wilcoxon signed-rank test.
Single Class
Dual Class
Mean
Median
Test for
difference
in Means
Test for
difference
in Medians
Std Dev.
Mean
Median
Std Dev.
16.50
25.70
15.60
25.20
2.15
23.60
23.60
17.10
17.10
1.00
19.20
19.20
16.50
16.50
1.00
13.20
13.20
14.10
14.10
0.00
-0.71
34 33***
14.21***
44.25***
58.60****
-0.62
15 75***
447***
15.82***
82.55***
1.77
8.32
38.00
79.17
45.24
69.62
1.31
4.31
24.92
71.62
17.60
55.00
0.74
1.95
7.23
34.07
4.49
19.10
3.31
12.39
58.65
99.45
40.87
94.15
-2.64***
-2.17**
-1.91*
-3.05***
0.70
-3 03***
-2.88***
-1.89*
0.66
_j 99**
-1.24
-3 34***
Ownership Characteristics
Cash Flow Rights -Largest Shareholder %
22.50
18.20
Voting Rights - Largest Shareholder %
57.80
54.80
Cash Flow Rights - Management & Directors %
24.90
19.30
Voting Rights - Management & Directors %
58.30
57.30
Management Voting Leverage (Mgmt. Vote)
2.42
3.01
Distribution Characteristics
Cash Dividend/Market Value %
0.56
0.95
Total Distribution/Market Value %
3.15
1.60
Cash Dividend/Earnings %
8.28
20.15
Total Distribution/Earnings %
57.74
27.67
Cash Dividend/Cash Flow %
3.24
19.12
Total Distribution/Cash Flow %
12.09
42.31
Preferential Dividend Treatment of Restricted Voting Shares
Firms with Preferential Treatment %
13.50
Dividend Paying Firms %
65.03
N/A
56.8
140
N/A
N/A
N/A
N/A
Table 1 Cont'd
Mean
Median
Std Dev.
Mean
Median
Std Dev.
Test for
difference
in Means
4917.45
21.61
12.61
9.75
0.99
9.98
16.92
15.80
1523.90
20.34
8.35
8.87
0.77
6.67
13.60
12.50
14597.23
18.10
34.87
8.81
0.81
15.61
15.66
12.73
4653.87
21.81
10.80
9.57
1.06
9.37
21.49
5.62
1629.00
20.82
7.45
9.06
0.85
6.46
18.71
0.00
15302.38
18.96
38.52
9.30
0.81
16.64
14.96
10.60
0.78
-0.58
2.18**
0.86
-1.46
1.07
-13 24***
34.83***
Dual Class
Size (Sales)
Performance- (Ret)%
Performance - (ROA)%
Risk-(Beta)
Growth % - (Total Asset)
Family Directors %
Single Class
Test for
difference
in Medians
-0.89
-0.28
1.58
-0.04
-1.63
1.08
-15.65***
37.86***
Note: The number of observations for the dual class sample is 792 firm-year and 792 firm-year for the single class closely-held sample. Earnings are
defined as earnings after taxes and interest expenses. Cash flow is equal to net operating cash flow.
141
Table 2: Regression of dividend payout on management voting leverage

Payout ratio,, = P 0 + fi^Mgmt .Fbte,,_, + y ' X + s,,
The dependent variable is cash dividends plus special cash dividends (Div.) paid to common
shareholders scaled by market value of equity (dividend yield), earnings (dividend payout) and
operating cash flow. Management voting leverage (Mgmt. Vote) is the percentage of total votes
controlled divided by the percentage of total equity held by management and directors. X is a vector
of control variables including: capital rationing (the capital rationing dummy variable is equal to 1 if
the average increase in capital stock plus financial debt as a ratio of sales is below the sample median
and the company's growth rate is above the sample median, otherwise it is set equal to 0),
performance (ROA=EBIT divided by total assets), growth is the geometric mean growth in total
assets over the previous five year period, risk (beta is estimated using the CRSP equally weighted
index and the previous five year monthly stock returns ), financial leverage (total debt divided by
total assets), size (natural logarithm of sales) and institutional ownership (percentage of shares held
by institutional investors). T-statistics are reported below the estimated coefficients. Significance
levels at the 1%, 5% and 10% are indicated as *, **, and ***, respectively.
Fixed
Fixed
Fixed
Tobit
Tobit
Effects
Effects
Effects
Dependent
Div. / Mkt. Div. / Mkt. Div./
Div./
Div. / Cash Div. / Cash
Variable:
Cap.
Cap.
Earnings Earnings Flow
Flow
Estimates Estimates Estimates Estimates Estimates Estimates
t-statistics t-statistics t-statistics t-statistics t-statistics t-statistics
Mgmt. Vote
-0.001
-0.001
-0.021
-0.015
-0.006
-0.004
-5 12***
-2.56**
-3.04*** -1.68*
-1.98**
-3 83***
-0.004
-0.123
-0.109
-0.066
-0.099
Capital Rationing -0.004
_3 7]***
-5.88***
-2.46**
-3.60*** -3 09***
-3 29***
0.006
0.232
0.029
-0.137
0.310
Performance
0.018
1.89*
3 3j***
3.26***
0.26
1.05
-1.05
-0.008
-0.028
-0.866
-0.194
-0.304
-0.080
Growth
_7 42***
-5 27***
-5 77*** -3 03*** -4 93***
-1.45
-0.003
-0.273
Risk
-0.009
-0.067
-0.132
-0.060
-8.75***
-8 95*** -4.14*** -10.02*** -4.26***
-11.74***
-0.001
-0.198
-0.131
-0.054
Financial Leverage -0.002
0.063
-0.77
-1.82*
-0.66
1.17
-2 78***
-0.90
0.001
0.011
0.064
Size
0.019
0.007
0.003
340***
4 93***
1.23
g 23***
1.02
-0.203
Institutional
-0.435
-0.260
-0.013
-0.008
-0.279
-4.01***
-3.62***
-4.20***
-3 31*** -2.48**
-4.91***
Ownership
0.324
0.008
0.131
0.030
0.401
-0.001
Intercept
c 30***
4 j^***
4.40***
2.62***
-0.17
2.26**
Industry and Year
Effects
No
Yes
No
Yes
No
Yes
70.64*** n/a
LR Chi-Square
n/a
273.68*** n/a
415.5 ***
g32***
7 j5***
n/a
F-statistics
n/a
n/a
35.68***
0.036
0.085
0.056
0.188
R-squared
0.143
0.155
1496
1398
1398
1496
Observations
1584
1584
Note: The R-squared for the Tobit regression is the square of the correlation between the model's
predicted values and the actual values.
Tobit
142
Table 3: Regression of total distribution

Payout ratio,, = p 0 + p,Mgm/ .Vote, r_, + y ' X + s,,
The dependent variable is the total distribution (cash dividends + repurchases) to common
shareholders scaled by market value of equity (dividend yield), earnings (dividend payout) and
operating cash flow. Management voting leverage (Mgmt. Vote) is the percentage of total votes
controlled divided by the percentage of total equity held by management and directors. X is a vector
of control variables including: capital rationing (the capital rationing dummy variable is equal to 1 if
by institutional investors). T-statistics are reported below the estimated coefficients. Significance
Fixed
Fixed
Fixed
Tobit
Tobit
Tobit
Effects
Effects
Effects
Total
Total
Total
Total Dist.
Dependent
Total Dist./
Total Dist./
Dist./
Dist./
Dist./
/Cash
Mkt. Cap.
Variable:
Cash Flow
Mkt. Cap. Earnings Earnings
Flow
Est.
Est.
Est.
Est.
Est.
Est.
t-stat
t-stat
t-stat
t-stat
t-stat
t-stat
Mgmt. Vote
-0.002
-0.001
-0.027
-0.021
-0.025
-0.018
-2 98***
-2 78*** -2.53**
-2.11**
-2.31**
-2.52**
Capital Rationing
-0.008
-0.007
-0.233
-0.194
-0.188
-0.138
-2.33**
-2.45**
-3.48*** -3 47***
-3.04***
-2.67***
0.110
0.072
Performance
1.743
1.136
1.564
0.697
2 74***
4.82***
4 77***
7.08***
6.01***
5.84***
Growth
-0.051
-0.035
-0.685
-0.488
-0.321
-0.503
-6 17***
-2 99***
-5.66*** -4.61*** _4 21 * * * -3.68***
-0.004
Risk
-0.001
-0.038
0.042
-0.034
0.021
-2.02**
-0.42
-1.05
1.43
-1.02
0.78
Financial Leverage -0.023
-0.018
-0.903
-0.525
-0.416
-0.629
-4.59***
-2.88***
-2.68*** -6.08*** -419***
-3.58***
0.004
Size
0.005
0.078
0.048
0.031
0.005
5 41***
4
3g***
3
1^***
4.50***
0.34
1.87*
-0.008
-0.168
-0.118
-0.322
-0.265
Institutional
-0.012
-0.98
-0.94
-1.95*
-1.91*
Ownership
-1.30
-0.79
0.015
0.221
0.416
0.363
0.606
Intercept
0.001
3 lg***
2.51**
2.05**
1.41
4.86***
0.77
Industry & Year
No
Yes
Yes
No
Yes
No
Effects
n/a
87.72*** n/a
86.40***
LR Chi-Square
146.31*** n/a
n/a
5.36***
9.26***
n/a
9.12**
F-Statistics
n/a
0.044
0.064
0.036
0.056
0.056
0.076
R-squared
1496
1584
1398
1398
1496
1584
Observations
Note: The R-squared for the Tobit regression is the square of the correlation between the model's
predicted values and the actual values.
143
Table 4: Descriptive statistics of controlling and non-controlling shareholders-CEOs

Panel A: Controlling shareholder-CEOs vs. non-controlling shareholder-CEOs
Controlling
Cash Dividend / Market Value %

Total Distribution / Market Value %
Cash Dividend / Earnings %
Total Distribution / Earnings %
Cash Dividend / Cash Flow %
Total Distribution / Cash Flow %
Capex / Sales %
Capex / Assets %
Mean
0.88
2.86
16.88
37.14
10.78
25.49
4.98
4.38
Median
0.51
1.59
7.01
24.57
2.92
11.34
3.53
3.30
Non-Controlling
Std. Dev
1.22
4.21
23.47
36.77
18.92
32.00
5.55
5.51
Mean
1.04
4.16
19.27
46.60
17.27
38.80
5.52
4.98
Median
0.63
2.25
10.27
36.77
5.71
21.36
3.45
3.68
Std. Dev
1.29
6.67
25.23
40.00
26.73
39.43
6.54
4.85
Test for
difference
in Mean
-2.55**
-4 40***
i 07**
-4.86***
-5.38***
-7 21***
-1.79*
-2.30**
Test for
difference
in Median
-2.06**
-4.42***
-1 96***
-4 35***
-3 45***
-5.85***
1.26
-0.93
Note: There are 645 controlling shareholder-CEOs and 939 non-controlling shareholder-CEOs
Panel B: Controlling shareholder-CEOs vs. non-controlling shareholder-CEOs in dual and single class firms
Test for difference in mean and median
Dual Class
Controlling minus Non-controlling
Mean
Median

-1.40
-3.69***
-0.90
-3.38***
.7.09***
-6.94***
Single Class
Controlling minus Non-controlling
Mean
Median
-0.88
-3 97***
-1.06
-1.34
-2.07**
-2.16**
-2.51**
-1.13
-2.02**
-1.52
-2.02**
-2.11**
-2.28**
-1.51
-1.58
-3.50***
-3.64***
-5 53***
144
Table 4 Cont'd
Panel C: Dual class and single class firms with controlling shareholder-CEOs
Dual Class
Mean
Median
Single Class
Std. Dev

Median
Std. Dev
Test for
difference
in Mean
Test for
difference
in Median
0.84
0.98
0.41
0.55
0.54
1.47 -1.30
1.13
1.64
2.76
3.19
5.14 -1.15
-0.29
1.58
3.86
14.49
7.96
23.80
16.91
6.15
22.38 -1.98**
1.79*
34.00
24.85
37.59
21.87
36.49
37.73 -2.18**
1.87*
10.28
12.28
3.64
20.50 -1.96**
3.01
18.37
-0.17
24.18
31.72
29.51
17.06
-2.08**
8.79
32.63 -2.18**
In the total sample, there are 486 dual class controlling shareholder-CEOs and 159 controllin g shareholder-CEOs in single class closely-held.
145
Table 5: Tobit regression of dividend payout ratios

5
Ln( Compensation )
Payout Ratio,
= a + r\kFirm Characteristicsk
E
t_l + 8, Governancek ,_
=(3 0 + Pi(C,, x ExcessCompl

+ P 3 (ExcessComplt
x Mgmt. Votelt_x) + p 2 ( C , t x Mgmf. Fote,,_,)
x Mgmt. Votelt_x)
+ fi5Mgmt. Vote,t_x + ^Family
+ (34 ( C x ExcessComp)l
Director,
+ &7EC,t + P 8 Excess Comp.,
+ y ' X + s/>(
The dependent variable is cash dividends paid to common shareholders scaled by market value of equity,
earnings and operating cash flow. Management voting leverage (Mgmt. Vote) is the percentage of total votes
controlled divided by the percentage of total equity held by management and directors. Executive-controlling
shareholder (EC) is an indicator variable equal to 1 if the controlling shareholder is the CEO or President and
zero otherwise (controlling shareholder is defined as an individual with ownership or control of 15% of the total
equity or total voting rights). Excess CEO total compensation (Excess comp.) is the residual from the first
equation above where firm characteristics include: size, profitability, growth, risk and financial leverage and
governance variables are as follows: board size, percentage of independent, busy and grey directors, percentage
of institutional ownership, CEO tenure, CEO-Chairman duality and percentage of family members on the board
of directors. X is a vector of control variables including: capital rationing (the capital rationing dummy variable
is equal to 1 if the average increase in capital stock plus financial debt as a ratio of sales is below the sample
median and the company's growth rate is above the sample median, otherwise it is set equal to 0), performance
(ROA=EBIT divided by total assets), growth is the geometric mean growth in total assets over the previous five
year period, risk (beta is estimated using the CRSP equally weighted index and the previous five year monthly
stock returns ), financial leverage (total debt divided by total assets), size (natural logarithm of sales) and
institutional ownership (percentage of shares held by institutional investors). Significance levels at the 1%, 5%
and 10% are indicated as *, **, and ***, respectively.
Dependent Variable:
EC x Excess Comp.
x Mgmt. Vote
EC x Mgmt. Vote
Excess Comp. x Mgmt. Vote
EC x Excess Comp.
Mgmt. Vote
Family Director
EC
Excess Comp.
Capital Rationing
Performance
Growth
Risk
Financial Leverage
Size
Intercept
Observations
LRChi
R-squared
Div. / Market. Cap
Div. / Earnings
Div. / Cash Flow
Est.
Est.
Est.
-0.0010
0.0002
-0.0007
0.0012
-0.0009
-0.0038
-0.0013
-0.0028
-0.0044
0.0160
-0.0289
-0.0093
-0.0017
0.0008
-0.0131
0.0185
1584
452.45***
0.179
t-stat
-2.05**
0.04
-2.52**
0.92
-3.01***
-0.85
-0.77
-4.54***
-3.84***
2.54**
_7 47***
22 9***
-0.73
3 02***
-4.41***
10.04***
146
-0.0297
0.0067
-0.0225
0.0811
-0.0231
-0.0645
-0.0456
-0.1067
-0.0989
0.0142
-0.9803
-0.2718
-0.1827
0.0396
-0.3917
0.4155
1398
303.94***
0.103
t-stat
_i 99**
0.37
-2.37**
1.44
-2.25**
-0.38
-0.72
-3 82***
-2.26**
0.06
-6.65***
-10 1***
-2.03**
4 21***
-3 49***
6.03***
-0.0327
-0.0072
-0.0230
-0.0612
-0.0063
-0.0765
-0.0003
-0.0695
-0.1103
0.3090
-0.5510
-0.2517
-0.2486
0.0823
-0.3931
0.0696
1496
370.80***
0.138
t-stat
-2.48**
-0.47
-1.69*
-1.76*
-0.77
-0.51
0.01
-2.65***
-2.86***
1.45
_4 37***
-10 5***
-3 09***
9 79***
-3 93***
1.15
Table 6: Tobit regression of total distribution

5
Ln( Compensation )
= a + X r\kFirm Characteristicsk
k=\
+
lt_\+
X 5,Governancek
'
k=\
,J
t_x
*j.t
Payout Ratio, t = P 0 + Pi(Cr ( x ExcessComp,, x Mgmt. Vote,,_,) + $2(EC, t x Mgmt. Vote,,_,)

+ P 3 (ExcessComp,
x Mgmt. Vote,,_,) + P 4 (C x ExcessComp),
+ fi5Mgmt. Vote,t_x + fi6Family Director, t + fi7EC, t + P8 Excess Comp., t

The dependent variable is total distribution to common shareholders scaled by market value of
equity, earnings and operating cash flow. Management voting leverage (Mgmt. Vote) is the
percentage of total votes controlled divided by the percentage of total equity held by management
and directors. Executive-controlling shareholder (EC) is an indicator variable equal to 1 if the
controlling shareholder is the CEO or President and zero otherwise (controlling shareholder is
defined as an individual with ownership or control of 15% of the total equity or total voting rights).
Family director is percentage of family members on the board of directors. Excess CEO total
compensation (Excess comp.) is the residual from the first equation above where firm characteristics
include: size, profitability, growth, risk and financial leverage and governance variables are as
follows: board size, percentage of independent, busy and grey directors, percentage of institutional
ownership, CEO tenure, CEO-Chairman duality and percentage of family members on the board of
directors. X is a vector of control variables including: capital rationing (the capital rationing dummy
variable is equal to 1 if the average increase in capital stock plus financial debt as a ratio of sales is
below the sample median and the company's growth rate is above the sample median, otherwise it is
set equal to 0), performance (ROA=EBIT divided by total assets), growth is the geometric mean
growth in total assets over the previous five year period, risk (beta ), financial leverage (total debt
divided by total assets), size (natural logarithm of sales) and institutional ownership (percentage of
shares held by institutional investors). Significance levels at the 1%, 5% and 10% are indicated as *,
**, and ***, respectively.
Dependent Variable:
EC x Excess Comp.
x Mgmt. Vote
EC x Mgmt. Vote
Excess Comp. x Mgmt. Vote
EC x Excess Comp.
Mgmt. Vote
Family Director
EC
Excess Comp.
Capital Rationing
Performance
Growth
Risk
Financial Leverage
Total Dist. /
Market Cap.
Est.
t-stat
-0.0026
-0.0005
-0.0014
-0.0028
-0.0011
-0.0128
0.0011
-0.0045
-0.0088
0.1085
-0.0472
-0.0050
-0.0104
Total Dist. /
Earnings
Est.
t-stat
-2.28** -0.0646
-0.49
0.0208
-2.42** 0.0109
-0.95
-0.0855
-1.69* -0.0230
-0.2384
-1.11
-0.0892
0.28
-2.31** -0.0624
-3 07*** -0.2347
8.15*** 1.9587
-6.38*** -0.6581
_3 14*** -0.0611
-0.6726
-1.58
147
-2.22**
0.87
0.89
-1.99*
-1.78*
-0.96
-0.98
-1.68*
-3 82***
Total Dist. /
Cash Flow
t-stat
Est.
-0.0413
-0.0093
-0.0181
-0.1324
-0.0184
-0.0342
-0.0933
-0.0793
-0.2767
5 7 0 * * * 2.0099
-4 27*** -0.3408
-1.82*
-0.1133
_c 2 1 * * * -0.5007
-0.75
-0.42
-1.64
1.91*
-1.31
-0.14
-1.07
-2.10**
.4 45***
5.82***
-2.19**
-3 28***
-3 47***
Table 6 Cont'd
Dependent Variable:
Size
Intercept
Observations
LRChi
R-squared
Total Dist. /
Total Dist. /
Total Dist. /
Market Cap.
Earnings
Cash Flow
Est.
t-stat
Est.
t-stat
Est.
t-stat
0.0028
5.49*** 0.1073
7.87*** 0.1450
13.03***
-0.0054
-0.73
-0.2012
-1.25
-0.2593
-1.59
0.0218
4.58*** 0.3799
3,89*** -0.0440
-0.41
1584
1398
1496
223.08***
199.91***
326.73***
0.162
0.081
O082
148
Table 7: Tobit regression for a sub-sample of family firms only

The dependent variable is equal to cash dividend or total distribution (cash dividends + share
repurchases) to common shareholders scaled by market value of equity, earnings and operating cash
flow. Management voting leverage (Mgmt. Vote) is the percentage of total votes controlled divided
by the percentage of total equity held by management and directors. Executive-controlling
shareholder (EC) is an indicator variable equal to 1 if the controlling shareholder is the CEO or
President and zero otherwise (controlling shareholder is defined as an individual with ownership or
control of 15% of the total equity or total voting rights). Excess CEO total compensation (Excess
comp.) is the residual from equation 3. Capital rationing (the capital rationing dummy variable is
equal to 1 if the average increase in capital stock plus financial debt as a ratio of sales is below the
sample median and the company's growth rate is above the sample median, otherwise it is set equal
to 0). Performance (ROA=EBIT divided by total assets), growth is the geometric mean growth in
total assets over the previous five year period, risk (beta is estimated using the CRSP equally
weighted index and the previous five year monthly stock returns), financial leverage (total debt
shares held by institutional investors). ***, ** and * denote significance at the 1, 5 and 10 percent
levels, respectively.
Dependent
Variable
EC x Excess Comp.
Mgmt. Vote
EC
Excess Comp.
Capital Rationing
Performance
Growth
Risk
Financial Leverage
Size
Cash Distribution
Div./
Div./
Div./
Market.
Earnings Cash
Cap
Flow
Total Distribution
Total
Total
Total
Dist. /
Dist. /
Dist. /
Market
Earnings Cash
Cap.
Flow
Estimates Estimates Estimates Estimates Estimates Estimates
t-statistics t-statistics t-statistics t-statistics t-statistics t-statistics
-0.002
-0.097
-0.062
-0.006
-0.138
-0.158
-3
49***
-2.56**
-2.13**
-3 28***
-2.06**
-2.54**
-0.003
-0.078
-0.081
-0.002
-0.144
-0.029
_1 99**
-2 79***
-1.80*
-0.41
-1.85*
-0.07
-0.001
-0.017
-0.014
-0.014
-0.001
-0.027
-2 99***
-2.52**
-2.49**
-2.05**
-1.46
-1.25
-0.001
0.024
0.021
0.151
0.001
0.087
0.94
2.81***
-1.25
0.78
0.38
2.24**
-0.004
-0.112
-0.104
-0.304
-0.315
-0.011
-3 28***
-4.28***
-3 24***
-2.55**
-2.33**
-4.36***
0.454
1.284
0.41
0.081
1.467
0.018
3
53***
2
94***
1.84*
2.41**
1.62
4.05***
-0.031
-0.980
-0.551
-0.037
-0.587
-0.446
-6 33***
-3 98***
-3 07***
-2.25**
-5.89***
-3.40***
-0.009
-0.226
-0.249
-0.006
-0.058
-0.128
_9 j4***
.7 34***
-2
99***
-7 70***
-1.43
-2.81***
-0.005
-0.415
-0.433
-0.075
-0.173
-0.007
-2 72***
-2 70***
-2.23**
-0.78
-1.59
-0.67
0.102
0.026
0.003
0.187
0.002
0.059
4
j2***
10.72***
6.07***
4.36***
2.51**
6.69***
149
Table 7 Cont'd
Dependent
Variable
Institutional
Ownership
Intercept
Observations
LRChi
R-squared
Cash Distribution
Div. /
Div. /
Div. /
Market.
Earnings
Cash
Cap
Flow
Total Distribution
Total
Total
Total
Dist. /
Dist. /
Dist. /
Market
Earnings Cash
Cap.
Flow
Estimates
t-statistics
-0.013
-3.30***
0.010
292***
Estimates
t-statistics
-0.470
-3.51***
0.419
542***
Estimates
t-statistics
-0.357
-2.61***
0.168
2 01**
889
270.82***
0.178
874
198.15***
0.137
884
196.44***
0.136
Estimates
t-statistics
-0.011
-1.04
0.018
3.07***
889
107.48***
0.162
150
Estimates
t-statistics
-0.148
-0.70
0.283
2.28**
874
126.04***
0.106
Estimates
t-statistics
-0.042
-0.19
-0.085
-0.66
884
222.61***
0.177
Table 8: Tobit regression for a sub-sample of dual class family firms only
flow. Management voting leverage (Mgmt. Vote) is the percentage of total votes controlled divided
by the percentage of total equity held by management and directors. Executive-controlling
shareholder (EC) is an indicator variable equal to 1 if the controlling shareholder is the CEO or
President and zero otherwise (controlling shareholder is defined as an individual with ownership or
control of 15% of the total equity or total voting rights). Excess CEO total compensation (Excess
comp.) is the residual from equation 3. Capital rationing (the capital rationing dummy variable is
equal to 1 if the average increase in capital stock plus financial debt as a ratio of sales is below the
sample median and the company's growth rate is above the sample median, otherwise it is set equal
to 0). Performance (ROA=EBIT divided by total assets), growth is the geometric mean growth in
total assets over the previous five year period, risk (beta is estimated using the CRSP equally
weighted index and the previous five year monthly stock returns), financial leverage (total debt
shares held by institutional investors). ***, ** and * denote significance at the 1, 5 and 10 percent
levels, respectively.
Dependent Variable
EC x Excess Comp.
Mgmt. Vote
EC
Excess Comp.
Capital Rationing
Performance
Growth
Risk
Financial Leverage
Size
Cash Distribution
Div./
Div./
Div./
Market.
Earnings Cash
Cap
Flow
Est.
t-stat
-0.002
-1.96**
-0.001
-2.52**
-0.002
-1.45
0.010
0.28
-0.001
-0.79
0.010
1.32
-0.031
-5.76***
-0.008
-7.50***
-0.005
-1.58
0.001
1.82*
Est.
t-stat
-0.118
i 71***
-0.020
-2 74***
-0.046
-0.92
0.023
0.81
-0.028
-0.57
0.177
0.67
-0.867
-4.58***
-0.201
-5.87***
-0.073
-0.70
0.017
1.44
Est.
t-stat
-0.067
-2.21**
-0.017
-2.41**
-0.014
-0.28
0.016
0.59
-0.022
-0.44
0.426
1.62
-0.533
-2.86***
-0.206
-5.87***
-0.182
-1.56
0.066
6.83***
151
Total Distribution
Total
Total
Total
Dist. /
Dist. /
Dist. /
Market
Earnings Cash
Cap.
Flow
Est.
Est.
Est.
t-stat
t-stat
t-stat
-0.223
-0.173
-0.005
-2.14**
-4 23*** -2.48**
-0.001
-0.015
-0.028
-1.72*
-1.16
-2.60***
-0.001
-0.036
-0.186
-0.42
-0.31
-2.26**
0.001
0.106
0.170
0.50
2.26**
2.86***
-0.192
-0.214
-0.003
-0.75
-2.24**
-2.62***
1.139
1.346
0.063
2 90***
2
90***
2.36**
-0.334
-0.044
-0.311
-3.62***
-1.26
-1.25
-0.056
-0.113
-0.003
-1.24
-1.13
-2.32**
-0.364
-0.581
-0.009
-3 42***
-1.08
-2.05**
0.004
0.092
0.208
A TJ*#*
4.58***
11.12***
Table 8 Cont'd
, .,, . .,
Dependent Variable
Institutional
Ownership
Intercept
Observations
LRChi
R-squared
Cash Distribution
Div. /
Div. /
Div. /
Market.
Earnings Cash
_,
pi
Est.
t-stat
-0.006
-1.58
0.014
3.89***
667
151.61***
0.112
Total Distribution
Total Dist. Total
Total
/ Market
Dist. /
Dist. /
Cap.
Earnings Cash
Flow
Est.
Est.
Est.
Est.
Est.
t-stat
t-stat
t-stat
t-stat
t-stat
-0.379
-0.197
-0.009
-0.119
0.040
-2.65*** -1.40
-0.82
-0.48
0.17
0.450
0.036
0.013
0.275
-0.173
4.95***
0.37
1.78*
1.74*
-1.16
648
656
667
648
656
125.51*** 139.6*** 77.52
78.29*** 205.23***
0.124
0.140
0.082
0.086
0.117
152
Table 9: Tobit regression of industry adjusted cash dividends and total distribution.
flow adjusting of industry median. For example, cash dividend-to-market capitalization (Div./
Market Cap.) is defined as Div./ Market Cap. for each firm minus its SIC industry median Div./
Market Cap. Management voting leverage (Mgmt. Vote) is the percentage of total votes controlled
divided by the percentage of total equity held by management and directors, capital rationing (the
capital rationing dummy variable is equal to 1 if the average increase in capital stock plus financial
debt as a ratio of sales is below the sample median and the company's growth rate is above the
sample median, otherwise it is set equal to O)performance (ROA=EBIT divided by total assets),
growth is the geometric mean growth in total assets over the previous five year period, risk (beta is
estimated using the CRSP equally weighted index and the previous five year monthly stock returns),
financial leverage (total debt divided by total assets), size (natural logarithm of sales) and
institutional ownership (percentage of shares held by institutional investors). T-statistics are reported
below the estimated coefficient. Significance levels at the 1%, 5% and 10% are indicated as *, **,
and ***, respectively.
Cash Distribution
Div. /
Div. /
Div. /
Market
Earnings Cash
Cap.
Flow
Dependent Variable
Mgmt.Vote
Capital Rationing
Performance
Growth
Risk
Financial Leverage
Size
Institutional
Ownership
Intercept
Observations
LRChi
R-squared
Est.
t-stat
-0.001
-2.64***
-0.001
-0.92
0.028
4.71***
-0.035
-7 80***
-0.006
-7 35***
-0.001
-0.36
0.001
4.06***
-0.010
-2.76***
-0.003
-0.86
1584
417.17***
0.188
Est.
t-stat
-0.021
-2.40**
-0.113
-2.14**
0.199
0.86
-0.928
-5 86***
-0.216
-6 96***
-0.144
-2.27**
0.040
2.97***
-0.449
-3.20***
0.036
0.30
1398
230.82***
0.065
Total Distribution
Total
Total
Total
Dist. /
Dist. /
Dist.
Market
Earnings /Cash
Cap.
Flow
Est.
Est.
Est.
t-stat
t-stat
t-stat
-0.001
-0.023
-0.026
-2 58**
-1 99**
-2 33**
-0.001
-0.163
-0.166
-0.24
-2.24**
-2.46**
0.102
1.568
1.499
5 76***
4 99***
5 15***
-0.043
-0.641
-0.406
-4 72***
-4 00***
-2 76***
0.003
0.035
0.018
1.28
0.89
0.49
-0.020
-0.865
-0.597
-2 29**
-5 34***
-3 98***
0.003
0.045
-0.010
2.41**
2.34**
-0.56
-0.001
-0.199
-0.392
-0.07
-1.01
-2.13**
-0.011
0.161
0.444
-1.17
0.93
2.77***
Est.
t-stat
-0.005
-1.35
-0.054
-2.41**
0.322
3.28***
-0.323
-5 01***
-0.103
_7 7i***
-0.145
-2.96***
0.008
1.40
-0.277
-4.60***
0.128
2.47**
1584
1496
270.23*** 150 67***
0.089
0.133
153
1398
123 71***
0.087
1496
9342***
0.075
Table 10: Tobit regression with predicted management voting leverage

flow. Predicted management voting leverage (Pred. Mgmt Vote) is the predicted management voting
leverage using equation (5), capital rationing (the capital rationing dummy variable is equal to 1 if
by institutional investors). T-statistics are reported below the estimated coefficient. Significance
Dependent
Variable
Pred. Mgmt Vote

Capital Rationing
Performance
Growth
Risk
Financial Leverage
Size
Institutional
Ownership
Intercept
Observations
LRChi
R-squared
Cash Distribution
Div./
Div./
Div./
Market.
Earnings Cash
Cap
Flow
Est.
t-stat
-0.002
-3 38***
-0.006
-2.19**
0.027
2.35**
-0.049
-6.10***
-0.014
-8.60***
-0.003
-0.59
0.005
6.67***
-0.024
-3.50***
-0.009
-1.53
1584
255.30***
0.198
Est.
t-stat
-0.045
_4 11***
-0.125
-2.52**
0.196
0.89
-0.857
_5 7i***
-0.272
-8 92***
-0.210
-1.93*
0.071
5 42***
-0.444
_3 4i***
0.053
0.46
1398
242.37***
0.182
Total Distribution
Total
Total
Total
Dist.
Dist. /
Dist./
Market
Earnings /Cash
Flow
Cap.
Est.
Est.
Est.
Est.
t-stat
t-stat
t-stat
t-stat
-0.028
-0.024
-0.002
-0.032
-2 59*** -2 61***
-2.30**
-2.39**
-0.122
-0.008
-0.234
-0.189
-2.82*** -2.34**
-3.50***
-3.06***
0.370
1.728
1.551
0.109
5 95***
5 79***
7 02***
1.92*
-0.501
-0.516
-0.050
-0.683
_4 59***
-4.16*** -6.15***
-3.66***
-0.242
-0.033
-0.004
-0.037
-9 01 * * * _1 99**
-1.03
-0.99
-0.908
-0.633
-0.215
-0.023
-2 90***
-2.24**
-6.10***
-4.62***
0.082
0.034
0.035
0.005
3 04***
5 57***
4 55***
2.07**
-0.301
-0.526
-0.011
-0.145
-0.82
-1.83*
-4.58*** -1.14
0.350
0.191
0.011
0.208
2.42**
1.89*
0.04
1.33
1496
1496
1584
1398
212.02*** 148.62*** 123.03*** 92.42***
0.152
0.145
0.185
0.161
154
Chapter 4
The Valuation Effects of Dual Class Structure and Managerial

Entrenchment
1.0
Introduction
The separation of voting and cash flow rights in dual class firms allows the
controlling shareholder-manager to become entrenched by rendering the market for

corporate control ineffective. In the absence of the threat of hostile takeover, controlling
shareholders-managers in poorly performing dual class firms face no threat of job dismissal.
Ruback (1988) argues that dual class ownership structure may be the most effective
universal anti-takeover device ever invented.
Concentrated equity ownership in single
class firms can also serve to entrench managers. Managers who control a substantial fraction
of single class firm's equity may have enough voting power or influence to guarantee their
employment with the firm (Morck et al., 1988). However, unlike controlling managers of
dual class firms, managers in single class firms with concentrated control still face a risk that
they will lose their control if they were to issue substantial amounts of new shares. Thus, the
ability to remain entrenched is not as severe for single class managers. This leads to several
empirical questions. Do managers in dual class firms display characteristics of
entrenchment? Does entrenchment in turn, lead investors to discount the value of dual class
firms?
Similarly, Jarrell and Poulsen (1988) argue that dual class structure provides an effective defense against
hostile takeovers.
155
Managerial entrenchment, broadly defined, occurs when managers gain so much

power that they are able to use the firm to further their own interests rather than the interests
of shareholders (Weisbach, 1988). It is the extent to which managers fail to experience
discipline from the full range of corporate governance and control mechanisms, including
monitoring by the board and the threat of dismissal or takeover (Berger, et al., 1997). In this
research, I define managerial entrenchment more precisely as occurring when the CEO
remains on the job longer than an industry peer or a matching company CEO, especially
when the CEO's company is performing relatively poorly.
In this essay, I investigate the relationship between dual class share structure,
managerial entrenchment and dual class discount. Managerial entrenchment is expected to
reduce the value of the firm as there is a lack of discipline on managers to address the poor
performance. Several studies provide evidence indicating that dual class firms are
discounted compared to single class firms (King and Santor, 2008, Gompers et al., 2010,
Smith et al., 2009). Family ownership of dual class firms lead to a 17% discount relative to
single class firms (King and Santor, 2008). In addition, Smith et al., (2009) find that the
value of dual class companies is discounted relative to single class concentrated control
companies. One possible explanation for the dual class discount is managerial entrenchment.
Entrenched managers are more likely to extract private benefits of control and therefore,
investors are likely to discount dual class firms relative to single class concentrated control
companies. In the first essay, I show that dual class executives received excess
compensation. Entrenchment allows them to enjoy this excess compensation for a longer
time.
156
In addition to dual class share structure, studies have identified several corporate
governance provisions such as staggered boards which allow managers to entrench
themselves. Gompers et al., (2003); Bebchuk and Cohen, (2005); Faleye, (2007) and
Bebchuk et al., (2009) provide evidence that firms with these provisions are valued less. In
this paper I extend their research by showing the link between these corporate governance
provisions, dual class discount and managerial entrenchment.
Using a sample of dual class firms and a propensity matched sample of single class
companies with concentrated control, I show that CEOs and directors in dual class firms are
more entrenched. CEOs and directors of dual class firms tend to have longer tenure
compared to their counterparts in single class companies. After adjusting for industry
median, dual class CEOs and directors remain on the job longer than those in similar single
class companies even when the company is performing relatively poorly. The regression
results indicate that investors apply a larger discount on the value of dual class firms which
have a greater degree of managerial entrenchment. The results are robust to several proxies
of managerial entrenchment. Furthermore, entrenchment is defined in the context of under
performance by management and hence, it is important to account for past performance.
Therefore, conditional on past poor performance, I show that dual class firms with excess
CEO tenure, excess E-index or excess director tenure are discounted more by investors. This
implies that investors are aware of the impact of managerial entrenchment in firms with dual
class ownership structure.
Studies often assume that managers are also controlling shareholders in firms with
concentrated ownership. However, this is not always the case. My data allows me to
separate dual class firms into groups where the controlling shareholder or a member of his
157
or her family is the CEO and those where the controlling shareholder is a member of the
board or the Chairman. The evidence from the sub-sample analysis indicates that the greater
the degree of managerial entrenchment the larger the dual class discount, especially in firms
with poor past performance. However, the results do not differ from the sub-sample of firms
where the controlling shareholder is not the CEO. In dual class firms where the controlling
share holder is a director or the Chairman, the CEO can become entrenched as long as their
interest does not diverge from that of the controlling shareholder. Therefore, the results of
the impact of entrenchment on dual class discount are independent of whether the CEO is
the controlling shareholder.
2.0
Literature Review and Hypothesis Development
2.1.0
Literature Review
Entrenchment can have adverse effects on management behaviour and incentives

(Bebchuk et al., 2009). Managerial entrenchment may result in agency costs to shareholders
by allowing poor performing managers to remain on the job and allow for empire-building.
Entrenching mechanisms allow managers to pursue their own interests and extract wealth at
the expense of outside shareholders. Florackis and Ozkan (2009) provide evidence that firms
with high levels of managerial entrenchment exhibit higher agency costs.
Other studies
provide evidence that firms with entrenched managers significantly underperform, hold
large amounts of cash, pay lower dividends, and are less leveraged (Morck et al., 1988
66
They used the inverse of asset turnover as a measure of agency costs and interpret this ratio as an asset
utilization ratio which shows how effectively managers deploy firm assets.
158
Berger et al., 1997; Gompers et al., 2003; Ozkan and Ozkan, 2004; Hardford et al, 2008 and
Khan, 2006).
Studies have argued that entrenchment can also produce benefits to shareholders by
reducing the extent to which the threat of a takeover distorts investments in long-term
projects (Stein, 1988 and Bebchuck and Stole, 1993). Also, Stulz (1988) argues that
entrenching mechanisms allow managers to extract higher acquisition premiums in
negotiated transactions. Prior studies such as Wilcox (2002) and Faleye (2007) argue that
staggered elections of directors as an entrenching mechanism encourage board
independence by reducing the threat that a director who refuses to succumb to management
will not be renominated each year.
Managerial entrenchment can occur in several ways, including manager-specific
investment, concentrated ownership and control as well as various anti-takeover provisions
such as staggered boards. By making corporate investments that fit the expertise of a
particular CEO, that CEO can reduce the probability of being replaced, can extract higher
wages and larger perquisites from shareholders (Shleifer and Vishny, 1989). For example,
Shliefer and Vishny (1989) argue that excessive growth in sales in the direction of the CEOs
talents and experience is a means of entrenchment. Long tenured managers, because of
manager-specific assets, are more valuable to shareholders compared to an alternative
manager. As a result, these managers can negotiate for higher compensation and increase
their latitude in running the firm.
Concentrated ownership and control resulting from dual and single class share
structure may insulate managers from dismissal even in poorly performing firms. In dual
class firms, both internal and external corporate governance mechanisms may be ineffective
159
in removing poorly performing managers because managers control the firm with a
significant proportion of the votes while owning a small fraction of the equity. Furthermore,
in both dual and single class firms, there is a greater degree of managerial control and
entrenchment of managers as managerial ownership increases (Morck et al., 1988).
Therefore, the probability of replacing executives in firms with concentrated ownership and
control is significantly reduced. Executives in firms with concentrated ownership and
control may display several characteristics of entrenchment such as long tenure in office and
compensation that has low sensitivity to performance (Berger et al., 1997). In fact, Morck et
al. (1988) argue that entrenchment is not just a consequence of voting power. Some
managers, by virtue of their tenure with the firm, status as founder or even personality, can
be entrenched.
Managerial equity ownership or voting control makes it more difficult to remove a
poorly performing top executive. Denis et al. (1997) find that the probability of turnover is
significantly less sensitive to performance when officers and directors own between 5% and
25% of a firm's equity than when officers and directors own less than 5%. Similarly, Huson
et al. (2001) provide evidence that the likelihood of forced turnover is negatively related to
the CEO fractional ownership. This may be more pronounced in dual class firms since the
market for corporate control is virtually ineffective as a disciplinary mechanism as
executives often control the firm with a significant proportion of the voting rights and a
small fraction of the equity ownership.
In addition to dual class share structure, anti-takeover provisions such as staggered
boards and poison pills may serve to entrench managers and therefore, have a negative
impact on firm value. Several studies argue that staggered boards can insulate management
160
from the market for corporate control because of the boards' ability to adopt and maintain
poison pills (Bebchuk and Cohen, 2005; Faleye, 2007 Bebchuk et al., 2009). Bebchuk and
Cohen (2005) provide evidence that staggered boards are associated with an economically
meaningful reduction in the firm value. This result is stronger for firms which establish
staggered boards through corporate charter which shareholders cannot amend compared to
staggered boards established in the company's by-laws. Similarly, Faleye (2007) shows that
staggered boards destroy value by entrenching management and reducing the likelihood of
forced CEO turnover. He argues that staggered boards insulate management from market
discipline and diminish board accountability. Bates et al. (2008), on the other hand, argue
that staggered boards do not change the likelihood that a firm, once targeted, is ultimately
acquired. In fact, shareholders of target companies with staggered boards realize bid returns
that are equivalent to those of targets with a single class of directors. They provide evidence
that staggered boards reduce the likelihood of receiving a takeover bid. However, the
economic effect of the bid deterrence on the value of the firm is quite small.
A staggered board is only one of several provisions which may serve to entrench
managers. For example, limits to shareholder by-law amendments, golden parachutes and
supermajority requirements for mergers can also be considered as entrenching provisions. In
fact, Gompers et al., (2003) consider 24 such provisions followed by the Investor
Responsibility Research Centre (IRRC) in the construction of their governance index (Gindex). However, examining these provisions, Bebchuk, et al. (2009) construct an
entrenchment index (E-index) and provide evidence that the E-index level is monotonically
associated with reduction in firm valuation during the period 1990-2003.
67
They use six of the 24 provisions followed by the Investor Responsibility Research Centre which is used to
construct the G-index. Four of the six provisions (staggered boards, limits on shareholder amendments of the
161
2.2.0
Hypothesis
Dual class ownership structure can lead to managerial entrenchment by allowing
managers who control a majority of the voting rights in dual class firms to become
entrenched. In turn, the more entrenched dual class managers are, the more likely it is that
such managers will extract pecuniary and non-pecuniary benefits at the expense of outside
shareholders. As a result, investors are expected to apply a greater discount on the value of
dual class firms. Several characteristics of managerial entrenchment such as CEO tenure,
and directors' tenure are expected to be related to the observed valuation discount of dual
class firms. Given this argument, the hypothesis follows:
H1: The greater the managerial entrenchment, the larger the dual class discount.
3.0
Methodology and Data
3.1
Methodology
To examine the effects of managerial entrenchment on valuation discount of dual
class firms (HI). I estimate equation (1) below. I expect entrenchment proxies to be
negatively related to dual class discount, that is, the higher the level of managerial
entrenchment, the greater the discount.
by-law, supermajority requirement for mergers and supermajority requirements for charter amendments) set
constitutional limits on shareholders voting power. The remaining two provisions reduce the impact of market
for corporate control (poison pill and golden parachute). They argue that the remaining provisions were
uncorrected with firm valuation.
162
Dual Class Discount
t+x
= a + $ Entrenchment
+ (3j Excess Compensation
+ p2Mgra. Vote t + P3 {Mgmt. Votex Excess Cadi)

+ P Excess Cadi
+ P Financial Leverage
+ P ConversionRight
+ P Size
+ p Div. diff
+s
(1)
The dependent variable, dual class discount, is computed as the difference in

Tobin's Q ratio of dual class firms and their propensity score matched single class
concentrated control firms. Alternatively, dual class discount is the difference between
Tobin's Q ratio of dual class firms and the industry average Q ratio. There are several
managerial entrenchment proxies utilized in the specification (1) above. First, excess CEO
tenure is the difference in tenure for dual class CEOs and their matching counterpart in
single class firms with concentrated ownership. The second entrenchment proxy is industry
adjusted CEO tenure (IACEO Tenure). It is computed as the difference between tenure of
dual class CEOs and the median industry CEO tenure. The third measure, Excess E-index,
utilizes the E-index which consists of several anti-takeover provisions that may result in
managerial entrenchment. Excess E-index is the difference between the E-index value of
dual class firms and matching single class firms. Excess G-index value is also included as an
additional control variable. In constructing the excess G-index, I subtract the E-index value
for each firm from the G-index value and then take the difference between dual class firms'
G-index and matching single class firms' G-index.6 Finally, industry adjusted directors
tenure (IADirectors tenure) is used as a proxy for entrenchment. IADirectors tenure is
computed as the median tenure per director less the median industry tenure per director.
Entrenchment of directors is likely to be more pronounced in dual class firms because the
Using average industry CEO tenure produces similar results.
The G-index is corporate governance index which is constructed using 24 governance provisions followed by
Investor Responsibility Research Centre (Gompers et al., 2003).
70
Using mean tenure per director produces similar results.
69
163
controlling shareholders can use their voting power to elect directors who are less likely to
act against the interest of the controlling shareholders. Also, Morck et al. (1988) suggest that
outside board members are capable of becoming entrenched.
Since excess compensation is not given, I need to find a methodology to measure
excess CEO compensation. Thus, following Zingales (1995) and Masulis et al. (2009), I
estimate equation (2) using firm characteristics and governance variables that have been
proven to explain executive compensation and extract the residuals as a measure of CEO
total excess compensation using equation (2).
Ln(Compensation)
= a ; + YJr\lkFirm Characteristicsk
=i
'
'
+ 1 5 , governance
t_x
+s
t_x
'
,J
(2)
k=\
where Ln(Compensation) is the CEO total compensation for firm j at year t. Total
compensation (TDC1) is defined as salary + bonus + other compensation + stock options.
Firm characteristics and governance variables are based on prior studies such as Smith and
Watts (1992), Core et al. (1999) and Chalmers et al. (2006). Firm characteristics include:
size, profitability, growth, risk and financial leverage. The governance variables are as
follows: board size, percentage of independent, busy and grey directors, percentage of
institutional ownership, CEO tenure, percentage of family members on the board of
directors and CEO-Chairman duality dummy variable. It is possible that entrenchment and
excess compensation are correlated because entrenched managers have the ability to extract
71
Other compensation includes the value of restricted stock grants, long term incentive payouts, contributions
to pension plans, life insurance premiums, consulting fees and awards under charitable award programs.
72
A busy director is defined as a director with more than four board memberships and grey directors are
defined as outside directors who are related to the company through a transactional relationship.
164
higher compensation from their firm. However, I examined the correlation between excess
compensation and the various proxies for entrenchment. The correlation is relatively low
with the highest being 0.07 between excess compensation and industry adjusted CEO tenure
(IADCEO tenure).
Following Zingales (1995), conversion right is an indicator variable equal to 1 if
superior voting shares can be converted into restricted voting shares and 0 otherwise, market
value of equity (size) is used as a proxy for the probability of acquisition and dividend
difference (Div. diff) is an indicator variable equal to 1 if the dividend paid or payable to
restricted voting shares is greater than that of the superior voting shares and zero
otherwise.73
3.2
Data
The dataset used in this research is constructed from a variety of sources.

To construct a sample of U.S. dual class firms, I retrieve a list of firms with dual of share
structure from Corporate Library for 2005-2007. During this period, Corporate Library
identifies all firms with dual class common shares as a takeover defense mechanism for
2005-2007.1 extend this list of dual class firms over this period to 2001 using a list of dual
class IPOs and a list of dual class firms used in Gompers et al. (2010).74 The list of 1,910
dual class firms over the period of 2001-2007 is merged with Execucomp database to
73
In 13.5% of the dual class firms, holders of restricted voting (RV) shares are paid more or will receive more
dividends in the future relative to holders of superior voting (SV) shares.
74
165
determine whether compensation data is available for these firms. Execucomp database
contains executive compensation data for the top executives representing the S&P 1500
group of companies. For each dual class company with compensation data, I retrieve proxy
statements from the Securities and Exchange Commission (SEC) website and check the
proxy statement for each firm in the sample to ensure that they are in fact, dual class
companies. Next, using proxy statements, I collect voting rights per share and the number
of superior voting and restricted voting shares owned by the largest shareholders and
management and directors as a group.
information such as total assets, sales, long-term debt, common equity and operating
income. In addition, I collect several governance variables and equity ownership data from
Corporate Library and Execucomp. These include the number of directors, outside related
directors and unrelated directors. I use proxy statements, firm websites and internet search
engines such as Lexus Nexus and Google to identify family executives and family directors.
I then calculate the percentage of family members who are directors of the board. In order
to complete the set of control variables, I collect stock return data from CRSP. I obtain
monthly returns to estimate beta and annual returns to compute standard deviation. Finally,
I collect CEO total compensation (TDC1) from Execucomp.
The list of dual class companies is matched with a list of single class concentrated
control firms using propensity score matching. Propensity score matching methods were
developed by Rosenbaum and Rubin (1983), Heckman and Robb (1986) and Heckman et al.
166
(1998). One of the major benefits of propensity score matching is that it can accommodate a
larger number of matching variables which can correct for the bias due to systematic
differences between the treated and control groups. The greater the overlap in all
characteristics of the treated and control groups, the more comparable the groups are and the
smaller the bias (Heckman et al., 1997 and Heckman et al., 1998). As a result, propensity
score matching has become a popular matching technique applied to studies of financial
markets (see Hillion and Vermaelen, 2004 and Villalonga, 2004).
directors, grey directors, institutional ownership, company age, R&D-to-sales, capex-to-total
asset and family firms. This matching exercise results in a final sample of 792 dual class
firm-years over the period of 2001-2007. This represents an average of 113 dual class firms
per year. The final matched sample is made up of 1,584 firm-year observations.
167
4.0
Results
4.1.0
Table 1 Panel A, reports the descriptive statistics of several different characteristics
of managerial entrenchment. There is no difference in mean and median CEO age in dual
class firms compared to single class concentrated control companies. Similarly, there is no
difference in the mean (median) age of directors in dual and single class companies. The
median age of directors in both dual and single class companies is 59 years. The first proxy
for managerial entrenchment, CEO tenure, indicates that dual class CEOs tend to remain on
the job longer than their matching counterparts in single class firms. The difference is
positive and significant as indicated by the tests for difference in the mean and median. Dual
class CEOs, on average, retain their position for 5.3 years longer than CEOs in single class
firms. This is an indication of managerial entrenchment. Alternatively, since dual class firms
have a higher concentration of family involvement, the longer tenure may reflect such
involvement. Family CEOs remain longer in their position to give the next generation time
to mature enough to succeed them.
Using the second measure of entrenchment, directors' tenure, dual class directors
have longer tenure compared to directors in single class firms. The average tenure per
director in dual class firms is 1.9 years longer than those in single class concentrated control
companies. This suggests that controlling shareholders use their voting power to elect and
maintain a board of directors who will stay on longer and act in their interest. In Table 1
Panel A, I also report two industry adjusted measures of entrenchment (industry adjusted
CEO and directors' tenure). The results indicate that dual class CEOs and directors have a
168
longer tenure relative to single class CEOs and directors after adjusting for average industry
tenure. CEOs in dual class firms serve in this capacity for 5.66 years longer than their
industry peers. In comparison, CEOs in single class concentrated control firms serve in this
role only 0.51 years longer than their industry peers. This implies that controlling
shareholders of dual class firms are using their voting power to remain on the job longer or
keep in place a CEO who acts according to the interests of the controlling shareholder.
Therefore, CEOs of dual class firms are more likely to be entrenched. Investors, knowing
this, are more likely to discount the value of dual class firms relative to single class firms.
The third managerial entrenchment measure, E-index, is lower for dual class firms
than for single class concentrated control companies. One possible explanation is that dual
class structure is the most effective anti-takeover defense and therefore, dual class firms do
not need additional anti-takeover defenses such as classified boards. Bebchuk et al. (2009)
argue that holders of superior voting rights might be sufficient to provide incumbents with a
powerful entrenching mechanism that renders other entrenching provisions relatively
unimportant. Nevertheless, dual class firms typically have 2 anti-takeover provisions which
are identified by Bebchuk et al. (2009) as a part of their E-index and may serve to entrench
managers.
Conditional on poor past performance, dual class CEOs and directors typically
remain on the job longer than their single class counterparts (Table 1, Panel A). The
univariate test for difference in mean is statistically significant. CEOs of dual class firms
with poor past performance, relative to the industry, remain on the job 8.63 years longer
than similar CEOs in single class firms with concentrated ownership. Longer tenure is an
indication of managerial entrenchment especially when firms have prior poor performance.
169
The last two columns in Table 1 Panel B show the test statistics for the difference in
means and medians for the two samples. In the dual class sample, the largest shareholder
owns, on average, 22.5% of the equity stake compared to 57.8% of the voting rights.
Management and directors as a group, control 58.3 % of the total votes compared to 24.9%
of the equity stake. In comparison, the largest shareholder in single class firms owns, on
average, 23.6% of the equity outstanding. The disparity between voting and cash flow rights
in dual class firms is at the heart of the agency problems associated with this type of
ownership structure. It can allow managers to become entrenched with a small proportion of
the equity capital.
Table 1 Panel C reports descriptive statistics for firm characteristics. Based on the
mean and median tests, it is evident that the propensity score matching exercise produces
samples of dual and single class firms that are very similar. There is no difference in size
(sales), financial leverage (D/A), performance (ROA), risk (beta) and growth (total assets).
The tests for mean (median) difference are insignificant for these variables.
4.2.0 Regression Analysis
Table 2 reports the effects of various entrenchment proxies on dual class discount. In
this table, dual class discount is computed as the difference in the Q ratio of dual class firms
and the Q ratio of propensity score matched single class concentrated control firms. In
model (1), entrenchment (excess CEO tenure) is measured as the difference between CEO
tenure in a dual class firm and its matching single class firm. The coefficient for excess CEO
tenure is negative and statistically significant. As expected, the greater the degree of
entrenchment, the larger the discount of dual class firms relative to single class companies.
170
Furthermore, Goyal and Park (2002) show that the probability of CEO turnover is
significantly lower when the CEO also serves as the Chairman of the board. Therefore, I
include an indicator variable equal to 1 if the CEO is also the Chairman. The coefficient is
not statistically significant. This implies that the dual role of CEO and Chairman does not
seem to affect dual class discount. I include several other variables affecting the discount
such as excess compensation, management voting leverage and interaction between excess
cash and management voting leverage. These variables serve as proxies for the extraction of
private benefits. All of these variables are significant with the expected sign. Other control
variables are based on prior studies such as Zingales (1995).
In model (2), excess CEO tenure (IACEO Tenure) is measured as the difference
between CEO tenure in dual class firms and the median industry CEO tenure. This
entrenchment proxy is negative and statistically significant as hypothesized above. The
entrenchment proxy in Model (3) is excess E-index. The excess E-index is the difference in
the E-index value between a dual class firm and its matching single class counterpart. The
coefficient is negative and statistically significant at the 10% level. In Model (4),
entrenchment is measured using industry adjusted directors, tenure (IADirectors' tenure).
The coefficient is negative and significant indicating the greater the entrenchment, the larger
the dual class discount. Finally, since the correlation among the various entrenchment
ye
proxies are relatively low, I include all the proxies in Model (5). All of the entrenchment
variables are negative and significant which confirms my expectation that investors discount
the value of dual class firms which appear to have entrenched managers and directors.
The highest correlation is 0.40 between IACEO tenure and IADirectors tenure.
171
Table 3 reports the results using an alternative measure of dual class discount. It is
the difference in Q ratio of dual class firms and the industry average Q ratio. In Models (1)
to (5), the entrenchment coefficients are similar in significance levels and magnitude to
those reported in Table 2. The control variables proxying for private benefits are all negative
and significant. As for the other control variables, they are similar to those reported in Table
2 except for dividend difference which is positive and significant. In these firms, restricted
voting shareholders are entitled to receive higher dividends relative to superior voting
shareholders. As expected, higher dividends reduce the dual class discount.
4.3.0
Entrenchment Conditional on Past Performance
The prior literature on managerial entrenchment often utilized anti-takeover

provisions, executive tenure, and age to proxy for managerial entrenchment (Berger et al.,
1997; Yermack, 2006 and Bebchuk et al., 2009). However, age and tenure can also proxy
for valuable experience (Norburn and Birley, 1988). Salas (2010) argues that executive
tenure conditional on firm performance is a more suitable measure of managerial
entrenchment as managers are truly only entrenched when they are not removed in the face
of poor relative performance. In light of this argument, it is important to control for firms'
prior performance when examining entrenchment. Although, the entrenchment measures
utilized in this study are "excess" measures, entrenchment proxies conditional on prior
performance will provide a more robust measure of the impact of managerial entrenchment
on dual class discount.
Accordingly, I construct two dummy variables based on prior operating
performance. The first performance dummy variable (Perduml) is equal to 1 if dual class
172
firms' previous 3-year average ROA is less than the ROA of matching single class firms.
The second performance dummy variable (Perdum2) is equal to 1 if the firm's 3-year ROA
is less than the 3-year industry average ROA. The 3-year average ROA is utilized as a
performance in order to eliminate the impact of any transitory effects of operating
performance. In addition, Denis and Denis (1995) show that firms with 3-years of prior poor
operating performance are more likely to replace their CEO. Similarly, Huson et al. (2001)
provide evidence that executive turnover tends to occur when industry adjusted accounting
performance has declined and stock returns have recently been negative. Hence, CEOs with
long tenure conditional on poor past performance is a clear indication of managerial
entrenchment.
Table 4 presents the results of the effects on managerial entrenchment conditional on
prior operating performance. I include an interaction term between performance dummy and
various proxies for entrenchment. In models (1) and (4), the performance benchmark is
based on the matching group of control firms (Perduml). In models (2), (3) and (5), the
performance benchmark is based on the industry average performance (Perdum2).
In models (1) and (3), the performance dummy variables are interacted with excess
CEO tenure. The coefficient of the interaction term is negative and statistically significant.
This indicates that dual class firms with previous poor performance and greater managerial
entrenchment are valued less. In model (2), I use industry as the benchmark for both
performance (Perdum2) and CEO tenure (IACEO tenure). The results show that investors
apply a larger discount to dual class firms with worse performance relative to the industry
and when the CEO remains on the job longer than their industry counterpart. This implies
that investors are concerned with managerial entrenchment especially in firms that are
173
performing poorly. In model (5), I used industry adjusted directors tenure (IADirectors
tenure) as the proxy for entrenchment. The results show that firms with poor performance
and longer directors' tenure are valued less.
In Table 5, the dependent variable, dual class discount, is computed as the difference
between Q ratio of dual class firms and the industry average Q ratio. The results presented in
Table 5 are similar to those presented in Table 4. In addition, the interaction term between
performance and excess E-index is now significant. Investors apply a greater discount to
dual class firms with poor performance and excess E-index. This implies that investors view
anti-takeover provisions, which are a part of E-index, as entrenching provisions especially in
firms with poor performance.
Studies which examine concentrated ownership often simply assume that managers
are also the controlling shareholders. However, this is not always the case. In this study, I
am able to identify firms where the controlling shareholder is an executive and those where
the controlling shareholder is a director or Chairman of the board. This is important because
it is easier to identify cases of entrenchment when the CEO is the controlling shareholder or
a family member of the controlling shareholder. However, identifying cases of
entrenchment of outside CEOs in firms with concentrated ownership is a bit more
challenging. Nevertheless, controlling shareholders are less likely to hire an outside CEO
who will openly oppose them or act against the interest of the controlling shareholder. An
outside CEO can become entrenched as long as their interest does not diverge from the
interest of the controlling shareholder.
In Table 6, I provide evidence for managerial entrenchment in dual class firms
where the controlling shareholder is also an executive. The results are similar to those
174
presented in Table 4. Longer tenure of controlling shareholders-executives provides robust

evidence that investors are aware of the impact of managerial entrenchment on the
extraction of private benefits and agency costs. Controlling shareholders-executives are
more likely to be entrenched because they control dual class firms with majority voting
rights.
Table 7 presents the results for dual class firms where the controlling shareholder is
not a member of the executive team. In these firms, the controlling shareholders are
typically members of the board of directors. The results are similar to those presented in
Table 6 only for models (2) and (5). By separating the sample into firms with controlling
shareholders as CEOs and those with non-controlling shareholders as CEOs, I present
further evidence of the impact of managerial entrenchment on dual class discount.
In
addition, the results presented in Tables 2 to 5 are independent of whether the executives are
also the controlling shareholders. Executives who are not controlling shareholders can also
be entrenched because controlling shareholders are more likely to hire executives who are
less likely to act against the interest of the controlling shareholders.
4.4.0 Robustness
Potential endogeneity concerns are common in corporate governance literature.

Simultaneity and reverse causality can bias our results. Managerial entrenchment can lead to
dual class discount but it is unlikely that dual class discount influences managerial
entrenchment. Nevertheless, I utilized a two stage least square technique as a robustness
check. In the first stage, I estimate a model for the determinants of entrenchment. The
76
In firms with non-controlling shareholder as CEOs, the controlling shareholders are usually Chairmen or
directors of the board.
175
following regressors are included: dual class dummy, E-index, G-index (excluding the Eindex value) and management voting leverage. Dual class ownership structure is arguably
the most effective anti-takeover defense mechanism which can allow managers to become
entrenched. Similarly, corporate governance provisions which make up the E-index such as
poison pills may also render the market of corporate control ineffective and hence, lead to
entrenchment. In addition, the voting power of superior voting shares in dual class firms can
allow managers to become entrenched. Therefore, I utilized these variables in order to
predict management entrenchment. The predicted variable is then used to explain the
documented dual class discount.
In the second stage, I include the predicted entrenchment using the estimated
coefficients from the first stage. I utilized three measures of entrenchment including: CEO
tenure, industry adjusted CEO tenure and industry adjusted directors' tenure. The results of
the second stage estimation are presented in Table 8.1 created interaction terms between the
performance dummy and the predicted entrenchment variables. The performance dummy is
equal to 1 if a dual class firm's previous 3-year ROA is less than the ROA of a matching
single class firm with concentrated ownership and zero otherwise. In models (1) to (3), the
dependent variable, dual class discount, is the difference in Q ratio of dual class firms and
their matching single class counterpart. The interaction term in each model is negative and
significant. This is consistent with the above hypothesis and with the results presented in
Table 4. In models (4) to (6), the dependent variable is the difference between the Q ratio of
dual class firms and the industry average Q ratio. The results are similar to those presented
in Table 5.
176
5.0

Dual class ownership structure is one of the most effective anti-takeover defense
mechanisms. It can lead to entrenchment since controlling shareholders can maintain a

voting block of shares and still raise additional equity capital. Entrenchment can be viewed
as a benefit to controlling shareholders that imposes agency costs to outside shareholders.
Therefore, investors are likely to attach lower value to dual class firms with a greater degree
of entrenchment. In this essay, I investigate whether dual class firms have a greater degree
of managerial entrenchment compared to single class firms and whether investors discount
the value of dual class companies that have signs of entrenched managers. The evidence
provided in the research shows that dual class CEOs and directors are entrenched compared
to single class CEOs. After adjusting for industry median, dual class CEOs and directors
have longer tenure than CEOs in single class firms even when the firm is making losses.
Univariate tests indicate that dual class CEOs and directors have a longer tenure than
their counterparts in single class firms with concentrated control. Dual class CEOs on
average, remain on the job 5.32 years longer than CEOs in similar single class firms. This
evidence can be interpreted as managerial entrenchment. However, longer tenure may also
indicate experience and superior performance in running the firm. Therefore, I construct
tests conditional on poor past firm performance using the industry and a matching sample as
benchmarks. Longer tenure of CEOs and directors in dual class firms when these firms
perform poorly in the past is consistent with managerial entrenchment. The univariate test
shows that in dual class firms with poor past performance, CEOs have a longer tenure (2.71
years) compared to CEOs in similar single class firms.
177
In a panel regression specification, I use several measures of entrenchment and two

different measures of dual class discount and provide evidence that the greater the
entrenchment, the larger the dual class discount. Excess CEO tenure (compared to matching
firms and industry median) leads to a greater discount of dual class firms. Also, investors
apply a greater discount to dual class firms with excess E-index and longer directors' tenure.
Although CEO tenure is an indication of managerial entrenchment, it can also signify
experience and superior ability in running firms. Therefore, I provide evidence conditional
on poor past performance. When firms perform poorly, managers who are not entrenched
are more likely to lose their job. Therefore entrenchment measures should be conditional on
poor past performance. I provide evidence that investors apply a greater discount to dual
class firms with excess CEO tenure and excess directors' tenure for dual class firms with
poor past performance.
I provide further evidence by examining sub-samples of dual class firms with
controlling shareholders as CEOs and those where the controlling shareholder is a director
or Chairman of the board. Longer tenure of controlling shareholders-CEOs provides robust
evidence of the relationship between managerial entrenchment and dual class discount.
Longer tenure in dual class firms with controlling shareholders-CEOs results in a greater
discount of these firms especially when prior performance is less than a matching firm or the
industry average. The evidence suggests that investors are concerned with managerial
entrenchment, potential agency costs and extraction of private benefits and therefore, attach
lower value to dual class companies which have a higher degree of entrenchment.
178
6.0
References
Bates, T. W., Becher, D. A., and Lemmon, M. L., (2008), Board classification and
managerial entrenchment: Evidence from the market for corporate control, Journal of
Bebchuk, L., and Cohen, A., (2005), The costs of entrenched boards, Journal of Financial
Bebchuk, L., and Stole, L., (1993), Do short-term managerial objectives lead to under-orover-investments in long-term projects? Journal of Finance, 48, 719-729
Bebchuk, L., Cohen, A., and Ferrell, A., (2009), What matters in corporate governance?
Review of Financial Studies, 22, 783-827
Berger, P., Ofek, F and Yermack, D. L., (1997), Managerial entrenchment and capital
structure decisions, Journal of Finance, 52, 1411-1438
Coughlan, A., and Schmidt, R., (1985), Executive compensation, management turnover, and
firm performance: An empirical investigation, Journal of Accounting and Economics, 1,
43-66.
Denis D. J., Denis D. K., (1995), Performance changes following top management
dismissals, Journal of Finance, 50, 1029-1057
Denis D. J., Denis D. K., and Sarin, A., (1997), Ownership structure and top executive
turnover, Journal of Financial Economics, 45, 193-221.
Faleye, O., (2007), Classified boards, firm value and managerial entrenchment, Journal of
Florackis, C and Ozkan A., (2009), The impact of managerial entrenchment on agency
costs: An empirical investigation using U.K. Panel Data, European Financial
Gompers, P., Ishii, J., Metrick, A., (2003), Corporate governance and equity prices,
Quarterly Journal of Economics, 118, 107-155
Gompers, P. A., Ishii, J. and Metrick, A., (2010), Extreme governance: An analysis of dualclass firms in the United States, Forthcoming Review of Financial Studies, 23, 10511088
Goyal, V. K., and Park, C. W., (2002), Board leadership structure and CEO turnover,
179
Harford, J., Mansi, S. A., and Maxwell, W. F., (2008), Corporate governance and firm cash
holdings in the U.S., Journal of Financial Economics, 87, 535-555
Heckman, J., (1979), Sample selection bias as a specification error, Econometrica, 41, 153161
Heckman, J., and Robb R., (1986), Alternative methods for solving the problem of selection
bias in evaluating the impact of treatments on outcomes, Drawing Inference from SelfSelected Samples, ed. H. Wainer. Berlin: Springer-Verlag
Heckman, J., Ichimura, H., and Todd P., (1997), Matching as an econometric evaluation
estimator: Evidence from evaluating a job training programme, Review of Economic
Studies, 64, 605-654
Heckman, J., Ichimura, H., Smith, J., and Todd P., (1998), Characterizing selection bias
using experimental data, Econometrica, 66, 1017-1098
Hillion, P., Vermaelen, T., (2004), Death spiral convertibles, Journal of Financial
Huson, M. R., Parrino, R., Starks, L. T., (2001), Internal monitoring mechanisms and CEO
turnover: A long-term perspective, Journal of Finance, 56, 2265-2297
Jarrell, G. and Poulsen, A., (1988), Dual-class recapitalization as antitakeover mechanisms:
The recent evidence, Journal of Financial Economics, 20, 129-152
Khan, T., (2006), Corporate dividends and ownership structure: Evidence from UK panel
data, The Economic Journal, 116, 172-189
King, M., and Santor, E., (2008), Family values: Ownership structure, performance and
Masulis, R.W., Wang, C, and Xie, F., (2009), Agency problems at dual-class companies,
Morck, R., Shleifer, A., Vishny, R., (1988), Management ownership and market valuation:
An empirical analysis, Journal of Financial Economics, 20, 293-315
Norburn, D., and Birley, S., (1988) The top management team and corporate performance,
Strategic Management Journal, 9, 225-237
Ozkan, A., and Ozkan, N., (2004), Corporate cash holdings: An empirical investigation of
UK companies, Journal of Banking and Finance, 28, 2103-2134
Rosenbaum, P., and Rubin D., (1983), The Central role of the propensity score in
observational studies for causal effects, Biometrika, 70, 41-55.
180
Ruback, R. S., (1988), Coercive dual-class exchange offers, Journal of Financial

Salas, J. M , (2010), Entrenchment, governance and stock price reaction to sudden executive
deaths, Journal of Banking and Finance, 34, 656-666
Shleifer, A., and Vishny R. W., (1989), Management entrenchment: The case of managerspecific investments, Journal of Financial Economics, 25, 123-139
Smith, B. F., Amoako-Adu, B., and Kalimipalli M., (2009), Concentrated control and
corporate value: A comparative analysis of single and dual class structures in Canada,
Applied Financial Economics, 19, 955-974
Stein, J., (1988), Takeover threats and managerial myopia, Journal of Political Economy,
96,61-80
Stulz, R., (1988), Managerial control of voting rights: Financing policies and the market for
corporate control, Journal of Financial Economics, 20, 25-54
Villalonga, B., (2004), Does diversification cause the "diversification discount"? Financial
Warner, J. B., Watts, R. L., and Wruck, K. H., (1988), Stock prices and top management
changes, Journal of Financial Economics, 20,461-492
Weisbach, M. S., (1988), Outside directors and CEO turnover, Journal of Financial
Wilcox, J. C , (2002), Two cheers for staggered boards, Corporate Governance Advisor, 10,
1-5
Yermack, D., (2006), Flights of fancy: Corporate jets, CEO perquisites, and inferior
shareholder returns, Journal of Financial Economics, 80, 211-242
Zingales, L., (1995), What determines the value of corporate votes? The Quarterly Journal
of Economics, 110, 1047-1073
181

Directors' age is the average age of a company's director. Directors' tenure is the median number a company's directors serve in this role. IACEO
Tenure and lADirectors' tenure are defined as the industry adjusted tenure for CEOs and Directors, respectively based on industry median. Eindex is the sum of the 6 entrenching provisions identified by Bebchuk et al. (2009) and G-index is the sum of the 24 governance provisions used
in Gompers et al (2003). Perdum2 is a dummy variable equal to 1 if the firm's 3-year ROA is less than the 3-year industry average ROA. Cash
flow rights is defined as the percentage of equity owned by the largest shareholder or management and directors. Voting rights is the percentage of
votes held by the largest shareholder or management and directors. Management voting leverage (Mgmt. Vote) is the percentage of total votes
controlled by management and directors divided by the percentage of total equity held by management and directors, size (natural logarithm of
sales), financial leverage (total debt divided by total assets), performance (ROA=EBIT divided by total assets and RET is measured as the annual
stock returns), risk (beta is estimated using the CRSP equally weighted index and the previous five year monthly stock returns ), growth is the
geometric mean growth in total assets over the previous five year period, institutional ownership (percentage of shares held by institutional
investors) and family director is the percentage of family members on the board of directors. Significance levels at the 1%, 5% and 10% are
indicated as *, **, and ***, respectively. The test for difference in mean is the t-test and the test for difference in median is the Wilcoxon signedrank test.
Panel A: Entrenchment Characteristics
Mean
Dual Class
Median Std Dev.
Single Class
Mean
Median Std Dev.
Mean test
T-stat
Median test
Z-stat
56.00
CEO Age (years)
55.88
56.00
7.41
55.38
6.91 1.37
1.19
59.33
59.80
-0.31
Directors' Age (years)
59.65
4.68
59.43
3.79 -0.43
9 13***
14.12
6.00
CEO Tenure (years)
10.00
12.13
8.80
8.55 Q Q g * * *
3 go***
Directors' Tenure
8.58
7.62
7.00
4.14 4.14***
8.00
5.00
o oy***
IACEO Tenure
8.28
4.25
12.30
3.08
0.50
8.56 9.68***
1.74
0.00
3.65***
lADirectors' Tenure
1.00
4.25
1.09
3.93 3 13***
_1
1 QQ***
E-index
1.80
3.00
2.00
1.42
2.66
1.28 -12 57***
-7 09***
G-index
6.09
6.80
7.00
6.00
1.75
1.88 -7 69***
-j c c * * *
1.82
1.00
Perdum2* IACEO Tenure
10.45
8.50
12.76
8.68 4.64***
Perdum2*IADirector Tenure
1.74
4.15
1.07
0.50
4.75 2 94***
2.32**
1.50
Note: Only non-zero observations are used to calculate the summary statistics for the interaction terms Perdum2*IACEO Tenure and Perdum2*IADirector
Tenure. For the Perdum2*IACEO Tenure, there are 230 observations for the dual class sample and 157 for the single class sample. For the
Perdum2*IADirector Tenure, there are 200 observations for the dual class sample and 142 for the single class sample. The median test for the interaction
terms is the Wilcoxon rank sum test.
182
Table 1 Cont'd
Panel B: Ownership Characteristics
Mean
Cash Flow Rights of the Largest
Shareholder %
Voting Rights of the Largest
Shareholder %
Cash Flow Rights of Management &
Directors %
Voting Rights of Management &
Directors %
Management Voting Leverage (Mgmt
Vote)
Dual Class
Median Std Dev.
Single Class
Mean
Median Std Dev
Mean test
T-stat
Median test
Z-stat
22 50
18 20
16 50
23 60
19 20
13 20
-0 71
-0 62
57 80
54 80
25 70
23 60
19 20
13 20
34 33***
15 75***
24 90
19 30
15 60
17 10
16 50
14 10
1421***
4 47***
58 30
57 30
25 20
17 10
16 50
14 10
44 25***
15 82***
3 01
2 42
2 15
100
100
0 00
58 60****
82 55***
-3 02***
-2 00**
0 78
-0 58
2 18**
0 86
-146
107
-13 24***
34 83***
-4 53***
-1 98**
-0 89
-0 28
158
-0 04
-163
108
-15 65***
37 86***
Panel C: Firm Characteristics

Tobin's Q ratio
1 87
124
2 04
1 62
125
147
Industry Adjusted Q
-0 59
-0 41
2 26
173
-0 39
-0 35
Size (Sales - $ million)
4917 45
4653 87
1523 90 14597 23
1629 00 15302 38
2161
20 34
2181
20 82
18 96
18 10
Performance - (Ret) %
38 52
12 61
8 35
34 87
10 80
7 45
Performance - (ROA)%
9 75
9 30
8 87
8 81
9 57
9 06
Risk-(Beta)
0 99
0 77
0 81
0 85
0 81
106
Growth % - (Total Asset)
9 98
16 64
6 67
15 61
9 37
6 46
16 92
13 60
15 66
2149
1871
14 96
Family Directors %
15 80
10 60
12 50
12 73
5 62
0 00
Note The number of observations for each of the dual and single class samples is 792 firm-year
Industry Adjusted Q-Mean is defined as the difference between the firm's Q ratio and the average SIC Industry Q ratio
183

single class firms
The dependent variable is computed as the difference between the Q ratio of dual class firms and the
Q ratio of matching single class concentrated control firms. Excess CEO tenure is the difference in
tenure for dual class CEOs and their matching counterparts in single class firms. LACEO tenure is
the difference between tenure of dual class CEOs and the average industry CEO tenure. Excess Eindex is the difference between the E-index value of dual class firms and matching single class firms.
In constructing the excess G-index, I subtract the E-index value for each firm from the G-index value
and then take the difference between dual class firms' G-index and matching single class firms' Gindex. IADirectors' tenure is computed as the median tenure per director less the median industry
tenure per director. The numbers below the estimated coefficients are t-statistics, with ****** being
significant at the 1%, 5% and 10% levels.
Excess CEO Tenure
m
-0.011
-1.82*
IACEO Tenure
(2)
(3)
-0.013
-2 93***
Excess E-index
-0.055
-1.71*
0.05
1.60
Excess G-index
IADirectors' Tenure
Excess Compensation
Mgmt. Vote x Excess Cash
Mgmt. Vote
Excess Cash
Financial Leverage
Conversion Rights
Size
Dividend Difference
CEO Chairman Duality
Intercept
Industry and Year effects
Adjusted R2
Obs.
(4)
-0.229
-3 38***
-0.293
-1.19
-0.109
-3.50***
0.829
1.03
-0.571
-1.53
-0.265
-1.72*
0.143
2 95***
-0.333
-1.32
0.143
1.11
-0.689
-1.53
Yes
0.242
792
-0.224
-3 31***
-0.242
_\ 97**
-0.102
_3 34***
0.736
0.90
-0.872
-3 13***
-0.204
-1.34
0.123
2.60***
-0.334
-1.32
0.148
1.17
-0.643
-1.53
Yes
0.237
792
184
-0.23
_3 4i***
-0.261
-1.05
-0.113
-3.68***
0.771
0.96
-0.938
-3 7 7 * * *
-0.255
-1.64
0.133
9 79***
-0.233
-0.93
0.124
0.99
-0.757
-1.75*
Yes
0.267
792
-0.032
-2.24**
-0.236
-3.50***
-0.241
-1.96**
-0.110
-3.62***
0.741
0.90
-0.875
-3 15***
-0.183
-1.23
0.123
2.54**
-0.296
-1.17
0.105
0.84
-0.577
-1.32
Yes
0.256
792
(5)
-0.012
-2.34**
-0.063
-1.95*
0.05
1.53
-0.019
-2.04**
-0.229
-3 38***
-0.252
-2.02**
-0.108
-3 61***
0.716
0.89
-0.902
-3 24***
-0.215
-1.43
0.107
2.12**
-0.289
-1.15
0.132
1.05
-0.447
-0.96
Yes
0.289
792
Table 3: Effects of entrenchment on discount of dual class Q ratio versus the industry
average
The dependent variable is computed as the difference between the Q ratio of dual class firms and
industry average Q ratio. Excess CEO tenure is the difference in tenure for dual class CEOs and their
matching counterparts in single class firms. IACEO tenure is the difference between tenure of dual
class CEOs and the average industry CEO tenure. Excess E-index is the difference between the Eindex value of dual class firms and matching single class firms. In constructing the excess G-index, I
subtract the E-index value for each firm from the G-index value and then take the difference between
dual class firms' G-index and matching single class firms' G-index. IADirectors' tenure is computed
as the median tenure per director less the median industry tenure per director. The numbers below the
estimated coefficients are t-statistics, with ****** being significant at the 1%, 5% and 10% levels.
Excess CEO Tenure
(1)
-0.022
-1.98**
IACEO Tenure
(2)
(3)
IADirectors' Tenure
Mgmt. Vote
Excess Cash
Financial Leverage
Conversion Rights
Size
Dividend Difference
CEO-Chairman Duality
Intercept
Adjusted R2
Obs.
-0.001
-2.07**
-0.425
-6.52***
-0.248
-1.87*
-0.027
-2 89***
0.786
1.21
-0.628
-2.57**
-0.143
-1.24
0.218
-0.012
-2.27**
-0.035
-1.98**
0.025
1.16
-0.013
-1.96**
-0.413
-6.29***
-0.258
-1.89*
-0.025
-2 83***
0.777
1.21
-0.633
-2.60***
-0.165
-1.41
0.207
-0.028
-1.88**
0.024
1.13
Excess G-index
(5)
-0.009
-2.03**
Excess E-index
Excess Compensation
(4)
-0.425
-6.53***
-0.249
-1.83*
-0.028
_1 99**
0.787
1.21
-0.629
-2.57**
-0.146
-1.27
0.22
4.48***
0.41
2.25**
0.044
0.41
-1.545
-4 29***
Yes
0.144
792
-0.417
_6.44***
-0.246
-2.05**
-0.024
-1.78*
0.768
1.20
-0.609
-2.49**
-0.134
-1.16
0.207
4 j]***
0.363
2.03**
0.063
0.59
-1.382
-3 75***
Yes
0.147
792
185
-0.424
-6.50***
-0.254
-2.25**
-0.029
-2.94***
0.786
1.20
-0.648
-2.66***
-0.16
-1.38
0.215
4.46***
0.425
2.33**
0.046
0.43
-1.492
.4.22***
Yes
0.146
792
T^***
0.403
2.22**
0.046
0.44
-1.519
-3 98***
Yes
0.144
792
414***
0.385
2.12**
0.075
0.69
-1.412
-1 79***
Yes
0.178
792

single class firms' conditional on past performance
Q ratio of matching single class concentrated control firms. Excess CEO tenure is computed as the
difference in tenure for dual class CEOs and their matching counterparts in single class firms.
IACEO tenure is computed as the difference between tenure of dual class CEOs and the average
industry CEO tenure. Excess E-index is the difference between the E-index value of dual class firms
and matching single class firms. In constructing the excess G-index, I subtract the E-index value for
each firm from the G-index value and then take the difference between dual class firms' G-index and
matching single class firms' G-index. lADirectors' tenure is computed as the median tenure per
director less the median industry tenure per director. Perdum 1 is a dummy variable equal to 1 if dual
firms' previous 3-year average ROA is less than the ROA of matching single class firms. (Perdum2
is a dummy variable equal to 1 if the firm's 3-year ROA is less than the 3-year industry average
ROA. The numbers below the estimated coefficients are t-statistics, with ****** being significant
at the 1%, 5% and 10% levels.
(1)
Perdum 1 x Excess CEO Tenure

Excess CEO Tenure
Perdum 1
Perdum2 x IACEO Tenure
IACEO Tenure
Perdum2
(2)
(3)
(4)
5)
-0.035
-3.61***
0.008
1.51
-0.08
-0.58
-0.019
-2.02**
-0.009
-1.79*
0.003
0.02
Perdum2 x Excess CEO Tenure
-0.016
-1.79*
-0.001
-0.21
-0.135
-1.01
Excess CEO Tenure

Perdum2
Perdum 1 x Excess E-index
0.004
0.08
-0.063
-1.53
0.051
1.92*
-0.107
-0.87
Excess E-index
Excess G-index
Perdum 1
Perduml*lADirectors'Tenure
-0.084
-4.05***
0.005
0.32
-0.001
lADirectors' Tenure
Perduml
-0.01
186
Table 4: Cont'd
(2)
-0.236
-3.44***
-0.295
-1.89*
-0.109
-3.44***
-0.288
-1.79*
-0.11
-3.46***
0.787
0.99
-0.043
-1.16
-0.278
-1.81*
0.152
2 n***
-0.315
-1.27
0.162
1.26
-0.773
-1.73*
Yes
(3)
-0.233
-3 43***
-0.304
-1.93*
-0.113
-3 47***
0.795
0.98
-0.053
-1.44
-0.268
-1.75*
0.13
2.68***
-0.339
-1.35
0.14
1.09
-0.576
-1.28
Yes
0.842
1.05
-0.045
-1.21
-0.284
-1.84*
0.144
2 95***
-0.313
-1.24
0.142
1.10
-0.682
-1.50
Yes
-0.227
-0.91
0.13
1.04
-0.733
-1.71*
Yes
(5)
-0.221
-3.28**
-0.268
-1.87*
-0.111
-3.61**
0.802
0.98
-0.079
-2.87**
-0.206
-1.39
0.123
2.55**
-0.273
-1.10
0.089
0.71
-0.533
-1.20
Yes
Adjusted R2
Obs.
0.211
792
0.236
792
0.229
792
0.276
792
0.290
792
Excess Compensation
(1)
-0.211
-3 10***

Mgmt. Vote
Excess Cash
Financial Leverage
Conversion Rights
Size
Dividend Difference
Intercept
187
(4)
-0.233
-3 41***
-0.264
_] 99**
0.770
0.95
-0.091
-3 24***
-0.269
-1.73*
0.134
-0.119
-3 75***
9 7Q***
Table 5: Effects of entrenchment on discount dual class Q ratio versus the industry
conditional on past performance
The dependent variable is computed as the difference between the Q ratio of dual class firms and
industry average Q ratio. Excess CEO tenure is the difference in tenure for dual class CEOs and their
matching counterparts in single class firms. IACEO tenure is the difference between tenure of dual
class CEOs and the average industry CEO tenure. Excess E-index is the difference between the Eindex value of dual class firms and matching single class firms. In constructing the excess G-index, I
subtract the E-index value for each firm from the G-index value and then take the difference between
dual class firms' G-index and matching single class firms' G-index. IADirectors' tenure is computed
as the median tenure per director less the median industry tenure per director. Perduml is a dummy
variable equal to 1 if dual firms' previous 3-year average ROA is less than the ROA of matching
single class firms. Perdum2 is a dummy variable equal to 1 if the firm's 3-year ROA is less than the
3-year industry average ROA. The numbers below the estimated coefficients are t-statistics, with
***, **, * being significant at the 1%, 5% and 10% levels.
Perduml x Excess CEO Tenure

Excess CEO Tenure
Perduml
IACEO Tenure
Perdum2
(1)
-0.033
-3 57***
0.018
2.47**
0.089
0.63
(2)
(3)
(4)
(5)
-0.006
-0.73
0.001
-0.01
0.074
0.51
-0.024
-2 90***
0.012
1.95*
0.052
0.37
Excess CEO Tenure

Perdum2
-0.147
-3.06***
-0.102
-2 89***
0.022
0.99
0.117
0.85
Excess E-index
Excess G-index
Perduml
Perduml x IADirectors'Tenure
-0.093
.531 ***
IADirectors' Tenure
0.039
2.66***
0.117
0.88
Perduml
188
Table 5 Cont'd
Excess Compensation
Mgmt Vote x Excess Cash
Mgmt Vote
Excess Cash
Financial Leverage
(1)
-0 139
-2 07**
-0 321
-194*
-0 033
-1 69*
0 973
121
-0 034
-0 86
-0 208
-170*
0218
(2)
(3)
-0 150
-2 18**
-0 149
-2 18**
-0 135
_1 97**
-0 140
-6 38***
-0 325
-1 96**
-0 037
.j 99**
-0 349
_1 97**
-0 319
-1 30
-0 038
.] 97**
-0 278
-1 78*
-0 022
-0 71
0 859
135
1 006
125
-0 04
-102
-0 206
-1 64
0 205
-0 037
-2 03**
1 052
131
4 i i ***
"2 * 7 0 * * *
0 578
3 08***
0 048
0 43
-1492
-3 72***
Yes
0 571
3 05***
0 035
031
-1353
-3 34***
Yes
-0 037
-0 95
-0 221
-1 78*
0 206
3 92***
0 59
3 15***
0 03
0 27
-1 366
-3 45***
Yes
Adjusted R2
Obs
0 107
792
0 091
792
0 100
792
Conversion Rights
Size
Dividend Difference
Intercept
189
(4)
0 958
1 17
-0 059
-2 36**
-0 183
-1 50
0 206
(5)
0 529
9 77***
0 059
0 55
-1 452
-4 04***
Yes
-0 056
-2 33**
-0 155
-1 34
0218
4 4]***
0 422
234**
0 023
0 22
-1 502
-3 96***
Yes
0 109
792
0 168
792
4 j j***
controlling shareholders-executives
tenure for dual class CEOs and their matching counterparts in single class firms. IACEO tenure is
the difference between tenure of dual class CEOs and the average industry CEO tenure. Excess Eindex is the difference between the E-index value of dual class firms and matching single class firms.
and then take the difference between dual class firms' G-index and matching single class firms' Gindex. IADirectors' tenure is computed as the median tenure per director less the median industry
tenure per director. Perduml is a dummy variable equal to 1 if dual firms' previous 3-year average
ROA is less than the ROA of matching single class firms. (Perdum2 is a dummy variable equal to 1
if the firm's 3-year ROA is less than the 3-year industry average ROA. The numbers below the
estimated coefficients are t-statistics, with ****** being significant at the 1%, 5% and 10% levels.
Perdumlx Excess CEO Tenure

Excess CEO Tenure
Perduml
IACEO Tenure
Perdum2
(1)
-0.033
-3 03***
0.011
1.46
0.123
0.55
(2)
(3)
(4)
(5)
-0.015
-2.13**
-0.010
-0.59
0.159
0.55
-0.021
-2.28**
0.006
0.68
0.058
0.27
Excess CEO Tenure

Perdum2
Perdumlx Excess E-index
0.025
0.35
-0.10
-1.73*
0.045
0.92
-0.029
-0.15
Excess E-index
Excess G-index
Perduml
Perduml x IADirectors'Tenure
-0.103
-3.88***
0.002
0.09
0.201
0.91
IADirectors' Tenure
Perduml
190
Table 6 Cont'd
Excess Compensation
Mgmt. Vote
-0.901
-1.98**
-0.070
-0.402
-1.94
-3 35***
Yes
(2)
-0.133
-1.31
0.043
0.13
-0.049
-1.88*
-0.307
-0.32
-0.063
-1.23
-0.161
-0.73
0.250
3 96***
-0.865
-1.94*
-0.119
-0.69
-1.477
-2.54**
Yes
(3)
-0.264
-2.86***
0.069
0.21
-0.062
-1.40
-0.401
-0.4
-0.03
-0.61
-0.172
-0.75
0.281
4.62***
-0.851
-1.86*
-0.105
-0.61
-1.804
-3 07***
Yes
(4)
-0.245
-2.54**
-0.006
-0.02
-0.065
-2.38**
-0.196
-0.19
-0.048
-0.96
-0.267
-1.05
0.289
4.58***
-0.732
-1.74*
-0.086
-0.51
-1.889
Yes
(5)
-0.137
-1.42
-0.024
-0.07
-0.073
-2.63***
-0.076
-0.08
-0.051
-1.12
-0.152
-0.67
0.281
4 4g***
-0.876
-2.02**
-0.168
-0.99
-1.566
-2.61***
Yes
0.102
486
0.106
486
0.120
486
0.103
486
0.113
486
(1)
-0.238
-2.58**
0.089
0.27
-0.049
. ] gg**
Excess Cash
Financial Leverage
Conversion Rights
Size
-0.456
-0.45
-0.035
-0.71
-0.161
-0.71
0.293
4 79***
Dividend Difference
Intercept
2
Adjusted R
Obs.
191
-3 18***
non-controlling shareholders-executives
tenure for dual class CEOs and their matching counterparts in single class firms IACEO tenure is
the difference between tenure of dual class CEOs and the average industry CEO tenure Excess Eindex is the difference between the E-index value of dual class firms and matching single class firms
and then take the difference between dual class firms' G-index and matching single class firms' Gmdex. IADirectors' tenure is computed as the median tenure per director less the median industry
tenure per director. Perduml is a dummy variable equal to 1 if dual firms' previous 3-year average
ROA is less than the ROA of matching single class firms. (Perdum2 is a dummy variable equal to 1
if the firm's 3-year ROA is less than the 3-year industry average ROA. The numbers below the
estimated coefficients are t-statistics, with ***5**5* being significant at the 1%, 5% and 10% levels.
Perdum 1 x Excess CEO Tenure

Excess CEO Tenure
Perduml
IACEO Tenure
Perdum2
(1)
-0 024
-1 51
0 002
0 02
-0 014
-0 06
(2)
(3)
(4)
(5)
-0 055
-3 57***
0 005
0 52
0161
0 79
0 021
1 33
-0 022
-2 10**
-0 148
-0 69
Excess CEO Tenure

Perdum2
-0 046
-0 55
0 021
0 36
0 009
0 26
-0 046
-0 22
Excess E-index
Excess G-index
Perduml
Perdumlx IADirectors' Tenure
-0 059
-2 08**
-0 038
-1 28
-0 006
-0 03
IADirectors' Tenure
Perduml
192
Table 7 Cont'd
(2)
-0.368
-3 45***
-0.695
-2.62***
-0.078
-3.46***
2.344
2.56**
-0.034
-0.58
0.166
0.83
0.137
1.98**
0.223
0.71
0.569
3 45***
-0.458
-0.75
Yes
(3)
-0.295
-3.02***
-0.641
-2.54**
-0.156
-3 58***
2.326
2.63***
-0.019
-0.30
0.165
0.81
0.160
2.30**
0.093
0.34
0.483
(4)
-0.285
-2.92***
-0.695
-2.48**
-0.088
-3 54***
2.435
2 Q]***
2 14***
(1)
-0.275
-2.88***
-0.684
-2.41**
-0.078
-3.08***
2.467
2.64***
-0.042
-0.66
0.167
0.82
0.158
2.31**
0.215
0.75
0.514
3.16***
-0.6252
-1.04
Yes
-0.407
-0.64
Yes
-0.565
-0.91
Yes
0.422
1.44
0.439
2.64***
-0.233
-0.40
Yes
Adjusted R2
Obs.
0.203
306
0.224
306
0.205
306
0.194
306
0.224
306
Excess Compensation
Mgmt. Vote
Excess Cash
Financial Leverage
Conversion Rights
Size
Dividend Difference
Intercept
193
9 <jQ***
-0.013
-0.20
0.169
0.82
0.152
2.19**
0.258
0.89
0.515
(5)
-0.360
-3 37***
-0.641
-2.47**
-0.087
-3 R^***
2.194
2.50**
-0.048
-0.69
0.234
1.16
0.128
1 07**
Table 8: Effects of predicted entrenchment on dual class discount conditional on past

performance
The dependent variable in models (l)-(3) is computed as the difference between the Q ratio of dual
class firms and the Q ratio of matching single class concentrated control firms and the dependent
variable in models (4)-(6) is the difference between the Q ratio of dual class firms and industry
average Q ratio. Predicted Excess CEO Tenure is the difference in predicted tenure for dual class
CEOs and their matching counterparts in single class firms. Predicted IACEO tenure is predicted
industry adjusted tenure of dual class CEOs. Predicted IADirectors' tenure is the predicted median
industry adjusted tenure per director. Perdum is a dummy variable equal to 1 if dual firms' previous
3-year average ROA is less than the ROA of matching single class firms. The numbers below the
estimated coefficients are t-statistics, with ***, **, * being significant at the 1%, 5% and 10% levels.
(1)
Perdum x Predicted
Excess CEO Tenure
Predicted Excess CEO
Tenure
(2)
(3)
-0.136
-5 70***
-0.100
-5 32***
0.094
0.059
2.23**
0.123
0.93
9 Qg***
Perdum
0.015
0.12
Perdum x Predicted
IACEO Tenure
Perdum
Perdum x Predicted
IADirectors' Tenure
Predicted IADirectors'
Tenure
Perdum
Mgmt. Vote x Excess

Cash
Mgmt. Vote
Excess Cash
Financial Leverage
Conversion Rights
-0.186
-2 70***
-0.183
-2.64***
-0.275
-1.98**
-0.133
-3 96***
0.641
0.83
-0.039
-1.10
-0.323
-2.09**
-0.261
-1.81*
-0.117
-3.46***
0.623
0.80
-0.029
-0.82
-0.323
-2.14**
194
(5)
(6)
-0.073
-5.80***
0.088
2.61***
0.12
0.92
-0.105
-6 13***
0.044
1.15
0.065
0.49
Predicted IACEO Tenure
Excess Compensation
(4)
-0.497
_5 97***
-0.381
-6.23***
-0.524
-1.16
0.057
0.41
-0.198
-2.82***
0.960
2.57**
0.149
1.13
-0.376
-5.82***
-0.223
1 QQ**
-0.170
-3 79***
0.53
0.66
-0.025
-0.72
-0.296
-2.02**
-0.384
-5 95***
-0.382
-5 89***
-0.266
-2.35**
-0.035
-1.01
0.72
1.14
-0.026
-0.72
-0.194
-1.66*
-0.268
-2.37**
-0.054
-1.55
0.703
1.12
-0.027
-0.75
-0.220
-1.89*
-0.303
_| 07**
0.014
0.37
0.826
1.29
-0.026
-0.75
-0.226
-1.91*
Table 8 Cont'd
Size
(1)
0.155
o2^***
Dividend Difference
Intercept
Adjusted R2
Obs.
-0.419
-1.71*
0.165
1.30
-0.95
-2.12**
Yes
0.112
792
(2)
0.148
3.01***
-0.43
-1.76*
0.154
1.21
-0.744
-1.60
Yes
0.124
792
195
(3)
0.144
2 Qg***
-0.44
-1.83*
0.148
1.17
0.733
0.68
Yes
0.128
792
(4)
0.228
4.58***
0.304
1.69*
0.062
0.59
-1.613
-4.04***
Yes
0.163
792
(5)
0.217
4 4]***
0.306
1.72*
0.06
0.57
-1.849
-4 ] ? * * *
Yes
0.169
792
(6)
0.225
4.58***
0.343
1.87*
0.057
0.55
-3.022
-3.60***
Yes
0.173
792
Chapter 5
Contributions and Conclusions
This dissertation makes several contributions to the finance and corporate

governance literature. The first essay adds to several streams of literature. First, it
contributes to the ownership and executive compensation literature by presenting evidence
in support of the extraction of private benefits of control hypothesis. Second, the research
contributes to the study of family companies by presenting evidence that all family
executives, not family CEOs alone, receive higher compensation. Finally, this research helps
to explain why dual class companies sell at a discount compared to single class companies.
In the first essay, I investigate the salient agency problems associated with
concentrated control and the extraction of private benefits. I propose three channels- excess
compensation, excess cash holdings and excess capital expenditure- through which private
benefits can be extracted and relate these channels to the documented dual class discount. I
provide evidence of two channels through which controlling shareholders can extract private
benefits at the expense of minority shareholders. I show that excess compensation and
excess cash holdings result in a larger discount of dual class firms relative to a matching
group of single class companies with concentrated control. This excess compensation is
greatest when executives are family members of the controlling shareholder. The evidence is
consistent with family members extracting perquisite consumption at the expense of
minority shareholders.
196
The second essay examines the relationship between concentrated control and
dividend policy. I proposed three explanations- extraction of private benefits, managerial
reputation and family legacy- for dividend policy in firms with concentrated control. I
provide evidence that, even in the U.S. with its stringent investor protection, controlling
shareholders of dual class firms are extracting private benefits and hence, instituting a lower
payout policy. Controlling shareholders of dual class firms, because they control the firm
with votes rather than equity, receive only a small fraction of the corporate distribution.
However, they can extract full benefits from assets retained within the firm. The results of
the second essay complement the findings of the first essay.
This essay adds to our
understanding of dual class share structure and why investors value these firms lower than
similar single class companies.
Private benefits and the agency costs of dual class share structure are not always
tangible. Dual class ownership structure can allow managers to remain on the job even when
the company is performing poorly relative to its peers. Therefore, managerial entrenchment
is a form of private benefit flowing from dual class share structure. In essay 3,1 contribute to
the corporate governance and finance literature by investigating the relationship between the
documented dual class discount and entrenchment. Dual class share structure is an effective
anti-takeover defense mechanism that can render the market for corporate control
ineffective. Hence, managers of dual class firms can become entrenched. Investors are likely
to apply a greater discount to the value of dual class firms which appear to have a higher
degree of managerial entrenchment. Using several proxies for entrenchment, I show that
dual class firms with poor past performance and a greater degree of managerial
entrenchment have a larger discount.
197
In summary, this research contributes to the corporate governance and finance

literature by providing evidence that dual class share structure allows controlling
shareholders and their family members to extract private benefits at the expense of outside
shareholders, institute lower dividend policy which benefits themselves and allow managers
to become entrenched. Investors are aware of the perquisite consumption flowing from
control established by dual class share structure and hence, discount the value of dual class
firms.
198

Three Essays in Corporate Governance

Uploaded by

Copyright:

Available Formats

Three Essays in Corporate Governance

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Three Essays in Corporate Governance

Uploaded by

Copyright:

Available Formats

What are the three essays about?

What are the three essays about?

What are the three proposed explanations for dividend policy in firms with concentrated control?

What are the three proposed explanations for dividend policy in firms with concentrated control?

Wilfrid Laurier University

Scholars Commons @ Laurier

Three Essays in Corporate Governance

Follow this and additional works at: http://scholars.wlu.ca/etd

Library and Archives

395 Wellington Street

395, rue Wellington

The author has granted a nonexclusive license allowing Library and

L'auteur a accorde une licence non exclusive

The author retains copyright

L'auteur conserve la propriete du droit d'auteur

In compliance with the Canadian

Conformement a la loi canadienne sur la

While these forms may be included

Bien que ces formulaires aient inclus dans

THREE ESSAYS IN CORPORATE GOVERNANCE

Vishaal Rabindranauth Anand Baulkaran

Vishaal Baulkaran 2010

I would like to express my sincere gratitude to my dissertation committee chair,

September 16, 1979

Born - New Amsterdam, Berbice, Guyana

Bachelor of Business Administration, University of New Brunswick, Canada

Master of Business Administration, University of New Brunswick, Canada

Major Field: Business Administration

Conceptual Framework of Corporate Governance

2.0 CORPORATE GOVERNANCE MECHANISMS

Laws and Regulations

BRIEF HISTORY OF THE DUAL CLASS MARKET IN THE UNITED STATES

FURTHER AREAS OF RESEARCH AND UNRESOLVED QUESTIONS

DUAL CLASS DISCOUNT

LITERATURE REVIEW AND HYPOTHESIS DEVELOPMENT

METHODOLOGY AND DATA

CHAPTER SUMMARY AND CONCLUSIONS

AGENCY THEORY, PRIVATE BENEFITS AND DIVDDEND POLICY

Managerial Reputation Explanation for Dividend Policy

WHY STUDY DUAL CLASS SHARE STRUCTURES OF FffiMS IN THE U.S.?

HYPOTHESES, METHODOLOGY AND DATA

CHAPTER SUMMARY AND CONCLUSIONS

LITERATURE REVIEW AND HYPOTHESIS DEVELOPMENT

CHAPTER SUMMARY AND CONCLUSIONS

Table 2: Number of dual class firms by two-digit SIC Industry code

Table 3: Summary statistics of governance characteristics, firm characteristics

Table 4: Regression of executive compensation on management voting leverage,

Table 5: Panel regression of executive compensation on family executives, management voting

Table 6: Panel regression for a sub-sample of family executives

Table 8: Panel regression of dual class discount on capital expenditure

Table 9: Fixed effects regression of dual class discount on excess compensation

Table 12: Panel regression of total compensation on management voting leverage,

Table 2: Regression of dividend payout on management voting leverage

Table 3: Regression of total distribution

Table 4: Descriptive statistics for contronlling and non-controlling shareholdersexecutives

Table 5: Tobit regression of dividend payout ratios

Table 6: Tobit regression of total distribution

Table 7: Tobit regression for a sub-sample of family firms only

Table 10: Tobit regression with predicted management voting leverage

Table 2: Effects of entrenchment on discount of dual class Q ratio to that of matching