Electronic copy available at: http://ssrn.com/abstract=1497205 Electronic copy available at: http://ssrn.

com/abstract=1497205



Deviations from Expected Stakeholder Management, Firm
Value, and Corporate Governance


Bradley W. Benson*
Department of Economics and Finance
Louisiana Tech University
Ruston, LA 71270
318-257-2389
[email protected]

Wallace N. Davidson III
Finance Department Mailcode 4626
Southern Illinois University
Carbondale, IL 62901
618-453-1429
[email protected]

Hongxia Wang
Institute for Contemporary Financial Studies
College of Business and Economics
Ashland University
Ashland, OH 44805
419-289-5222
[email protected]

Dan L. Worrell
Dean and Sam Walton Leadership Chair
Sam Walton College of Business
University of Arkansas
Fayetteville, AR 72701-1201
479-575-5949
[email protected]



August 12, 2010

* Corresponding Author


The authors would like to thank seminar participants at the University of Mississippi for many
helpful and insightful comments. All errors remain the responsibility of the authors.
Electronic copy available at: http://ssrn.com/abstract=1497205 Electronic copy available at: http://ssrn.com/abstract=1497205
1


Deviations from Expected Stakeholder Management, Firm
Value, and Corporate Governance


Abstract

In this paper, we examine the relation between deviations from expected investment in
stakeholder management and corporate governance. Building good relations with various
stakeholders may help create firm value, but investment in stakeholder management beyond what
is necessary to create shareholder value may be an agency cost. We propose that high quality
corporate governance may mitigate value-destroying investments in stakeholder management.
Using an unbalanced panel of 9,051firm-year observations for 1,631 firms, we find that
deviations from expected SM are increasing in CEO portfolio delta. We find, however, that
deviations from expected stakeholder management are negatively related to proxies for effective
board monitoring. More independent boards effectively control deviations from expected
investment in stakeholder management. These results are consistent with both the CEO perquisite
and the board monitoring hypotheses. We also document that the effect of governance
mechanisms varies by industry (consumer or industrial orientation) and SM dimension; their
influence is similar in dimensions expected to provide similar benefits across industry orientation,
but differs when the benefits are expected to benefit one industry more than the other. Consistent
with Jensens Enlightened Value Maximization theory, the results show that corporations with
good governance pursue shareholder value maximization while constraining unnecessary
investment in stakeholders.

JEL classification: G34; J33; L21; M14; M52

Keywords: Corporate Governance; Corporate Social Responsibility; Enlightened Value
Maximization; Institutional Ownership; Firm Value; Managerial Ownership Incentives;
Stakeholder Theory

2

Deviations from Expected Stakeholder Management, Firm
Value, and Corporate Governance
1. Introduction
Proponents of stakeholder management argue that a firm should manage the relations with
all of its stakeholders rather than focus on shareholder wealth. Researchers in a wide range of
fields have examined the implications of stakeholder management (e.g. strategic management,
organizational behavior, business ethics, and sustainable development) (Laplume, Sonpar, and
Litz, 2008). The corporate scandals of the early 2000s have brought corporate social performance
and stakeholder management to the forefront (Neubaum and Zahra, 2006)
1
. Furthermore,
constituency statutes have been enacted in over half of US states in recent years that either require
or permit directors to consider the interests of non-shareholders when conducting their duties
(Keay, 2009).
2

The principal argument of stakeholder management (SM) is that organizations should be
operated and managed in the interests of all their constituents who can affect or be affected by the
achievement of the organizations objectives (Freeman, 1984; Donaldson and Preston, 1995).
Stakeholder management captures various firm-stakeholder relationships. Proponents maintain
that stakeholder management is strategically important, and firms can benefit from properly
managing the relationship with these important groups (Fong, 2009; Bhattacharya, Korschun, and

1
Stakeholder management and social performance are similar in their construct such that firms which manage
stakeholder relations may be seen as socially responsible.
2
The legal duty of directors to consider the interest of additional stakeholders has also taken hold in the UK. For
example, section 172 of the Companies Act 2006 states that directors have a duty to promote the success of the
company by considering such factors as: (a) the likely consequences of any long term decision, (b) the interests of the
companys employees, (c) the need to foster the business relationships of the company with suppliers, customers and
other organizations, (d) the impact of the companys operations on the community and the environment, (e) the
desirability of the company maintaining a reputation for high standards of business conduct, and (f) the need to act
fairly between members of the company.
3

Sen, 2009). Empirical studies, however, have documented that only some stakeholders are critical
to firm value creation (such as Galbreath, 2006; Hillman and Keim, 2001).
One of the downsides to investing in stakeholder management is that it consumes a firms
limited resources. Directing excess resources from shareholders to other stakeholders may hurt
firm value. Some shareholders view stakeholder management as being at odds with profit making
and value maximization. For example, two shareholders at Goldman Sachs protested the
companys donation of land for a nature preserve. The shareholders based their protest on the
idea that while the donation would have benefitted one of the firms stakeholders (the world
environment), it would be costly to shareholders because the land could have been put to
profitable use (Kelly and Davis, 2007).
Stakeholder theory seems to be at odds with value maximization. Jensen (2001; 2002)
addresses this problem and proposes enlightened value maximization. He argues that a firms goal
is to increase firm value, but a firm cannot maximize its value unless it takes care of its
stakeholders. Jensens theory implies that there is an optimal level of stakeholder management
that maximizes shareholder wealth. Similarly, Brickley, Smith, and Zimmerman (2002) argue
that creating shareholder wealth involves allocating resources to all constituencies that affect the
process of shareholder value creation, but only to the point at which the benefits from such
expenditures do not exceed their additional costs (p.113). That is, when the marginal cost of
stakeholder management exceeds the marginal benefit to shareholders, it would not be optimal to
pursue stakeholder management any further. Excess stakeholder management would be
investments in stakeholders beyond this optimal level. Benson and Davidson (2009) find that
while stakeholder management is positively related to firm value, firms do not compensate
managers for stakeholder management; instead, firms compensate their CEOs for achieving the
4

firms ultimate goalvalue maximization. Their results indicate that effective managers optimize
relations with stakeholders to accomplish value maximization.
When firms allocate excess resources to stakeholders, this can be considered an agency
cost and shareholder wealth would suffer. As with other agency conflicts, there are various
governance mechanisms that should protect shareholders, and therefore, the quality of corporate
governance may affect a firms stakeholder management policy.
In this paper, we examine how various corporate governance mechanisms affect the
amount of resources diverted toward stakeholders. We propose that firms will invest in
stakeholder management (SM) until the marginal cost of doing so exceeds the marginal benefit to
shareholders. This level of investment in stakeholder management is each firms expected
investment in stakeholder management (ESM). Good corporate governance, proxied by
managerial ownership incentives, board monitoring, and monitoring from institutional investors,
should control deviations from this expected level of stakeholder management.
Using an unbalanced panel of 9,051 firm-year observations for 1,631 firms, we find that
deviations from expected SM are increasing in CEO portfolio delta (scaled wealth-performance
sensitivity), but they are negatively related to proxies for effective board monitoring. More
independent boards effectively control deviations from expected investment in stakeholder
management. These results are consistent with both the CEO perquisite and the board monitoring
hypotheses. We also document that the effect of governance mechanisms varies by industry
(consumer or industrial orientation) and SM dimension; their influence is similar in dimensions
expected to provide similar benefits across industries, but differs when the benefits are expected
to benefit one industry more than the other.
5

This study is unique. To the best of our knowledge, we are the first to introduce the
concept of deviation from expected investment in stakeholder management and empirically test
how various governance mechanisms are related to unexpected investment in stakeholder
management. We organize the remainder of this paper as follows. In Sections 2 and 3, we
motivate our research, review the literature, and develop our hypotheses. We discuss our data and
sample in Section 4. In Section 5, we test our hypotheses on the effect of internal governance
mechanisms on excess stakeholder management. We provide concluding remarks in Section 6.
2. Background and motivation
2.1. Stakeholder theory
Under stakeholder theory, the firm is a collection of groups and individuals with a stake in
the firm. The purpose of the firm is to manage the interests of these various stakeholders.
Stakeholder theory has diverged into at least two approaches. The first is the strategic approach
in which firms manage stakeholder relations in the pursuit of value maximization. Here, value
maximization is the fundamental purpose of the firm with stakeholder management as a means to
an end (Freeman et al., 2004). The second approach is the moral approach where firms manage
stakeholder relations for ethical or moral reasons (Freeman, 1984; Evan and Freeman, 1988).
3

One issue in stakeholder theory is the identification of stakeholders. Freeman (1984, p. 46)
is credited with the classic definition in which a stakeholder is "any group or individual who can
affect or is affected by the achievement of the organization's objectives". In a literature review,
Friedman and Miles (2006) find 55 definitions of what constitutes a stakeholder. Stakeholder

3
Donaldson and Preston (1995) define these approaches as instrumental and normative stakeholder theories.
Instrumental stakeholder theory is concerned with how managers act if they are to further goals of the organization,
such as long-run profit maximization or shareholder value. On the other hand, normative stakeholder theory is used to
explain how managers should behave and arrive at the purpose of the organization based on ethical principles.
Freeman (1994; 1999) maintains that business and ethical decisions cannot be separated.
6

definitions range from those who have a vested interest in the firm (Carroll, 1993) to any entity
that affects or is affected by the firm (Starik, 1993).
Another issue in stakeholder theory/management is the allocation of resources between
stakeholders. Clarkson (1995) argues that all stakeholders have intrinsic value and no stakeholder
is more important than another. However, if managers treat all stakeholders with the same
priority, it would be difficult to manage a firm because various stakeholders have different
interests. For example, suppliers would want to receive high prices, and customers would want
low prices; accomplishing these goals would be unlikely to produce value for another stakeholder
group, shareholders (Sundaram and Inkpen, 2004).
2.2. Value Maximization
Milton Friedman stated that the social responsibility of business is to increase its profits
(Friedman, 1970)
4
. Brickley et al. (2002) argue that a company must focus its attention on
shareholder wealth to survive in a competitive and technology-oriented business world.
Shareholder wealth maximization has been taught in colleges as a basic concept in finance
courses for decades and is a dominating view of many if not most financial scholars. One of the
major differences between shareholder wealth maximization and stakeholder theories is the role
of managers in the resolution of a firms internal conflicts (Laurent, 2007). Shareholder value
theory emphasizes that the role of managers is to make decisions that enhance shareholder value;
this differs from normative stakeholder theory which maintains that managers should focus on the
welfare of all stakeholders (Laurent, 2007). Proponents of value maximization believe that other
constituencies can become better off if companies pursue shareholders interests. Jensen (2001;

4
Hillman and Keim (2001) argue that social responsibility and stakeholder management are closely related and that
social responsibility is a measure of how well a firm manages the relations with its stakeholders. Graves and
Waddock (2000) state that social performance can be defined as stakeholder relations. The groups identified in social
responsibility discussions are often the same groups identified as stakeholders. As such, corporate social
responsibility and stakeholder theory may be directly linked and are at a minimum based on a similar foundation.
7

2002) argues that a business cannot maximize wealth simply by stating this as its goal; it requires
working with stakeholders to achieve this objective. So, simply focusing on shareholder value
maximization may be like putting blinders on a horse, and it may not be applicable in todays
business world.
5

3. Enlightened value maximization and development of hypotheses
Jensen (2001; 2002) argues that while the goal of the corporation is to maximize
shareholder wealth, this goal cannot be met by treating stakeholders poorly. Companies should
not try to maximize the welfare of all stakeholders, but work with stakeholders to produce
shareholder wealth. More specifically, companies should improve stakeholder welfare until the
marginal cost of doing so exceeds the marginal benefit to shareholders. Thus, Jensens approach
is somewhat consistent with the management of stakeholders as a means to an end where the end
is shareholder value maximization
6
.
There are at least two questions raised by this logic. The first question is: how would
actions aimed at improving the welfare of stakeholders contribute to shareholder wealth; and the
second is: does stakeholder management actually improve company financial performance and
increase shareholder wealth.
To address the first question, we must realize that the effect of stakeholder management
on shareholder wealth could be bi-directional. There may be rewards for positive actions, as well
as penalties for irresponsible ones. Companies that build better relations with primary

5
In the recent business downturn and stock market drop, the press has increasingly focused on business social
responsibility and the relations between business and stakeholders. Value maximization, as the sole firm goal, is
viewed very negatively in the popular press and even in the financial press.
6
Instrumental stakeholder theory posits that stakeholder management will lead to better financial performance.
Jensens (2001; 2002) enlightened value maximization does differ from instrumental stakeholder theory. Jensen
argues that stakeholder theory fails to provide a mechanism for making the tradeoffs between competing stakeholder
claims. Enlightened value maximization implies that managers will make tradeoffs between stakeholders using the
effect on firm value as the decision criteria.
8

stakeholders such as employees, customers, suppliers, and communities may be able to increase
their financial gains (Freeman, 1984). Prior research has documented these rewards (Turban and
Greening, 1996; Greening and Turban, 2000; De Luque, Washburn, Waldman, and House, 2008).
Godfrey (2005) argues that corporate social responsibility and philanthropic activities can
generate positive moral capital among stakeholders which may provide an insurance-like effect
for firm reputation if social problems occur. Companies that mistreat stakeholders may face
penalties. Researchers have documented these penalties (Davidson and Worrell, 1988; Karpoff
and Lott, 1993; Cohen, Fenn, and Naiman, 1995; Davidson, Worrell, and El-Jelly, 1995; Hart and
Ahuja, 1996; Karpoff, Lott, and Rankine, 1998; Karpoff, Lee and Vendrzyk, 1999; Klassen and
Whybark, 1999; Konar and Cohen, 2001; Davidson, Worrell, and El Jelly, 2000).
While stakeholder theory posits that stakeholder management may generate increases in
income or shareholder wealth, companies that spend resources directed at stakeholder welfare, in
excess of the marginal benefits, may find that income and/or shareholder wealth decreases
(Hillman and Keim, 2001). Therefore, whether companies actually increase or decrease
shareholder wealth by managing stakeholder relations is an empirical question.
7

Jensens theory of enlightened value maximization suggests that value maximization
cannot be achieved if companies ignore or mistreat their stakeholders. Others have made similar
propositions (Graves and Waddock, 2000; Davis, 2005). Jensens model differs from traditional
instrumental stakeholder theory because his model provides a basis for making tradeoffs between
stakeholder groups. The companys relations with its stakeholders must be rooted in the

7
Numerous researchers have addressed this question with equivocal findings. Many have found a positive relation
between stakeholder management and financial performance (Berrone, Surroca, and Trib, 2007; Barbara, Myron,
and Bruce, 2006; Hillman and Keim, 2001; Moore, 2001; Ogden and Watson, 1999; Ruf, Muralidhar, Brown, Janney,
and Paul, 2001; Waddock and Graves, 1997; Moneva, Rivera-Lirio, and Munoz-Torres, 2007) while others have
found a negative or a mixed relation (Meznar, Nigh and Kwok, 1994; Berman et al., 1999; Omran, Atrill, and
Pointon, 2002; Brammer, Brooks, and Pavelin , 2006; and Bird, D. Hall, Moment, and Reggiani, 2007).

9

companys strategies and must be a means to an endvalue maximization. Accordingly,
managers optimize relations with stakeholders to maximize firm value. Managing stakeholder
welfare may be a partial determinant of firm value. However, deviations from expected
investment in stakeholder management may be detrimental to a firms value.
3.1. Deviations from expected stakeholder management as an agency cost
Once an owner-manager owns less than 100 percent of the corporation, the costs of
providing the manager with pecuniary and non-pecuniary benefits are borne, in part, by
shareholders (Jensen and Meckling, 1976). These costs are called agency costs. If management
spends resources on managing stakeholder relations and the expenditures do not lead to increased
firm value, then these expenditures would be an agency cost. What incentive would management
have to spend excess resources on stakeholders?
Cespa and Cestone (2007) argue that stakeholder management can become a powerful
entrenchment strategy for incumbent CEOs because inefficient managers may commit themselves
to socially responsible behavior to gain stakeholders support against shareholders. Firms have
limited resources, and there are conflicts between shareholders and other constituents over these
resources. Managers must solve these conflicts. Diverting resources from shareholders may either
hurt shareholders or benefit shareholders in the long run. Proponents of the strict shareholder
value maximization philosophy argue that philanthropic activities and maintaining stakeholder
welfare would waste company resources that would be better spent to produce value for
shareholders. However, diverting some resources in the management of stakeholder welfare may
produce long-run value maximization (Jensen, 2001, 2002; Brickley et al., 2002). The question
may not be whether a firm should invest in stakeholder management, but how much should be
invested in the pursuit of shareholder wealth maximization. Finding the optimal level of
10

investment likely challenges all firms, but may be easier when there is high quality corporate
governance.
3.2. Internal governance mechanisms and agency costs
3.2.1. Managerial ownership incentives and agency costs
From a theoretical standpoint, higher pay-performance sensitivity should result in a
convexity payoff that encourages managers to work hard and take on value enhancing risky
projects (Jensen and Meckling, 1996; Myers 1977; Haugen and Senbet, 1981). Accordingly,
increasing the sensitivity of CEO wealth to stock price should better align managerial behavior
with shareholder interests. An unintended effect of high managerial ownership is that it may
expose managers to more risk relative to diversified shareholders. Here, the concavity of the
managers utility function may overcome the convexity of the payoff making managers more risk
adverse (Guay, 1999; Ross, 2004). High managerial ownership may also increase the risk of
managerial perquisite consumption (Baker and Hall, 2004; Edmans, Gabaix, and Landier, 2009).
Despite significant research on this issue, there is substantial debate over the theoretical and
empirical effect of higher pay-performance sensitivity on financial performance and its effect in
deterring various agency-related issues.
8,9


8
For example, studies find a positive relation between managerial ownership and firm performance (Mehran, 1995;
Core and Larker, 2002), or a positive but decreasing relation between managerial ownership and firm performance
(Morck, Schleifer, and Vishney, 1988; McConnell and Servaes 1990, 1995; Hermalin and Weisbach, 1991; Hubbard
and Palia, 1995; Holderness, Krosner, and Sheehan, 1999; Anderson and Reeb, 2003; Tian, 2004; Davies, Hillier, and
McColgan, 2005; Adams and Santos, 2006; Pukthuanthong, and Roll, Walker, 2007; McConnell, Servaes, and Lins,
2008; Tong, 2008; Benson and Davidson, 2009). Others find no relation between managerial ownership and firm
performance (Demsetz, 1983; Demsetz and Lehn, 1985; Agrawal and Knoeber, 1996; Lorderer and Martin, 1997;
Cho, 1998; Demsetz and Villalonga, 2001; Himmelberg, Hubbard, and Palia, 1999; Palia, 2001; Coles, Lemmon, and
Meschke, 2003; Brick, Palia, and Wang, 2005; and Cheung and Wei, 2006).
9
The remainder of this section focuses on the effects of increased sensitivity of wealth to stock price, delta. However,
it is important to note that increasing the sensitivity of CEO wealth to stock return volatility, vega, is seen as one
potential mechanism to offset managerial risk aversion. Coles, Daniel, and Naveen (2006) provide evidence that
managers with greater convexity in their contracts engage in riskier investment. However, theoretical and empirical
studies suggest that increasing convexity does not necessarily increase managerial risk taking (Lambert, Larker, and
11

At the heart of the debate is the empirical puzzle of a negative and significant relation
between the CEOs effective ownership percentage (dollar-dollar incentives), b, and firm size
(Demsetz and Lehn, 1985; Jensen and Murphy, 1990; Gibbons and Murphy, 1992; Garen, 1994;
Hadlock and Lumer 1997; Hall and Liebman, 1998; Schaefer 1998; Murphy, 1999; Jin, 2002;
Baker and Hall, 2004; Cichello, 2005; Benson and Davidson, 2009; and Edmans et al. 2009).
Nonetheless, the concept that CEOs of large companies have trivial incentives relative to those of
small companies seems inconsistent with the observation of the very large wealth swings induced
by substantial stock and option holdings of large-company CEOs (Hall and Liebman, 1998; Baker
and Hall, 2004; Core and Guay, 2010). Baker and Hall (2004), however, note that a key
assumption made by researchers examining this puzzle is the marginal product of CEO action on
firm value, . Viewed from two polar extremes, if total incentives (i.e. the marginal return to
effort) are equal to b*, this suggests that either = 0 and is independent of firm size or = 1 and
scales proportionally with firm size. The former assumption is best modeled using dollar-dollar
incentives, b. Dollar-dollar incentives, consequently, are most effective at deterring agency costs
that have a dollar effect on firm value (i.e. perquisite consumption). The later assumption,
however, is best modeled using dollar-percent incentives. This implies that dollar-percent
incentives are most effective at deterring agency costs that scale with firm size (i.e. corporate
strategy) and their model predicts that for tasks whose marginal products scale with firm size,
optimally low b in large firms should not incur significant agency costs. However, given the
infeasibility of high b in large firms, optimally low b will also induce significant agency costs for
those tasks whose marginal products do not scale with firm size. Therefore, some agency costs are

Verrecchia, 1991; Guay, 1999; Carpenter, 2000; Ju, Leland, and Senbet, 2002; Ross, 2004; Tian, 2005; Parrino,
Poteshman, and Weisbach, 2005; Lewellen, 2006; and Hayes, Lemmon, and Qiu, 2010). Furthermore, it is not
entirely clear how increasing managerial sensitivity to risk would influence deviations from expected SM.
12

best deterred by active board monitoring and bureaucratic control systems rather than managerial
incentives.
Edmans et al. (2009) propose a multiplicative specification for the CEOs utility and
production function. Their model predicts that the dollar increase in firm value has size elasticity
of 1, or is proportional to firm size. They measure CEO incentives using percent-percent
incentives, the dollar change in wealth for a one-percent change in firm value, divided by total
annual pay (i.e. scaled wealth-performance sensitivity).
10
Dollar-dollar incentives equal percent-
percent incentives multiplied by CEO wage and divided by firm value. Even a small b can induce
high levels of CEO effort since firm value is substantially greater than the CEOs wage, making
the dollar gains from effort larger than the dollar costs. A key implication of the Edmans et al.
(2009) model, however, is that equity based incentives are only effective at addressing agency
costs that have a multiplicative effect on firm value, or actions that are proportional to firm value,
such as corporate strategy. In contrast, CEO incentives cannot solve additive value-destructive
actions, such as perquisite consumption, as these actions have a minimal effect on stock price and
consequently, CEO wealth. Perks, instead, are best controlled by corporate governance. As the
authors suggest, active monitoring [via corporate governance] and incentives should be used in
tandem; the former to deter additive value-destructive actions, and the latter to encourage
multiplicative value-enhancing efforts (p. 4902).
Our testable hypotheses, therefore, assume that firms will choose CEO equity-based
ownership incentives, delta, to implement value-maximizing decisions. If CEO incentives are in
alignment with shareholder interests, deviations from expected investment in stakeholder

10
Edmans et al. (2009) also demonstrate that this measure has several empirical advantages over previous measures,
including size and firm risk insensitivity.
13

management should be negatively related to CEO portfolio delta. As a result, the managerial
ownership incentive hypothesis states:

H
1
: Deviation from expected stakeholder management is negatively related to CEO
portfolio delta.

The empirical implication is that higher delta deters agency costs related to activities that
scale firm size, such as corporate strategy. If investment in stakeholder management is necessary
for maximizing firm value from a strategic standpoint, then we would then expect that deviations
from expected stakeholder management are negatively related to CEO portfolio delta.
However, it is also possible that managers may undertake some investment in stakeholder
management for personal interests. For example, non-strategic investments in stakeholder
management may benefit the CEO by increasing job security (Cespa and Cestone, 2007) or
support moral or personal beliefs incongruent with those of the firm. As a result, deviations from
expected investment in stakeholder management may also represent perquisite consumption by
the CEO. As a result, the managerial perquisite hypothesis states:
H
2
: Deviation from expected stakeholder management is positively related to CEO
portfolio delta.

The empirical implication of this hypothesis is that while a higher delta may induce
greater alignment between CEO incentives and investment in stakeholder management related to
corporate strategy, CEOs may invest in superfluous stakeholder management for personal reasons
as a form of perquisite consumption. These investment are likely to be more additive in nature,
and will have a minimal effect on stock price. While they will be costly in dollar terms to the
firm, they will not appreciably affect CEO wealth. We therefore expect that deviations from
expected stakeholder management may also be increasing in CEO portfolio delta.
14

3.2.2. Board monitoring and agency costs
Executive perquisites have attracted a lot of attention and criticism from the public and
regulators, especially in the current financial crisis
11,12
. There are two competing views on
executive perquisites: the outcome of an optimal employment contract, designed to motivate
executives to work harder and enhance productivity (Fama, 1980), and an agency problem leading
to reduced firm value either directly or indirectly (Jensen and Meckling 1976). Empirical studies
have documented mixed evidence. For example, Yermack (2006) find that firms which disclose
the private use of corporate jets witness negative market reactions, and they tend to underperform
their peers with no private jet use disclosures. Similarly, Chen, Chen, and Hui (2009) find that
family ownership reduces managerial perquisite consumption and stock prices drop around the
disclosure of a descendent CEOs perk consumption and Andrews, Linn, and Yi (2009) find that
firms that hid large amounts of CEO perquisites experienced a negative market reaction after their
proxy statements were released. On the other hand, Rajan and Wulf (2006) provide evidence that
executive perks can increase managerial productivity.
Both theories, however, are consistent on one point; weak corporate governance and
executive personal tastes affect perquisite awards and consumption. Yermack (2006) documents
that variables related to CEO characteristics have significant explanatory power for executives
personal use of corporate jets. Andrews, Linn, and Yi (2009) find that firms with weak corporate
governance are more likely to award perquisites to executives. These results imply that

11
For example, there are a lot of negative reports criticizing the CEOs of the big three auto makers who flew to
Washington in their luxury corporate jets to request 25-billion-dollar taxpayer bailout money on November 19, 2008,
and the 1.22 million dollar office refurbishing expenses of the Merrill Lynch CEO in early 2008 while the securities
firm was cutting thousands of jobs, and etc.
12
To improve the transparency of disclosure of executive perquisite consumption to investors, the SEC tightened up
its disclosure requirements. The new rule not only lowers the minimum threshold amount for reporting from $50,000
to $10,000, but also provides interpretive guidelines on the classification of perquisite. There is loophole in this rule
because it does not specify clearly between an expense and a perquisite. We cannot completely rule out the possibility
that executives report some perks, for instance, investment in stakeholder management because of their own personal
interests, as normal business expenses.
15

executives personal tastes may play a significant role in stakeholder management; executives
may tend to invest in stakeholder management unrelated to firm value; and ineffective governance
may not be able to deter but encourage this kind of stakeholder management behavior. As
indicated earlier, this kind of behavior may become more additive. The rationale is simple. If
CEOs can spend corporate resources for their personal use (e.g. luxury office, traveling), we can
reasonably argue that they may invest in stakeholder management based on their personal
interests. This investment would simply be another form of perquisite. This type of investment
may be more harmful to a firm because of the hidden nature of this type of perquisites.
A boards attitude towards various stakeholders is likely to become an important
determinant of its stakeholder policy. In addition, many state statutes have challenged the
traditional view of the shareholder-only fiduciary duty of the board and specified that boards have
the right to take the interests of all stakeholders into account (McDonnell, 2004). Under this
pressure, boards may be more likely to apply enlightened stakeholder management in dealing
with stakeholders while maximizing shareholder wealth. So, effective monitoring by boards may
mitigate agency problems related to the amount of resources diverted to other stakeholders.
Board composition may affect the boards management policy. Regulators and academics
believe that outside directors are generally more effective monitors than inside directors. The
Sarbanes-Oxley Act and the exchange rules in 2002 require that the majority of the board be
independent. Numerous studies link the proportion of outside directors to financial performance
and shareholder wealth (e.g. Rosenstein and Wyatt, 1990; Byrd and Hickman, 1992; Brickley,
Coles, and Terry, 1994; Cotter, Shivdasani, and Zenner, 1997). Outside dominated boards may
be less likely to provide other stakeholders with excess resources, and they would be more likely
16

to optimize the level of stakeholder management. In other words, outside dominated boards
should be in a better position to control unwarranted investment in stakeholder management.
13

Board size is another major characteristic of corporate boards (Lipton and Lorsch, 1992;
Jensen, 1993; Yermack, 1996). Even though there is debate regarding the effect of board size on
its decisions, the commonly accepted idea is that small boards are better than large ones because
large boards place a greater emphasis on politeness and courtesy, making it easier for CEOs to
control. Empirical studies support this idea (e.g. Yermack, 1996; Eisenberg, Sundgren, and Wells,
1998). If small boards are more effective in monitoring management, small boards may be more
likely to control superfluous resources diverted to stakeholder management.
Thus, more effective monitoring by the board is expected to reduce agency costs between
managers and shareholders by preventing managers from making unnecessary investment in
stakeholder management as a potential means of perquisite consumption. Therefore, the board
monitoring hypothesis is as follows:
H
3
: Deviation from expected stakeholder management is negatively related to proxies for
effective board monitoring (i.e. smaller board size and more independent directors).

The empirical implication of this hypothesis is that higher delta may increase agency costs
between managers and shareholders related to investment in stakeholder management as a form of
perquisite consumption. This type of additive value-destructive investment is best deterred by
active monitoring by the board of directors. We therefore expect that deviations from expected
stakeholder management may be decreasing in proxies for effective board monitoring.

13
Other research has related board composition to stakeholder management. Johnson and Greening (1999) find that
outside director dominated boards tend to pursue greater stakeholder management. Similarly, Kassinis and Vafeas
(2002) find that the likelihood of becoming an environmental lawsuit defendant decreases with the number of outside
directors.
17

3.2.3. The Equivocal Effect of Duality
When the CEO also holds the title of board chair, CEO duality, power may concentrate in
the CEOs position. Duality may allow the CEO to control information available to other directors
impeding effective monitoring (Jensen, 1993). However, empirical work has not supported the
concept that duality can lead to agency costs. For example, Brickley, Coles, and Jarrell (1997)
show that CEOs are awarded the chair position as a normal part of the succession process;
successful CEOs later become CEO/Chair. While duality may have an impact on deviations from
expected stakeholder management, we do not have strong priors with respect to empirical
outcomes. It is possible that duality proxies for the CEOs power over the board of directors, thus
impeding effective monitoring by the board. It is also possible that duality is endogenously
determined, such that CEOs that are less likely to engage in unnecessary investment in
stakeholder management are awarded the chair position. We therefore treat the expected impact of
duality as an open empirical issue.
3.2.4. Institutional and blockholder ownership and agency costs
Given their large financial stake, institutional investors and blockholders have incentives,
resources, and the ability to mitigate agency problems (Edwards and Hubbard, 2000) and monitor
a firms stakeholder policy. Empirical research supports this idea (Brickley, Lease, and Smith,
1988; Hartzell and Starks, 2003; Holderness, 2003; Jiambalvo, Rajgopal, and Venkatachalam,
2002; McConnell and Servaes, 1990). Empirical studies have also explored the role of
blockholders in the conflict between stakeholder management and financial performance. For
example, institutional ownership plays an important role in monitoring and influencing CEOs
attention to corporate social performance (Graves and Waddock, 1994; Wright et al., 2002;
18

Wright, Ferris, Sarin, and Awasthi, 1996; Johnson and Greening, 1999; Neubaum and Zahra,
2006).
Our focus, however, is on the role of institutional investors in preventing needless
investment in stakeholder management. A higher concentration of institutional investor ownership
should also lead to lower levels of unnecessary investment in stakeholder management.
Consequently, the institutional ownership hypothesis is as follows:
H
4
: Deviation from expected stakeholder management is negatively related to the
percentage of shares held by institutional investors.

3.2.5. Operationalizing expected stakeholder management
Our theoretical arguments allude to a value maximizing, or optimal, level of stakeholder
management. Our measurement of optimal stakeholder management is limited, however, by a
lack of empirical literature on this concept. Consequently, as a starting point we assume that
management has an incentive to invest in stakeholder management to increase firm value either
voluntarily or due to pressure from regulators, investors, the board of directors, the media, or the
product market. And like other financial decisions, management has a target level of investment
in stakeholder management which is believed to maximize shareholder wealth. Given the
difficulty in accurately assessing the optimal level of stakeholder management, empirical
evidence and corporate practice suggest that industry average provides a useful starting point for
arriving at this value. Masulis (1983), for example, finds that when firms issuing debt move
toward the industry average from below, the market reacts more positively than when the firm
moves away from the industry average, indicating that industry average provides a proper
benchmark for optimal capital structure. Similarly, Frank and Goyal (2009) find that the median
industry leverage is one of most reliable factors in the capital structure decisions of publicly
19

traded American firms. Extending this logic from the capital structure literature, an optimal level
of stakeholder management exists, but not every firm achieves this value-maximizing level.
Industry average provides an approximate estimate for the optimal level of stakeholder
management
14

However, given the difficulty in measuring actual investment in stakeholder management
and the unique characteristics of each firm, a simple industry average is likely to be a noisy proxy.
Specifically, in addition to industry, the optimal level of stakeholder management may be
explained by firm value and other observable firm specific characteristics, such as firm size,
financial constraints, and growth opportunities. Such an interpretation is also more consistent with
Jensens (2001, 2002) argument that value is the criterion to stakeholder management. A more
detailed discussion of the operationalization of our measures used for deviation from expected
stakeholder management is provided in section 4.2.1.
Finally, given our operationalization of optimal stakeholder management, our conclusions
depend on a key assumption. Namely, that the sample mean, conditional upon industry and other
observable firm specific factors, is a reasonable proxy for optimal stakeholder management. To
the extent that this assumption is violated, our measures of deviation from expected stakeholder
management may be biased.
4. Sample selection and data
4.1. Sample selection
We obtain our initial sample from the KLD Statistical Tool for the Analysis of Trends in
Social and Environmental Performance (KLD) database, an annual statistical database of over 90

14
The use of industry average as a proxy for the optimal level is also validated by common practice in financial
statement analysis, in which greater deviations from the industry average are a potential indicator of financial
problems within the firm.
20

social and environmental indicators provided by KLD Research and Analytics, Inc. The database
reports social, environmental, and governance performance indicators for S&P 500, Domini 400
Social, and Russell 1000 (starting in 2001) and Russell 3000 (starting in 2003) companies
between the years of 1991 to 2008. We use the KLD database because it is a measure of social
performance (relations with stakeholders) that has been developed independently of this studys
researchers, and therefore, does not suffer from potential researcher bias that might occur if we
use our own definition of stakeholder management and corporate social performance.
Furthermore, the KLD database is a well-established measure of both stakeholder
management and corporate social performance. Researchers have certified its quality and it has
been widely used in empirical studies.
15

16

Our initial KLD sample has 26,575 firm years from 1991 to 2008. We match the KLD
data with COMPUSTAT accounting data, resulting in a sample composed of an unbalanced panel
of 24,813 firm year observations for 4,552 firms.
17
We match this set of firms to Standard and
Poors EXECUCOMP database, which includes annual compensation data from proxy statements
for the five highest paid executives for firms in the S&P 500, the S&P MidCap 400, and the S&P
SmallCap 600. We lose one year of data because the EXECUCOMP data begins in 1992. This
results in an unbalanced panel of 15,761 firm year observations for 2,297 firms from 1992 to
2008. To account for the influence of internal governance mechanisms, we obtain board of

15
For example, Chatterji, Levine, and Toffel (2009) conclude that while the KLD database does not reflect all
available information on stakeholder management and corporate social performance, it is a good predictor of more
narrow measures such as compliance with environmental regulations.
16
See Agle, Mitchell, and Sonnenfeld (1999), Berman, Wicks, Kotha, and Jones (1999), Graves and Waddock
(1994), Hillman and Keim (2001), Johnson and Greening (1999), Turban and Greening (1996), Waddock and Graves
(1997), and Coombs and Gilley (2005).
17
A primary problem with the KLD data is that the data lacks a numeric unique identifier (i.e. CUSIP) for years prior
to 1995 in which to merge with COMPUSTAT. We try several methods to obtain a COMPUSTAT unique identifier
(GVKEY) for the initial sample. First, we backfill early years of data (prior to 1995) with CUSIPs from the same
firm if it appears in the database in a later year. Next, we attempt to merge companies with COMPUSTAT using
CUSIP, ticker, and the first 15 letters in the company name. Lastly, we hand verify our merge results from the steps
above and hand match firms with COMPUSTAT for any remaining firms not captured in previous steps.
21

director characteristics from the RiskMetrics IRRC (IRRC) database and institutional ownership
data from the Thomson 13F (13F) database. Due to data limitations in IRRC, board characteristics
are only available from fiscal year 1995 onward.
18
We obtain stock return volatility, volume, and
shares outstanding data from the Center for Research in Securities Prices (CRSP) database. We
omit regulated utilities (SIC codes 4910-4949), depository institutions (SIC codes 6000-6099) and
holding or investment companies (SIC codes 6700-6799) from our sample. The final matched
sample, after combining KLD, COMPUSTAT, EXECUCOMP, IRRC, 13F, and CRSP data,
contains an unbalanced panel of 9,051 firm year observations for 1,631 firms from 1995 to 2008.
4.2. Variable descriptions
We rely on the KLD database for stakeholder management and social issue performance
data, COMPUSTAT for accounting data, EXECUCOMP for managerial ownership,
compensation data, and CEO data, IRRC for board and CEO data, 13F for institutional ownership
data, and CRSP for volume traded and shares outstanding data. We winsorize all variables at the
upper and lower 1% to reduce any possible impact from outliers.
4.2.1. Deviation from expected stakeholder management
Enlightened value maximization posits that firms will invest in stakeholder management
(SM) until the marginal cost of doing so exceeds the marginal benefit to shareholders, and we call
this the expected investment in stakeholder management (ESM). While we can observe the effect
of a firms actual investment in stakeholder management, SM, ESM is unobservable. Therefore,
we use two proxies for ESM to measure the deviation of each firm from expected SM.

18
The Risk Metrics IRRC proxy database reports the year using the meeting date. We assume that the proxy date is
approximately 3 months after fiscal year end and the meeting date follows by 1 month. So, firms with fiscal years
ending in December 1995 will be matched with IRRC observations with meeting dates in April 1996 or earlier.
22

To begin with, we create a measure of the deviation of each firm from expected
stakeholder management (DESM), which is the residual from a cross-sectional regression of SM
on size, financing constraints, performance, and growth opportunities estimated at time t for all
firms in the same 4-digit Global Industry Classification System (GICS) industry
19
:
SH
t
= o +
1
(sizc)
t
+
2
(inoncing constroints)
t
+
3
(pcrormoncc)
t

+
4
(growtb opportunitics)
t
+e
t
(1)
where SM is the firms stakeholder management score, size is LNSALES, financing constraints
are measured by CASH/TA and TD/TA, performance is measured by ROA and TRS, growth
opportunities is measured by Q and DESM = e
t
. The definitions of variables in regression
equation (1) are described below. The residuals are a proxy of deviations from expected
investment in SM given industry, year, and observable firm specific factors. To construct
measures for SM, we follow the procedure used in prior research in the area (Waddock and
Graves, 1997; Hillman and Keim, 2001; Coombs and Gilley, 2005), and construct a measure of
the firms stakeholder management (SM) performance using the KLD categories of employee
relations (EMP), diversity issues (DIV), product issues (PRO), community relations (COM), and
environmental issues (ENV). These five categories parallel the primary stakeholder groups with
regard to employees (including diversity initiatives), customers (product safety/quality), the

19
Following Benson and Davidson (2009), we use 4-digit GICS industry group rather than the more common 2-digit
SIC code or Fama-French (FF) industry classification for several reasons. First, GICS categorization is based on both
a firms operational characteristics and information on investors perceptions as to what constitutes the firms main
line of business. Second, Chan, Lakonishok, and Swaminathan (2007) demonstrate that a 4-digit GICS code
classification using 24 categories achieves the same level of discrimination as the FF procedure using 48 categories.
Finally, the ability to effectively discriminate between industries using a smaller number of categories results in more
industry peers each year than the FF procedure, resulting in potentially more accurate industry averages. A
breakdown of 4-digit GICS and FF industry classifications by year for the KLD database is provided in Appendix C
(available from the authors upon request).
23

natural environment, the community, and suppliers (to the extent that certain diversity initiatives
are directed toward suppliers). We detail the calculation of the SM variables in Appendix A.
As a second proxy for the deviation from expected SM, we assume that ESM is close to
the industry average for each firm in each year. Therefore, we use industry adjusted SM (IASM)
as an additional measure of the deviation of each firm from expected stakeholder management:
IASH
t
= SN
t
- IAI0SH
t
(2)
where industry adjusted SM (IASM) is equal to the difference between each firms actual
investment in SM at time t and the industry average SM (IAVGSM), the average value of peer
firms within each companys 4-digit GICS industry group at time t.
4.2.2. Managerial ownership incentives
We use delta as a proxy for CEO equity-based compensation and ownership incentives.
As in Edmans et al. (2009), delta is the dollar change in CEO portfolio wealth for a 1% change in
firm value scaled by total annual compensation. We estimate each CEOs portfolio of stock and
options using data from the EXECUCOMP database as in Core and Guay (2002). We detail the
calculation of the CEO portfolio delta in Appendix B. We use log of delta (LNDELTA) in our
analysis to account for the high skewness and kurtosis in the variable.
4.2.3. Board monitoring
We proxy for effective monitoring by the board of directors using four measures: board
size, the percentage of independent directors on the board, board independence, and CEO duality.
First, we measure board size (BRDSIZE) as the number of directors serving on the board during
the year. Next, we classify directors as insiders (employed by the firm), affiliated (e.g. former
employees, family members of employees, or those with business relations with the firm) and
24

independent (Baysinger and Butler, 1985). We then create a variable for the percentage of outside
directors serving on the board (PIND) calculated as the percentage of outside directors relative to
total directors on the board. We also create a composite measure of board independence
(BRDIND), calculated as the ratio of independent directors to insider and affiliated directors
multiplied by the inverse of board size. Higher values of this measure indicate more effective
board structure (i.e. higher ratio of independent directors and/or smaller board size). Finally, we
determine whether the CEO holds the title of CEO and COB (DUALITY) and create a dummy
variable equal to one in such cases, and zero otherwise.
4.2.4. Institutional and blockholder ownership
We develop a measure of institutional investor ownership percentage (INSTOWN), which
is the percentage of shares held by institutions during the year.
4.2.5. Controls
We include several additional control variables to proxy for various firm specific factors
which have been shown to be determinants of stakeholder management, internal governance
structure, and firm value.
Cash. High cash balances may exacerbate agency problems (Jensen, 1986) leading to
inefficient investment. We control for cash using the ratio of cash to total assets (CASH).
Capital structure (Leverage). Theoretical and empirical research has found that leverage is
positively related to firm value. However, Titman (1984) and Maksimovic and Titman (1991)
argue that some firms also consider the costs imposed on non-financial stakeholders in
undertaking relationship-specific investments when making capital structure decisions. For
example, research has shown that firms selling unique products (Titman and Wessels, 1988), or
25

firms entering into bilateral relationships (Banerjee et al., 2008) or strategic alliances (Kale and
Shahrur, 2007) maintain or lower leverage. We proxy for the firms capital structure using the
debt ratio (TD/TA); we calculate it as book value of total debt to book value of total assets.
Firm size. Research has shown that firm size is positively related to board size and CEO
compensation, but negatively related to firm value. We control for firm size using the log of sales
(LNSALES).
Firm Performance. We measure accounting performance using return on assets (ROA),
calculated as earnings before interest and taxes divided by total assets. As an additional measure
of financial performance, we use the 1-year total return to shareholders (TRS), including the
monthly reinvestment of dividends. We obtain this measure from the COMPUSTAT database.
4.2.6. Instruments
We use several instruments to proxy for various firm specific factors which have been
shown to be determinants of internal governance structure.
Growth opportunities and information asymmetry. Jensen (1993) argues that monitoring
high growth firms is costly, and Fama and Jensen (1983) suggest that firms with higher stock
return volatility have higher levels of information asymmetry. Prior research has suggested that
the monitoring costs associated with higher levels of information asymmetry are inversely related
to board size and independence (Linck et al., 2008; Adams and Ferreira, 2007; Raheja, 2005;
Maug, 1997). Similarly, Demsetz and Lehn (1985) note that higher levels of information
asymmetry (as proxied by higher volatility) are associated with more managerial discretion,
necessitating higher levels of variable compensation. Following Linck, et al. (2008) we proxy for
growth opportunities and information asymmetry using Tobins Q, the level of R&D
expenditures, and total firm risk. As in Smith and Watts (1992), we calculate Tobins Q (Q) as the
26

market value of equity minus the book value of equity plus the book value of assets all divided by
the book value of assets.
20
R&D expenditures (R&D/TA) are the dollar value of R&D
expenditures scaled by total assets. Many firms in the COMPUSTAT database have missing
values for R&D. We set R&D expenditures equal to zero when the value is missing. Total equity
risk is measured using volatility (VOLATILITY), which is the annualized standard deviation of
monthly stock returns calculated over 36 months.
Capital intensity. Previous research has found that managerial ownership and
compensation are related to the capital intensity of the firm. We include a variable for capital
intensity (CAP), measured as the net property, plant, and equipment to total assets.
CEO characteristics. Hermalin and Weisbach (1998) suggest that firms will add insiders
to the board of directors as the CEO approaches retirement as part of the succession process. We
use CEO age (AGE) as a proxy for the length of time to retirement of the CEO.
Risk aversion. Following Coles, Naveen, and Naveen (2006), we include CEO tenure
(TENURE), measured as the time the CEO has been in the current position, and CEO cash
compensation (TCC), measured as salary plus bonus, as proxies for the CEOs level of risk
aversion. CEOs with longer tenures may be entrenched, making them more likely to avoid risk
(Berger et al., 1997). However, Guay (1999) argues that higher levels of cash compensation
reduces manager risk aversion because cash compensation is unrelated to performance and
permits their personal diversification.
Turnover. As in Hartzell and Starks (2003), we use turnover (TURNOVER), the log of
one plus the ratio of monthly volume to number of shares outstanding in the month prior to the
institutional ownership observation, as an instrument for institutional investor ownership.

20
We use Q rather than MKBK as in Linck, et al. (2008) to account for the problem of many firms having negative or
small values of book value of equity in the denominator. This leads to negative values of MKBK or abnormally large
values of MKBK that are driven by small book values of equity rather than large market values of equity.
27

4.3. Descriptive statistics
Table 1, Panel A, presents descriptive statistics. The mean (median) deviation from
expected stakeholder management score (DESM) is 0.02 (0.02). The measure ranges from a
minimum score of 0.88 to a maximum score of 0.93. The mean (median) industry adjusted
stakeholder management score (IASM) is 0.04 (0.05). The measure ranges from a minimum score
of 1.01 to a maximum score of 1.14. The mean (median) stakeholder management score (SM) is
0.05 (0.00). The measure ranges from a minimum score of 1.12 to a maximum score of 1.16.
-----Insert Table 1 About Here-----
The mean (median) value for DELTA is 446.49 (85.41). Our sample boards average 9.66
directors (BRDSIZE). Approximately 29.82% of directors are independent (PIND), and the CEO
is also Chair (DUALITY) approximately 66% of the time. Institutional ownership (INSTOWN)
averages 75.41% of shares outstanding. Finally, our sample consists of large industrial firms. The
mean (median) value of total sales is $6.669 ($2.142) billion. However, the sample includes a
broad range of firms with total sales ranging from $94.06 million to $81.303 billion.
Table 1, Panel B, presents correlations of key variables for our sample. The correlations
between the two measures of deviation from expected stakeholder management (DESM and
IASM) and SM are significant. However, the correlations between DESM and IASM or SM are
less than 90%, suggesting that they may capture different dimensions of overall deviation from
expected stakeholder management. Lastly, we find that less financially constrained firms (higher
CASH/TA and lower TD/TA) and firms with higher ROA are associated with higher deviations
from expected SM (IASM and SM).
21
This finding may also reflect that firms with slack financial
resources are associated with higher levels of SM (Waddock and Graves, 1997).

21
Note that we control for CASH/TA and TD/TA in equation 1 when calculating DESM. So, it is not surprising that
DESM does not exhibit a similar significant correlation with these measures.
28

The correlations between LNDELTA and two of three measures of deviation from
expected and actual SM are positive and significant. Interestingly, we also find that dollar/dollar,
or effective percentage owned, incentives, (Jensen and Murphy, 1990) have a negative and
significant correlation with both measures of deviation from expected SM and actual SM (un-
tabulated). This finding is consistent with the predictions of Hall and Baker (2004) who note that
dollar/dollar incentives are more effective at deterring tasks that do not scale with firm size.
Deviations from expected SM, therefore, may be additive value destructive and reflect a form of
perquisite consumption by the CEO. Combined, these findings provide initial support for the CEO
perquisite consumption hypothesis (H
2
). Conversely, more effective corporate governance is
associated with lower deviations from expected SM; it has positive and significant correlations
with BRDSIZE, negative and significant correlations (DESM and IASM) with PIND, and
negative and significant correlation (SM) with INSTOWN. Furthermore, higher LNDELTA is
associated with more effective board governance (smaller BRDSIZE, higher PIND, separation of
CEO and COB title) and lower levels of institutional ownership (INSTOWN). Overall, this initial
evidence supports the board monitoring hypothesis (H
3
). As noted by Edmans et al. (2009), more
effective corporate governance is associated with higher managerial incentives, the latter being an
effective means of addressing agency costs that are proportionate with firm value. However,
incentive pay is ineffective at deterring perquisite consumption, which is largely independent of
firm value; additive value destructive actions are best controlled by direct monitoring, rather than
incentive pay.
22
Our initial observations suggest that firms recognize this tradeoff and adjust
accordingly.

22
As a final note, we find that percent/percent incentives (LNDELTA) are independent of firm size (LNSALES),
whereas the dollar/dollar (Jensen and Murphy, 1990) and dollar/percent (Baker and Hall, 2004) measures have
negative and significant correlations with firm size (un-tabulated). This makes the percent/percent measure an
empirically desirable measure of managerial incentives (Edmans et al., 2009).
29

5. Analysis of the effect of internal governance mechanisms on the deviation from expected
stakeholder management
We begin by examining whether internal governance mechanisms are related to deviations
from expected stakeholder management. Previous research has explored the relation between SM
and firm performance using pooled ordinary least squares regressions, which assumes that a
firms level of SM is exogenous to the firm (Hillman and Keim, 2001). Modeling the relation in
this manner fails to control for unobserved firm heterogeneity, or omitted variable bias. As a
result, it is possible that these results are biased due to multiple occurrences of the omitted
variable across time periods. We, therefore, use firm fixed effects (FE) regression as the
estimation method to control for the presence of unobserved firm effects.
Specifically, we test whether changes in internal governance mechanisms affect deviations
from expected stakeholder management or are endogenously determined by the firm specific
factors. The model we test is as follows:
BESN
t
= +
1
[
LNLL1A
+

t
+ [
2
[
BRSIzL
+

t
+[
3
[
PIN
-

t
+ [
4
[
0ALI1
+-

t
+ [
5
[
INS10wN
-

t
+[
k
(Controls)
t
+ [

(ycor Jummy :orioblcs)

t
+ e
t
(S)
where deviation from expected stakeholder management (DESM) is a function of managerial
ownership incentives (LNDELTA), board monitoring (BRDSIZE, PIND, DUALITY),
institutional ownership (INSTOWN), controls (CASH/TA, TD/TA, LNSALES, ROA, TRS), and
dummy variables for year. The definition of variables in regression equation (3) is as described in
section 4.2. The sign beneath each variable indicates the expected relation between the dependent
variable and relevant independent variables.
30

Table 2, column 1 reports results of fixed-effects (FE) estimates of regression equation
(3).
23
Here and throughout the table we compute the t (z)-statistics using robust standard errors
clustered at the firm level (see Rogers, 1993; Wooldridge, 2002). The estimated coefficient for
LNDELTA is positive and significant at the 5% level (t = 2.18). Higher managerial incentives are
associated with larger deviations from expected stakeholder management. This result supports the
CEO perquisite hypothesis (H
2
). While the signs on BRDSIZE and PIND are as predicted, only
the estimated coefficient for BRDSIZE is significant at the 1% level (t = 3.01). Larger boards are
associated with larger deviations from expected stakeholder management. This provides partial
support for the board monitoring hypothesis (H
3
). The estimated coefficients for DUALITY and
INSTOWN are insignificant. Column 2 includes our composite measure for BRDIND rather than
BRDSIZE and PIND. While the sign on BRDIND is as predicted, the estimated coefficient for
this variable is not significant. The significance and direction of the estimated coefficients for
LNDELTA, DUALITY, and INSTOWN are similar to those in column 1.
-----Insert Table 2 About Here-----
While our initial finding provides some evidence that higher levels of managerial
incentives are associated with greater deviations from expected SM and that more effective
internal governance mechanisms work to lessen the deviations from expected SM, these
deviations may also reflect either an under- or over-investment from expected SM. While a
negative value of DESM (a negative residual) suggests an under-investment in SM, a positive
value of DESM (a positive residual) suggests an over-investment. Internal governance
mechanisms may influence under-investment and positive deviations from expected SM

23
Pooling tests for firm fixed effects and year fixed effects are significant at the 1% level, supporting the inclusion of
firm and time effects in the model. The conclusion of subject specific parameters in the model is also supported by a
modified Wald statistic for group-wise heteroskedasticity in the residuals of a fixed effect regression model,
following Greene (2003, p. 598). This statistic is consistent at the 1% level. Finally, a Hausman test for random
effects is significant at the 1% level, supporting the use of the fixed effects model over the random effects model.
31

differently. To test this directly, we first sort DESM into quintiles. Firm observations in the
bottom two quintiles, those with the most negative values, are classified as under-investing in SM.
Firm observations in the top two quintiles, those with the most positive values, are classified as
over-investing in SM. We classify firms in the middle quintile as near expected SM and treat
them as a benchmark. We then estimate a multinomial logit model that predicts the likelihood that
a firm will be in one of the two extreme quintile groups as opposed to the middle quintile.
24

Columns 3 and 4 contain results for a multinomial logit regression of equation (3).
Column 3 contains estimated results for the under-investment in SM firms. The estimated
coefficient for BRDSIZE is negative and significant at the 1% level (z = 3.77). Larger boards are
associated with a lower relative risk of under-investment in SM. Column 4 contains estimated
results for the over-investment in SM firms relative to the normal investment in SM firms. The
estimated coefficient for BRDSIZE is now positive and significant at the 5% level (z = 2.25),
while the estimated coefficient for PIND is negative and significant at the 5% level (z = 2.40).
More effective board monitoring (i.e. smaller boards and more independent directors) are
associated with a lower relative risk of over-investment in SM. Columns 5 and 6 report results
using BRDIND rather than BRDSIZE and PIND. Column 5 contains estimated results for the
under-investment in SM firms relative to the normal investment in SM firms. The estimated
coefficient for BRDIND is positive and significant at the 10% level (z = 1.90). Column 6 contains
estimated results for the over-investment in SM firms relative to the normal investment in SM
firms. The estimated coefficient on BRDIND is now negative and significant at the 1% level (z =
2.68). More independent boards are associated with a higher relative risk of under-investment in

24
We use a multinomial logit model rather than an ordered logit model because we are uncertain of the ordinality of
the ESM measure. As noted by Long (1997), using a nominal model when the dependent variable is ordinal results in
a loss of efficiency since information is ignored. However, applying an ordinal model to a nominal dependent
variable may result in biased estimates. In cases where there is a question about the ordinality of the dependent
variable, the loss of efficiency of using a nominal model is outweighed by the potential bias of ordinal model.
32

SM and lower relative risk of over-investment in SM. Overall, these results suggest that more
effective board monitoring (i.e. smaller boards and more independent directors) may be effective
at constraining over-investment, or positive deviations from expected SM. Our results continue to
provide support for the board monitoring hypothesis (H
3
).
Columns 7 and 8 contain estimated results for regression equation (3) replacing DESM
with IASM as a proxy for the deviation from expected investment in SM. The significance and
direction of the estimated coefficients for LNDELTA, BRDSIZE, PIND, BRDIND, and
INSTOWN are similar to those in columns 1 and 2. The estimated coefficients for DUALITY,
however, are now significant. This finding is consistent with prior research; successful CEOs,
such as those better at maximizing shareholder wealth, become COB (Brickley et al., 1997).
Lastly, the deviation of observed SM from the expected level of SM may be conditional
upon the financing constraints faced by each firm. Under this assumption, SM is more likely to
deviate from expected SM when managers are less financially constrained and agency problems
are more likely to exist. We use cash and leverage as proxies for financial constraints. Large cash
balances may exacerbate agency problems (Jensen, 1986). As such, managers of firms with large
cash balances may have a tendency to overinvest in SM. Conversely, firms with low leverage may
be less financially constrained, which may cause managers to over-invest in SM. We follow the
procedure used in Biddle et al. (2009) to proxy for the financial constraints faced by each firm.
We rank firms into deciles based on cash and leverage (we first multiply leverage by minus one
so that it is decreasing in the level of financial constraint). We then create a composite measure
for the propensity to over-invest in stakeholder management (OVER), which is the average of the
ranked values of cash and negative leverage re-scaled so that the values range between zero and
one. Higher values for OVER reflect less financially constrained firms, which may increase the
33

likelihood that observed SM deviates from expected SM; managers of these firms may have a
greater propensity to over-invest in SM. We test this by creating interaction terms between each
of our internal governance mechanism variables (LNDELTA, BRDSIZE, PIND, DUALITY, and
INSTOWN) and OVER.
Specifically, we test whether changes in internal governance mechanisms, conditional on
whether the firm is financially constrained and more likely to over-invest in SM, affect observed
SM or are endogenously determined by the firm specific factors. The model we test is as follows:
SN
t
= +
1
(INEIIA)
t
+ [
2
(BRSIZE)
t
+[
3
(PIN)
t
+ [
4
(uAIII)
t
+ [
5
(INSI0wN)
t
+
6
[
LNLL1A X 0vLR
+

t
+ [
7
[
BRSIzL X 0vLR
+

t
+[
8
[
PIN X 0vLR
-

t
+ [
9
[
0ALI1 X 0vLR
+
-

t
+ [
10
[
INS10wN X 0vLR
-

t
+[
k
(Controls)
t
+ [

(ycor Jummy :orioblcs)

t
+ e
t
(4)
where stakeholder management (SM) is a function of managerial ownership incentives
(LNDELTA), board monitoring (BRDSIZE, PIND, DUALITY), institutional ownership
(INSTOWN), interactions between internal governance mechanisms and OVER, controls
(LNSALES, ROA, TRS), and dummy variables for year. The definition of variables in regression
equation (4) is as mentioned in section 4.2. We exclude CASH and TD/TA from our list of
controls because they are used to calculate OVER. The estimated coefficients (
1
,
2,

3,

4,
and

5
)
measure the relation between internal governance mechanisms and SM for the most financially
constrained firms. The sum of the coefficients (
1
+
6
,
2
+
7
,
3
+
8
,
4
+
9
,
5
+
10
)
measure the relation between internal governance mechanisms and SM for the least financially
constrained firms. The sign beneath each variable indicates the expected relation between the
dependent variable and relevant independent variables.
34

Column 9 contains results of fixed-effects (FE) estimates for regression equation (4). With
the exception of the estimated coefficient for DUALITY, which is positive and significant at the
10% level (t = 1.66), none of the estimated coefficients on the main effects (i.e. the most
financially constrained firms) for LNDELTA, BRDSIZE, PIND, and INSTOWN are significant.
However, as predicted, the estimated coefficients for the interaction between OVER and
BRDSIZE is positive and significant at the 1% level, while the estimated coefficients for the
interaction between OVER and DUALITY is negative and significant at the 1% level. This
suggests that the relation between internal governance and investment in SM is conditional upon
whether a firm is more likely to over-invest, and suggests that these mechanisms may be more
effective at constraining excess SM. Furthermore, the overall relation between internal
governance and SM for the least financially constrained firms (measured by the sum of the
coefficients for the internal governance mechanism variables and the interaction terms between
internal governance mechanisms and OVER) for LNDELTA and BRDSIZE are positive and
significant at the 5% and 1% levels, respectively, while the sum of coefficients for DUALITY is
negative and significant at the 5% level. While we find that the relation between INSTOWN and
SM is significantly higher for less financially constrained firms, the overall relation is not
significant. Our conclusion using BRDIND in column 10 is similar to that in column 9.
5.1.1. Robustness tests
In the presence of unobserved firm effects, firm fixed effects (FE) regression is commonly
suggested. However, FE estimation may be unsuitable in our unbalanced panel for several
reasons. First, several of our primary variables of interest, such as BRDSIZE and PIND, are
relatively time invariant and cannot be estimated with FE regression as they would be absorbed in
the within transformation of the variable. Second, FE estimation requires significant within panel
35

variation for important right-hand side variables to produce consistent and efficient estimates
(Wooldridge, 2002, p. 286). For example, the within standard deviations for BRDSIZE (1.04) and
PIND (9.13) are substantially smaller than the between standard deviations (2.21) and (12.55),
respectively. Furthermore, while our sample spans 13 years, on average firms remain in our
sample for approximately half of this period. Consequently, for robustness we also run random-
effects (RE) GLS regression on the unbalanced panel, and we use FE estimation on a balanced
panel of firms (153 firms and 1,989 firm-year observations) that remain in our sample for all 13
years of our sample (reported in Appendix C and available from the authors upon request).
25
Our
results are similar to those using the FE model on the unbalanced panel, except that the estimated
coefficients for PIND and BRDIND remain negative, but are generally significant.
5.2. Exploring causality
The estimated coefficients in Table 2 may be biased as the various internal governance
mechanisms are endogenously formed (e.g. Hermalin and Weisbach, 2003). We address
endogeneity concerns in two ways. First, to ensure that causality runs from internal governance to
deviation from expected stakeholder management, we re-estimate regression equations (3) and (4)
replacing contemporaneous internal governance variables with their lagged values and also using
contemporaneous internal governance variables while including a lagged dependent variable in

25
In cases where researchers are interested in the inferring significance of key variables in the model, Wooldridge
(2002, pg. 290) suggests a t-statistic Hausman test computed as (

FE
/

RE
)/{[se(

FE
)]
2
-[se(

RE
)]
2
}
1/2
to determine
whether a random effects estimator is appropriate. While the overall Hausman (1978) test is significant at the 1%
level, the t-statistic Hausman tests are insignificant on all of the variables of interest, supporting the use of the
random-effects model in our analysis. Consequently, we use random-effects (RE) GLS regression as the estimation
method for our unbalanced panel. However, as noted by Mundlak (1978), one assumes that the researcher is making
the assumption that the omitted variable, c
i
, is uncorrelated with x
it
when using a random-effects model. If this
assumption is violated, the random-effects estimators are inconsistent. To minimize this issue, we include dummy
variables for industry and year in our models to control for part of the c
i
correlated with x
it
. Still, a limitation of using
the random effects model in our context is that we can only infer the significance of the relations for key variables
(i.e. governance), rather than all variables in our model.
36

our regression equation. Our results are similar to those reported in Table 2 (un-tabulated and
available from the authors upon request).
While lagging the measurement of the dependent and/or independent variables partially
controls for simultaneity, it does not provide a complete remedy. Our analysis to this point has
treated the internal governance variables as exogenous, but if any one of these variables is
determined simultaneously with deviation from expected stakeholder management, it violates the
least squares regression assumption that the regressors are uncorrelated with the error term. To
eliminate the endogeneity problem from simultaneity bias, we endogenize LNDELTA, BRDIND,
and INSTOWN given the existing literature on managerial ownership incentives (Coles et al,
2006), board structure determinants (Linck et al., 2008), and institutional ownership (Hartzell and
Starks, 2003) by developing the following four regression equations:
26

(ESH)
t
= +[
1
_
INEIIA
+
]
t
+ [
2
_
BRIN
+
]
t
+ [
3
_
uAIII
_
]
t
+ [
4
_
INSI0wN
-
]
t

+ [
5
(INICC)
t
+ [
6
(IuRN0IER)
t
+[
k
(C0NIR0IS)
t
+ e
t
(S)
(INEIIA)
t
= +[
1
(ESH)
t
+ [
2
()
t
+ [
3
(CAP)
t
+ [
4
(RIA)
t
+ [
5
(I0IAIIIII)
t
+ [
6
(INICC)
t
+ [
7
(IENIuRE)
t
+[
k
(C0NIR0IS)
t
+ u
t
(6)
(BRIN)
t
= + [
1
(ESH)
t
+ [
2
()
t
+ [
3
(RIA)
t
+ [
4
(I0IAIIIII)
t
+ [
5
(INEIIA)
t
+ [
6
(A0E)
t
+ [
7
(uAIII)
t
+[
k
(C0NIR0IS)
t
+ :
t
(7)

26
While DUALITY may also be endogenously formed, we treat it as exogenous because it is a binary variable. We
also use the composite measure of effective board monitoring (BRDIND), rather than BRDSIZE and PIND separately
to simplify interpretation of our results and due to difficulties identifying a set of instruments that is sufficiently
correlated with PIND. The set of instruments is strongly correlated with BRDIND.
37

(INSI0wN)
t
= + [
1
(ESH)
t
+ [
2
(IuRN0IER)
t
+[
k
(C0NIR0IS)
t
+ p
t
(8)
where the definition of variables is as mentioned in section 4.2 and the list of controls remains the
same as in regression equation (3). The four equations, regression equations (5) (8) are solved
using system of simultaneous equations using first-difference three-stage least squares (FD3SLS)
to control for unobserved firm heterogeneity, or omitted variable bias.
27

Table 3 reports results of FD3SLS estimates of the system of four equations.
28
Column 1
contains FD3SLS results for regression equation (5). The estimated coefficient for BRDIND
remains negative and significant at the 1% level (z = 6.26). More effective monitoring by the
board of directors (smaller boards with more independent directors) leads to lower deviations
from expected SM. This continues to support the board monitoring hypothesis (H
3
). The
estimated coefficient for DUALITY also remains negative and significant at the 5% level (z =
2.73). However, the estimated coefficient for LNDELTA, while positive, is no longer
significant. Furthermore, the estimated coefficient for INSTOWN is now negative but not
significant.
-----Insert Table 3 About Here-----
The results in columns 2 and 3 for regression equations (6) and (7), respectively, also
provide some additional useful information. For example, the negative and significant estimated

27
In the absence of specification problems, the use of a systems procedure, such as 3SLS, is asymptotically more
efficient than a single-equation procedure such as 2SLS. But single-equation methods are more robust to
misspecification (Wooldridge, 2002 p. 222). Our conclusions are unchanged using 2SLS (un-tabulated and available
from the authors upon request).
28
A DavidsonMacKinnon (1993) test for endogeneity is significant at the 1% level supporting the use of the 3SLS
model in Table 4. However, when conducting 2SLS/3SLS the set of instruments, z, must also be highly correlated
with the endogenous regressors, x
k
, but uncorrelated with the disturbance process, u. The first condition, that the set
of instruments is sufficiently correlated with the endogenous regressors, is supported by estimated coefficients and F
statistics for the set of instruments that are significant at the 1% level, high Shea (1997) partial R
2
values, and a
Kleibergen-Papp (2006) LM statistic that is significant at the 1% level. The latter condition, that the set of
instruments is uncorrelated with (orthogonal to) the disturbance process is supported by a non-significant Hansen-
Sargan J-statistic.

38

coefficients for DESM in columns 2 and 3 suggest that there is an inverse relation between
deviations from expected SM and the level of CEO incentives and board independence. Firms
with lower deviations from expected SM have CEOs with higher incentives. Higher levels of
managerial incentives potentially expose the firm to perquisite consumption in the form of
unnecessary investment in SM. However, board independence also increases in tandem with the
level of CEO incentives. The negative and significant estimated coefficient for BRDIND and
positive but not significant estimated coefficient for LNDELTA in the main equation suggest that
greater monitoring by the board effectively constrains this perquisite consumption by the CEO in
the form of unwarranted investment in SM.
5.2.1. Industry level analysis of the effect of internal governance mechanisms on deviations from
expected investment in stakeholder management and components
Although we find some support for a relation between various internal governance
mechanisms and deviations from expected SM, industry likely plays an important role in
determining the nature of the firm-level stakeholder environment, and as a result, influences the
relative importance and allocation of resources to various competing stakeholder groups. For
example, investment in diversity, community, and environment related stakeholder management
may play a role in differentiating ones product for consumer oriented firms. As a result, some
consumers may be willing to pay a premium for such products, leading to better financial
performance. Higher investment in these dimensions of SM for industrial oriented firms,
however, may provide limited financial benefits. Conversely, strong relations with employee,
supplier, and product related stakeholder groups are likely to benefit both consumer and industrial
oriented industries more equally via either the attraction of more talented or motivated employees
or stronger relations with either upstream or downstream customers/suppliers. It may also be the
39

case that internal governance mechanisms are only related to deviations from the expected
stakeholder strengths (abnormally good relations) or concerns (abnormally bad relations).
Therefore, we test for differences between consumer and industrial firms by running regressions
on a consumer subsample which includes firms classified as operating in 2-digit GICS codes
25(Consumer Discretionary), 30 (Consumer Staples), 35 (Health Care), 40 (Financials), and 45
(Information Technology) and an industrial subsample which includes firms classified as
operating in 2-digit GICS codes 10 (Energy), 15 (Minerals), 20 (Industrial), 50
(Telecommunications), and 55 (Utilities).
Table 4 reports results of FD3SLS estimates of the system of four equations (5) (8)
using the consumer and industrial subsamples and replacing DESM with the deviation from
expected investment in managing stakeholder strengths (DESMSTR) or concerns (DESMCON),
or deviation from expected investment in social issue participation (DESIP).
29
Columns 1 and 2
contain estimated results for regression equation (5) with DESM as the dependent variable for the
consumer and industrial samples, respectively. For the consumer subsample in column 1, the
estimated coefficient for BRDIND is negative and significant at the 1% level (z = 4.79). None of
the estimated coefficients for LNDELTA, DUALITY, or INSTOWN are significant. For the
industrial subsample in column 2, none of the estimated coefficients for LNDELTA, BRDIND,
DUALITY, or INSTOWN are significant.
-----Insert Table 4 About Here-----
Columns 3 and 4 contain estimated results with DESMSTR as the dependent variable for
the consumer and industrial subsamples, respectively. The estimated coefficients for BRDIND are
negative and significant at the 1% level for both subsamples. Columns 5 and 6 contain estimated

29
To conserve space in tables 5 and 6, we only report results for the main equation. Results from the first stage
regressions are reported in Appendix C and available from the authors upon request.
40

results with DESMCON as the dependent variable for the consumer and industrial subsamples,
respectively. The estimated coefficients for BRDIND are no longer significant. However, the
estimated coefficients for DUALITY are now negative and significant for both the industrial and
consumer subsamples. The estimated coefficient for INSTOWN remains negative and significant
for the consumer subsample, but is now positive and significant for the industrial subsample.
With the exception of a negative and significant estimated coefficient for INSTOWN in the
industrial subsample, our results with DESIP as the dependent variable in columns 7 and 8 are
similar to those in columns 1 and 2.
Overall, these results suggest that the effect of governance mechanisms varies by industry
and whether the deviation in SM is related to strengths or concerns. For example, proxies for
effective board monitoring are more effective at reducing deviations from expected SM strengths
than concerns, while duality is more effective at reducing deviations from concerns. Furthermore,
higher concentrations of institutional ownership lead to significantly lower deviations from both
strengths and concerns for consumer oriented firms, whereas, they increases deviations from both
strengths (insignificant) and concerns for industrial oriented firms.
Table 5 reports results of FD3SLS estimates of the system of four equations (5) (8)
using the consumer and industrial subsamples and replacing DESM with measures of deviation
from expected investment in the 5 primary stakeholder dimension categories of employee
relations (DEEMP), diversity issues (DEDIV), product issues (DEPRO), community relations
(DECOM), and environmental issues (DEENV). Columns 1 and 2 contain estimated results for
regression equation (5) with DEEMP as the dependent variable for the consumer and industrial
samples, respectively. In both the consumer and industrial subsamples, the estimated coefficients
for BRDIND are negative and significant, while the estimated coefficients for INSTOWN are
41

insignificant. The estimated coefficient for DUALITY for the consumer subsample is also
positive and significant. As noted earlier, both consumer and industrial firms are likely to benefit
from better employee relations equally. This is likely to also influence the effect of governance
mechanisms on investment in the employee dimension of SM.
-----Insert Table 5 About Here-----
Columns 3 and 4 contain estimated results for regression equation (5) with DEDIV as the
dependent variable for the consumer and industrial samples, respectively. The estimated
coefficient for INSTOWN in the consumer subsample is now positive and significant.
Conversely, the estimated coefficient for BRDIND in the industrial subsample is negative and
significant. While higher levels of institutional ownership lead to greater deviations from
expected diversity related SM for consumer oriented firms; more effective board structure reduces
such deviations for industrial oriented firms. One possible explanation for this result is that
abnormal investment in diversity related SM may benefit consumer oriented firms more than
industrial oriented firms in the marketplace via differentiation or by attracting customers willing
to pay a premium for such products. As a result, institutional investors may exert a positive
influence in this dimension for consumer oriented firms, while effective boards may seek to
constrain similar investments in this dimension for industrial oriented firms.
Columns 5 and 6 contain estimated results for regression equation (5) with DEPRO as the
dependent variable for the consumer and industrial subsamples, respectively. The estimated
coefficient for BRDIND in the consumer subsample is negative and significant, whereas it is
positive and significant for the industrial subsample. This suggests that board monitoring is a less
effective mechanism for controlling deviations from expected product related SM for industrial
oriented firms. The estimated coefficient for LNDELTA, however, is also negative and significant
42

for the industrial subsample. Therefore, while board monitoring may not effectively constrain
deviations from product related SM for industrial firms, CEO incentives are more effective at
doing so for this subset of firms. Lastly, the estimated coefficients for INSTOWN are negative
and significant for both the consumer and industrial subsamples, suggesting that institutional
ownership also plays an important role in constraining needless investment in product related SM.
Columns 7 and 8 contain estimated results for regression equation (5) with DECOM as the
dependent variable for the consumer and industrial samples, respectively. The estimated
coefficients for BRDIND, DUALITY, and INSTOWN are negative and significant in the
consumer subsample. None of the estimated coefficients for LNDELTA, BRDIND, DUALITY,
or INSTOWN are significant for the industrial subsample. A closer examination of the KLD
strength indicators for the community dimension shows that it includes screens for charitable
giving, innovative giving, non-US charitable giving, support for education, support for housing,
and volunteer programs. Investments in these areas may play a role in differentiating ones
product or attracting customers willing to pay a premium. As a result, consumer oriented firms
should benefit more from investment in community related SM than industrial firms. However,
consumer firms may also be exposed to a greater risk of overinvestment in this dimension.
Consequently, the negative and significant estimated coefficients on BRDIND and INSTOWN
suggest that direct monitoring by the board and institutional investors may provide one
mechanism for limiting unnecessary investment in community related SM investment for
consumer oriented firms.
Columns 9 and 10 contain estimated results for regression equation (5) with DEENV as
the dependent variable for the consumer and industrial samples, respectively. The estimated
coefficient for INSTOWN is now positive and significant in the consumer subsample. Examining
43

the first stage regressions for INSTOWN the estimated coefficient for DEENV is positive and
significant for the consumer subsample. In contrast to the community related SM regressions, this
indicates that institutional investors, or at least some subset of institutional investors, perceive
extra investment in environmental related SM as providing value for consumer oriented firms.
Overall, our results suggest that the effect of governance mechanisms varies by industry
and SM dimension. With a few exceptions, we find that governance mechanisms, especially
board independence, play a similar role in constraining deviations from expected SM in
dimensions expected to influence consumer and industrial firms similarly (i.e. employee
relations). Conversely, we find that governance mechanisms play a differential role in
constraining deviations from expected SM in dimensions expected to influence consumer and
industrial firms differently (e.g. diversity, community, environment). Furthermore, while we fail
to find a significant relation between institutional ownership and overall measures of deviations
from expected SM, in many cases we find a significant negative (or positive) relation on the
estimated coefficient for individual dimensions. However, this effect varies depending upon
whether the firm has a consumer or industrial orientation. Our results suggest that institutional
owners constrain needless investment in only specific dimensions of SM and that their influence
varies by consumer or industrial orientation of the firm.
6. Conclusions
In this paper, we examine the relation between deviations from expected stakeholder
management and corporate governance. The goal of the corporation is to maximize shareholder
wealth and this goal cannot be achieved by ignoring the interests of other constituents (Jensen,
2002; Brickley, Smith, and Zimmerman, 2002). To cater to the interests of various stakeholders
inevitably consumes a firms resources. Hillman and Keim (2001) document evidence that
44

building good relations with primary stakeholders (such as employees, customers, suppliers, and
communities) lead to increased shareholder wealth and consuming resources in social issue
participation that is unrelated to a firms primary stakeholders results in decreased shareholder
value. Even though Hillman and Keim (2001) do not explicitly use the word excess or
insufficient stakeholder management, their empirical results support the notion that there should
be an expected level of stakeholder management. Inefficient managers may have incentives to
gain support from stakeholders and use stakeholder management as an entrenchment strategy
(Cespa and Cestone, 2007). We propose that deviations from expected investment in stakeholder
management lead to agency costs similar to executive perquisite consumption. Theoretically, high
quality corporate governance mechanisms, including managerial ownership incentives, effective
board monitoring, and institutional ownership, can mitigate the agency problems caused by
deviations from expected stakeholder management level.
Using two stakeholder management deviation measures, the fixed-effects (FE) estimates
show that high managerial incentives and large boards are related to larger deviation from
expected stakeholder management. These findings support both the CEO perquisite and the board
monitoring hypotheses. However, we do not find that more independent boards mitigate excess
stakeholder investment. Consistent with Brickley et al. (1997), we provide some evidence
supporting the idea that CEOs are awarded with COB titlesCEO/Chair duality significantly
control deviations from expected stakeholder management. We also find that the relation between
corporate governance and stakeholder management is contingent upon a firms financial
constraints. Specifically, managers are prone to over-investment in stakeholder management
when they are less financially constrained.
45

We investigate the different roles of governance mechanisms on under- (negative
deviation from the expected level) and over- investment (positive deviation from the expected
level) in stakeholder management by multinomial logit regression. Our results show that effective
board monitoring, proxied by small and more independent boards, reduce the risk of over-
investment in stakeholder management, which is consistent with board monitoring hypothesis.
Given that internal governance variables may be endogenously determined with excess
stakeholder management, we address this issue with two approaches: lagged value models and a
four-equation system with 3-stage least square regression. Our results support the board
monitoring hypotheses after controlling for endogeneity. We find that small, more independent
boards and CEO duality effectively control deviations from expected stakeholder management.
We also examine how governance mechanisms affect deviations from expected
stakeholder management, strength, concerns, and each individual dimensions by industry. We
estimate first difference 3SLS regression on the four-equation system by industry using
subsamples of customer oriented firms and industrial oriented firms according to 2-digit GICS
codes. We find that governance mechanisms affect management stakeholder investment behavior
differently by industry. For example, effective board monitoring significantly controls deviations
from expected stakeholder management in consumer oriented firms, but not in industrial oriented
firms. We find the institutional ownership significantly controls deviations from expected social
issue participation and encourages deviations from expected stakeholder management concerns in
industrial oriented firms, but significantly reduces the deviations from both expected stakeholder
management strength and concerns in consumer oriented firms. Similarly, we find that board
independence significantly controls deviations from expected investment in product issues in
consumer oriented firms, but board independence significantly increases deviations from expected
46

investment in product issues in industrial oriented firms. We also document some similarities.
For instance, the effects of board independence on deviations from expected stakeholder
management strength and expected employee relation are similar for the two subsamples.
CEO/Chair duality plays a similar role in constraining deviations from expected stakeholder
management concerns for both consumer and industrial oriented firms.
Our results show that good corporate governance is consistent with Enlightened Value
Maximization theory which says the goal of a corporation is to maximize shareholder wealth;
however, this goal can only be achieved by taking care of stakeholders without either under-or
over investing in stakeholder management. The results are also a reflection of stakeholder
salience, suggesting managers must consider power, legitimacy, and urgency (Mitchell, Agle, and
Wood, 1997) to determine the amount to invest in various stakeholders accordingly to avoid
investing in non-value increasing stakeholder management.
This study is important because we are the first to investigate whether and how
governance system controls value unrelated stakeholder management. From a practical point of
view, our results provide some guidelines for management and boards in dealing with various
stakeholders and effectively monitoring management policy on stakeholder management. Firms
operating in different industries should adopt different approaches in dealing with stakeholders
because the same governance mechanisms may play different roles in controlling deviations from
the expected stakeholder management level. Our conclusions depend, however, on the ability of
our measure of expected stakeholder management to accurately capture the optimal level of the
stakeholder management. Future research directed at measuring optimal stakeholder management
may provide additional insights on this issue.
47

The study also sheds light on the dilemma faced by managers when called to serve an
expanded role in society. Our results suggest managers should always keep expected level of
stakeholder management in mind to avoid either excessive or inadequate investment in
stakeholder management in the pursuit of shareholder wealth maximization.
This study may also provide important implication for regulators. The regulators may need
to further amend the current perquisite disclosure requirements and provide clearer guideline on
the definition of perquisites to avoid perquisite consumption on stakeholder management that
hurts shareholder interests.
48

References
Adams, R. B., and Santos, J. A., 2006. Identifying the effect of managerial control on firm
performance. Journal of Accounting and Economics, 41(1-2), 55-85.
Adams, R., Ferreira, D., 2007. A theory of friendly boards. Journal of Finance 62, 217250.
Agle, B. R., Mitchell, R. K., and Sonnenfeld, J. A., 1999. Who matters to CEOs? An investigation
of stakeholder attributes and salience, corporate performance, and CEO values. Academy
of Management Journal 42, 507-525.
Agrawal, A., and Knoeber, C. R., 1996. Firm Performance and Mechanisms to Control Agency
Problems between Managers and Shareholders. The Journal of Financial and Quantitative
Analysis, 31(3), 377-397.
Amihud, Y., and Lev, B., 1981. Risk Reduction as a Managerial Motive for Conglomerate
Mergers. The Bell Journal of Economics, 12(2), 605-617.
Anderson, R. C., and Reeb, D. M., 2003. Founding-Family Ownership and Firm Performance:
Evidence from the S&P 500. The Journal of Finance, 58(3), 1301-1328.
Andrews, A.B., Linn, S.C. and Yi, H., 2009. Corporate Governance and Executive Perquisites:
Evidence from the New SEC Disclosure Rules. Unpublished working paper, Wayne State
University
Baker, G., Hall, B., 2004. CEO incentives and firm size. Journal of Labor Economics 22, 767-
798.
Baltagi, B.H., 2005. Econometric Analysis of Panel Data. John Wiley and Sons.
Banerjee, S., Dasgupta, S., and Kim Y., 2008. Buyer-Supplier Relationships and the Stakeholder
Theory of capital structure. Journal of Finance, 63, 2507-2552.
Bartkus, B., Glassman, M., and McAfee, B., 2006. Mission Statement Quality and Financial
Performance. European Management Journal 24(1), 86-94.
Baysinger, B. D., and Butler, H. N., 1985. The Role of Corporate Law in the Theory of the Firm.
Journal of Law and Economics, 28(1), 179-191.
Bebchuk, L. A., Cremers, K. M., and Peyer, U. C., 2007. CEO Centrality. Working paper,
Harvard University.
Belsey, D. A., Kuh, E., and Welsch, R. E., 1980. Regression Diagnostics: Identifying Influential
Data and Sources of Collineartity. New York: Wiley.
Benson, B. W., and Davidson III, W. N., 2009. Reexamining the managerial ownership effect on
firm value. Journal of Corporate Finance, 15(5), 573-586.
Benson, B. W., and Davidson III, W. N., 2009. The Relation Between Stakeholder Management,
Firm Value, and CEO Compensation: A Test of Enlightened Value Maximization.
Financial Management, forthcoming.
Berger, P., Ofek, E., Yermack, D., 1997. Managerial entrenchment and capital structure decisions.
Journal of Finance 52, 14111438.
Berman, S. L., Wicks, A. C., Kotha, S., and Jones, T. M., 1999. Does stakeholder orientation
matter? The relationship between stakeholder management models and firm financial
performance. Academy of Management Journal 42, 488-506.
Berrone, P., Surroca, J., and Trib, J., 2007. Corporate Ethical Identity as a Determinant of Firm
Performance: A Test of the Mediating Role of Stakeholder Satisfaction. Journal of
Business Ethics 76(1), 35-53.
Biddle, G., Hilary, G., and Verdi, R., 2009. How does financial reporting quality relate to
investment efficiency? Journal of Accounting and Economics 48, 112-131.
49

Bhagat, S., Bolton, B., 2008. Corporate governance and firm performance. Journal of Corporate
Finance, 14, 257-273.
Bhattacharya, C., Korschun, D., and Sen, S., 2009. Strengthening StakeholderCompany
Relationships Through Mutually Beneficial Corporate Social Responsibility Initiatives.
Journal of Business Ethics, 85, 257-272
Bird, R.D., Hall, A., Moment, F., and Reggiani, F., 2007. What Corporate Social Responsibility
Activities are Valued by the Market? Journal of Business Ethics 76(2), 189-206.
Booth, J.R., Cornett, M.M. and Tehranian, H., 2002. Boards of directors, ownership, and
regulation. Journal of Banking and Finance, 26, 1973-1996.
Brammer, S., Brooks, C. and Pavelin, S., 2006. Corporate Social Performance and Stock Returns:
UK Evidence from Disaggregate Measures. Financial Management, 35(3), 97-116
Brick, I. E., Palia, D., and Wang, C., 2005. Simultaneous estimation of CEO compensation,
leverage, and board characteristics on firm value. Working Paper, Rutgers University.
Brickley, J. A., Coles, J. L., and Terry, R. L., 1994. Outside directors and the adoption of poison
pills. Journal of Financial Economics, 35(3), 371-390.
Brickley, J. A., Lease, R. C., Smith, C.W., 1988. Ownership Structure and Voting on
Antitakeover Amendments, Journal of Financial Economics 20, 267-291.
Brickley, J., Coles, J., Jarrell, G., 1997. Leadership structure: separating the CEO and chairman of
the board. Journal of Corporate Finance 3, 189220.
Brickley, J.A., Smith Jr. , C.W. and Zimmerman, J.L., 2002. Business ethics and organizational
architecture. Journal of Banking and Finance, 26(9), 1821-1835.
Byrd, J. W., and Hickman, K. A., 1992. Do outside directors monitor managers?: Evidence from
tender offer bids. Journal of Financial Economics, 32(2), 195-221.
Carpenter, J. N., 2000. Does Option Compensation Increase Managerial Risk Appetite? The
Journal of Finance, 55(5), 2311-2331.
Carroll, A. B., 1993. Business and Society: Ethics and Stakeholder Management (South. Western
Publishing, Cincinnati.
Cespa , G., and Cestone , G., 2007. Corporate Social Responsibility and Managerial
Entrenchment. Journal of Economics & Management Strategy, 16(3), 741-771.
Chatterji, A. K., Levine, D. I., and Toffel, M. W., 2009. How well do social ratings actually
measure corporate social responsibility? Journal of Economics and Management Strategy
18(1), 125-169.
Chen, K.C., Chen, T., Hui, K., 2009. CEO perquisites and family firms. Unpublished working
paper, Hong Kong University of Science and Technology.
Cheung, W. A., and Wei, K. J., 2006. Insider ownership and corporate performance: Evidence
from the adjustment cost approach. Journal of Corporate Finance, 12(5), 906-925.
Cho, M., 1998. Ownership structure, investment, and the corporate value: An empirical analysis.
Journal of Financial Economics, 47(1), 103-121.
Cichello, M., 2005. The impact of firm size on payperformance sensitivities. Journal of
Corporate Finance 11, 609627.
Clarkson, M. B. E., 1995. A Stakeholder Framework for Analyzing and Evaluating Corporate
Social Performance. The Academy of Management Review, 20(1), 92-117.
Cohen, M. A., Fenn, S. A., and Naiman, J. S., 1995. Environmental and Financial Performance:
Are they related? Washington D.C.: Investor Responsibility Research Center.
Coles, J. L., Daniel, N. D., and Naveen, L., 2006. Managerial incentives and risk-taking. Journal
of Financial Economics, 79(2), 431-468.
50

Coles, J. L., Lemmon, M. L., and Meschke, F., 2007. Structural Models and Endogeneity in
Corporate Finance: the Link Between Managerial Ownership and Corporate Performance.
Working paper, Arizona State University.
Coombs, J. E., and Gilley, K. M., 2005. Stakeholder management as a predictor of CEO
compensation: main effects and interactions with financial performance. Strategic
Management Journal 26, 827-840.
Core, J. E., and Larcker, D. F., 2002. Performance consequences of mandatory increases in
executive stock ownership. Journal of Financial Economics, 64(3), 317-340.
Core, J., and Guay, W., 1999. The use of equity grants to manage optimal equity incentive levels.
Journal of Accounting and Economics, 28(2), 151-184.
Core, J., and Guay, W., 2002. Estimating the Value of Employee Stock Option Portfolios and
Their Sensitivities to Price and Volatility. Journal of Accounting Research, 40(3), 613-
630.
Core, J.E., and Guay, W.R., 2010. Is pay too high and are incentives too low? A wealth-based
contracting framework. Working paper, University of Pennsylvania.
Cotter, J.F., Shivdasani, A., and Zenner, M., 1997. Do Independent Directors Enhance Target
Shareholder Wealth during Tender Offers? Journal of Financial Economics 43, 195-218.
Davidson, R. and MacKinnon, J., 1993. Estimation and Inference in Econometrics. New York:
Oxford University Press.
Davidson, W. N., and Worrell, D. L., 1988. The impact of announcements of corporate illigalities
on shareholder returns. Acedemy of Management Journal 31, 195-200.
Davidson, W. N., Worrell, D. L., and El-Jelly, A., 1995. Influencing managers to change
unpopular behavior through boycotts and divestitures: A stock market test. Business and
Society 34, 171-196.
Davidson, W. N., Worrell, D. L., and El-Jelly, A., 1995. Influencing managers to change
unpopular behavior through boycotts and divestitures: A stock market test. Business and
Society 34, 171-196.
Davies, J., Hillier, D., and McColgan, P., 2005. Ownership structure, managerial behavior and
corporate value. Journal of Corporate Finance, 11(4), 645-660.
De Luque, M.S., Washburn, N.T., Waldman, D.A. and House, R.J., 2008. Unrequited profit: How
stakeholder and economic value relate to subordinantes perspective of leadership and
firm performance. Administrative Science Quarterly 53, 626-654.
Demsetz, H., and Villalonga, B., 2001. Ownership structure and firm performance. Journal of
Corporate Finance, 7, 209233.
Demsetz, H., 1983. The Structure of Ownership and the Theory of the Firm. The Journal of Law
and Economics, 26(2), 375.
Demsetz, H., Lehn, K., 1985. The structure of corporate ownership: Causes and consequences.
Journal of Political Economy 93, 1155-1177.
Donaldson, T., and Preston, L. E., 1995. The Stakeholder Theory of the Corporation: Concepts,
Evidence, and Implications. The Academy of Management Review, 20(1), 65-91.
Edmans, A., Gabaix, X., and Landier, A., 2009. A multiplicative model of optimal CEO
incentives in market equilibrium. Review of Financial Studies, 22(12), 4881-4917.
Edwards, F.R., Hubbard, R.G., 2000. The growth of institutional stock ownership: a promised
unfulfilled. Journal of Applied Corporate Finance, 13(3), 92-104.
Eisenberg, T., Sundgren, S. and Wells, M.T., 1998. Larger board size and decreasing firm value
in small firms. Journal of Financial Economics, 48(1), 35-54.
51

Evan, W. M., and Freeman, R. E., 1988. A stakeholder theory of the modern corporation: Kantian
capitalism. Ethical theory and business, 3, 97106.
Fama, E.F., 1980. Agency Problems and the Theory of the Firm. Journal of Political Economy,
88(2), 288.
Fama, E. F., and Jensen, M. C., 1983. Separation of Ownership and Control. The Journal of Law
and Economics, 26(2), 301-326.
Frank, M.Z., and Goyal, V.K., 2009. Capital Structure Decisions: Which Factors Are Reliably
Important? Financial Management, 38(1), 1-37.

Freeman, R. E., 1984. Strategic Management: A stakeholder approach. Boston: Pitman.
Freeman, R. E., 1994. The Politics of Stakeholder Theory: Some Future Directions. Business
Ethics Quarterly, 4(4), 409-421.
Freeman, R. E., 1999. Divergent stakeholder theory. Academy of Management Review, 24, 233
236.
Freeman, R. E., Wicks, A. C., and Parmar, B., 2004. Stakeholder Theory and "The Corporate
Objective Revisited ". Organization Science, 15(3), 364-369.
Friedman, A. L., and Miles, S., 2006. Stakeholders. Oxford University Press.
Friedman, M., 1970. The social responsibility of business is to create profits. New York Times
Magazine, September 13, 122-126.
Fung, M.K. (2009). Is innovativeness a link between pay and performance? Financial
Management 38(2), 411-429.
Garen, J., 1994. Executive compensation and principal-agent theory. Journal of Political
Economy 102, 1175 1199.
Gibbons, R., and Murphy, K. J. (1992). Optimal Incentive Contracts in the Presence of Career
Concerns: Theory and Evidence. The Journal of Political Economy, 100(3), 468-505.
Godfrey, P. C., 2005. The Relationship Between Corporate Philanthropy and Shareholder Wealth:
A Risk Management Perspective. Academy of Management Review, 30(4), 777-798.
Gompers, P., Ishii, J., and Metrick, A., 2003. Corporate governance and equity prices. Quarterly
Journal of Economics, 118(1), 107-155.
Graves, S. B., and Waddock, S. A., 1994. Institutional owners and corporate social performance.
Academy of Management Journal 37, 1034-1046.
Graves, S. B., and Waddock, S. A., 2000. Beyond built to last...Stakeholder relations in "Built-to-
Last" companies. Business and Society 105, 393-418.
Green, W. H., 2003. Econometric Analysis (5th ed.). Upper Saddle River, NJ: Prentice Hall.
Greening, D. W., and Turban, D. B., 2000. Corporate social performance as a competitive
advantage in attracting a quality workforce. Business and Society 39, 254-280.
Guay, W., 1999. The sensitivity of CEO wealth to equity risk: An analysis of the magnitude and
determinants. Journal of Financial Economics, 53(1), 43-71.
Hadlock, C., Lumer, G., 1997. Compensation, turnover, and top management incentives:
historical evidence. Journal of Business 70, 153 187.
Hall, B., Liebman, J., 1998. Are CEOs really paid like bureaucrats? The Quarterly Journal of
Economics CXIII, 653-691.
Hart, S. L., and Ahuja, G., 1996. Does it pay to be green? an empirical examination of the
relationship between emission reduction and firm performance. Business Strategy and the
Environment, 5(1), 30-37.
Hartzell, J. C., and Starks, L. T., 2003. Institutional Investors and Executive Compensation. The
52

Journal of Finance, 58(6), 2351-2374.
Haugen, R., Senbet, L., 1981. Resolving the agency problems of external capital through options.
Journal of Finance 36, 629-47.
Hausman, J. A., 1978. Specification Tests in Econometrics. Econometrica, 46(6), 1251-1271.
Hayes, R.M., Lemmon, M., and Qiu, M., 2010. Stock options and managerial incentives for risk-
taking: Evidence from FAS 123R. Working paper, University of Utah.
Hermalin, B. E., and Weisbach, M. S., 1991. The Effects of Board Composition and Direct
Incentives on Firm Performance. Financial Management, 20(4), 101-112.
Hermalin, B., Weisbach, M., 1998. Endogenously chosen boards of directors and their monitoring
of the CEO. American Economic Review 88, 96118.
Hermalin, B.E., Weisbach, M.S., 2003. Boards of directors as an endogenously determined
institution: A survey of the economic literature. FRBNY Economic Policy Review 9, 7
26.
Hillman, A., and Keim, G., 2001. Shareholder value, stakeholder management, and social issues:
What's the bottom line? Strategic Management Journal, 22(2), 125-139.
Himmelberg, C. P., Hubbard, R. G., and Palia, D., 1999. Understanding the determinants of
managerial ownership and the link between ownership and performance. Journal of
Financial Economics, 53(3), 353-384.
Holderness, C. G., 2003. A Survey of Blockholders and Corporate Control. SSRN eLibrary.
Holderness, C. G., Kroszner, R. S., and Sheehan, D. P., 1999. Were the Good Old Days That
Good? Changes in Managerial Stock Ownership since the Great Depression. The Journal
of Finance, 54(2), 435-469.
Hubbard, R. G., and Palia, D., 1995. Benefits of Control, Managerial Ownership, and the Stock
Returns of Acquiring Firms. The RAND Journal of Economics, 26(4), 782-792.
Jensen, M. C., 1993. The Modern Industrial Revolution, Exit, and the Failure of Internal Control
Systems. The Journal of Finance, 48(3), 831-880.
Jensen, M. C., 2001. Value Maximization, Stakeholder Theory, and the Corporate Objective
Function. Journal of Applied Corporate Finance, 14, 821.
Jensen, M. C., 2002. Value Maximization, Stakeholder Theory, and the Corporate Objective
Function. Business Ethics Quarterly, 12(2), 235-256.
Jensen, M., and Meckling, W., 1976. Theory of the firm: Managerial behavior, agency costs, and
capital structure. Journal of financial economics, 3(4), 305360.
Jensen, M., Murphy, K., 1990. Performance pay and top management incentives. Journal of
Political Economy 98, 225-264
Jiambalvo, J. J., Rajgopal, S., and Venkatachalam, M., 2002. Institutional Ownership and the
Extent to Which Stock Prices Reflect Future Earnings, Contemporary Accounting
Research 19, 117-145.
Johnson, R. A., and Greening, D. W., 1999. The effects of corporate governance and institutional
ownership types on corporate social performance. Academy of Management Journal 42,
564-576.
Jin, L., 2002. CEO compensation, diversification, and incentives. Journal of Financial Economics
66, 2963.
Ju, N., Leland, H. E., and Senbet, L. W., 2002. Options, Option Repricing and Severance
Packages in Managerial Compensation: Their Effects on Corporate Risk. Working paper,
University of Maryland.
Kale, J. R., and Shahrur H., 2007. Corporate Capital Structure and the Characteristics of Supplier
53

and Customer Markets. Journal of Financial Economics, 83, 321-365.
Karpoff, J. M., Lott, J. R., and Rankine, G., 1998. Environmental violations, legal penalties, and
contractor influence. Working paper, University of Washington.
Karpoff, J.M. and Lott, J.R., 1993. The reputational penalty firms bear from committing criminal
fraud. Journal of Law and Economics 36, 757-802.
Karpoff, J.M., Lee, D.S. and Vendrzyk, V.P., 1999. Defense procurement fraud, penalties and
contractor influence. Journal of Political Economy, 107, 809-842.
Kassinis, G. and Vafeas, N., 2002. Corporate Boards and Outside Stakeholders as Determinants
of Environmental Litigation. Strategic Management Journal, 23(5), 399-415.
Keay, A., 2009. Shareholder primacy: Can it survive? Should it survive? Working paper,
University of Leeds.
Kelly, K. and Davis, A., 2007. Goldmans Green Streak is Questioned as Two Investors Seek
Focus on Profit. Wall Street Journal Online.(March 27, page c3).
Klassen, R., and Whybark, D., 1999. The impact of environmental technologies on manufacturing
performances. Academy of Management Journal 42, 599-615.
Kleibergen, F., and Paap, R., 2006. Generalized reduced rank tests using the singular value
decomposition. Journal of Econometrics, 133(1), 97-126.
Konar, S., and Cohen, M., 2001. Does the market value environmental performance? Review of
Economics and Statistics 83, 281-289.
Lang, L. H. P., and Stulz, R. M., 1994. Tobin's q, Corporate Diversification, and Firm
Performance. Journal of Political Economy, 102(6), 1248-1280.
Laplume, A. O., Sonpar, K., and Litz, R. A., 2008. Stakeholder Theory: Reviewing a Theory That
Moves Us. Journal of Management, 34(6), 1152-1189.
Laurent, 2007. Neither Shareholder nor Stakeholder Management: What Happens When Firms
are Run for their Short-term Salient Stakeholder? European Management Journal, 25(2),
146-162.
Lewellen, K., 2006. Financing decisions when managers are risk averse. Journal of Financial
Economics, 82(3), 551-589.
Linck, J., Netter, J., and Yang, T., 2008. The determinants of board structure. Journal of Financial
Economics, 87(2), 308-328.
Lipton, M. and Lorsch, J.W., 1992. A Modest Proposal for Improved Corporate Governance,
Business Lawyer (ABA), 48-59.
Loderer, C., and Martin, K., 1997. Executive stock ownership and performance Tracking faint
traces. Journal of Financial Economics, 45(2), 223-256.
Long, J.S., 1997. Regression models for categorical and limited dependent variables. Sage
Publications, Thousand Oaks.
Maksimovic, V., and Titman, S., 1991. Financial Policy and Reputation for Product Quality.
Review of Financial Studies, 4, 175-200.
Masulis, R.W., "The Impact of Capital Structure Change on Firm Value: Some Estimates,"
Journal of Finance 38, March 1983, pp. 107-126
Maug, E., 1997. Boards of directors and capital structure: alternative forms of corporate
restructuring. Journal of Corporate Finance 3, 113139.
McConnell, J. J., and Servaes, H., 1990. Additional evidence on equity ownership and corporate
value. Journal of Financial Economics, Journal of Financial Economics, 27(2), 595-612.
McConnell, J. J., Servaes, H., and Lins, K. V., 2008. Changes in insider ownership and changes in
the market value of the firm. Journal of Corporate Finance, 14(2), 92-106.
54

McConnell, J., and Servaes, H., 1995. Equity ownership and the two faces of debt. Journal of
Financial Economics, 39(1), 131-157.
McDonnelll, B., 2004. Corporate constituency statutes and employee governance. William
Mitchell Law Review, 30 (4), 1227-1259.
Mehran, H., 1995. Executive compensation structure, ownership, and firm performance. Journal
of Financial Economics, 38(2), 163-184.
Meznar, M. B., Nigh, D., and Kwok, C. C. Y., 1994. Effect of Announcements of Withdrawal
from South Africa on Stockholder Wealth. The Academy of Management Journal 37(6),
1633-1648.
Mitchell, R. K., Agle, B. R., and Wood, D. J., 1997. Toward a Theory of Stakeholder
Identification and Salience: Defining the Principle of Who and What Really Counts. The
Academy of Management Review, 22(4), 853-886.
Moneva, J. M., Rivera-Lirio, J. M., and Munoz-Torres, M. J., 2007. The corporate stakeholder
commitment and social and financial performance. Industrial Management & Data
Systems 107(1), 84 - 102.
Moore, G., 2001. Corporate Social and Financial Performance: An Investigation in the U.K.
Supermarket Industry. Journal of Business Ethics 34(3/4), 299-315.
Morck, R., Shleifer, A., and Vishny, R., 1988. Management ownership and market valuationAn
empirical analysis. Journal of Financial Economics, 20, 293-315.
Mundlak, Y., 1978. On the pooling of time series and cross section data. Econometrica, 46, 69-85.
Murphy, K., 1999. Executive compensation. In: Ashenfelter, O., Card, D. (Eds.), Handbook of
Labor Economics Vol. 3b, Elsevier Science, North Holland.
Myers, S. C. 1977. Determinants of corporate borrowing. Journal of Financial Economics, 5(2),
147-175.
Neubaum, D.O. and Zahra, S.A., 2006. Institutional Ownership and Corporate Social
Performance: The Moderating Effects of Investment Horizon, Activism, and
Coordination. Journal of Management, 32(1), 108-131.
Ogden, S., and Watson, R., 1999. Corporate performance and stakeholder management: balancing
shareholder and customer interests in the U.K. privatized water industry. Academy of
Management Journal 42(5), 526-538.
Omran, M., Atrill, P., and Pointon, J., 2002. Shareholders versus stakeholders: corporate mission
statements and investor returns. Business Ethics: A European Review 11(4), 318-326.
Palia, D., 2001. The Endogeneity of Managerial Compensation in Firm Valuation: A Solution.
The Review of Financial Studies, 14(3), 735-764.
Parrino, R., Poteshman, A. M., and Weisbach, M. S., 2005. Measuring investment distortions
when risk-averse managers decide whether to undertake risky projects. Financial
Management, 34(1),2160.
Pukthuanthong, K., Roll, R., and Walker, T., 2007. How employee stock options and executive
equity ownership affect long-term IPO operating performance. Journal of Corporate
Finance, 13(5), 695-720.
Raheja, C., 2005. Determinants of board size and composition: a theory of corporate boards.
Journal of Financial and Quantitative Analysis 40, 283306.
Rajan, R.G. and Wulf, J., 2006. Are perks purely managerial excess? Journal of Financial
Economics, 79(1), 1-33.
Rogers, W. H., 1993. Regression standard errors in clustered samples, Stata Technical Bulletin,
13, 1923.
55

Rosenstein, S., and Wyatt, J. G., 1990. Outside directors, board independence, and shareholder
wealth. Journal of Financial Economics, 26(2), 175-191.
Ross, S. A., 2004. Compensation, Incentives, and the Duality of Risk Aversion and Riskiness.
The Journal of Finance, 59(1), 207-225.
Ruf, B. M., Muralidhar, K., Brown, R. M., Janney, J. J., and Paul, K., 2001. An Empirical
Investigation of the Relationship Between Change in Corporate Social Performance and
Financial Performance: A Stakeholder Theory Perspective. Journal of Business Ethics
32(2), 143-156.
Schaefer, S., 1998. The dependence of payperformance sensitivity on the size of the firm.
Review of Economics and Statistics 80, 436 443.
Schrand, C., and Unal, H., 1998. Hedging and Coordinated Risk Management: Evidence from
Thrift Conversions. The Journal of Finance, 53(3), 979-1013.
Shea, J., 1997. Instrument Relevance in Multivariate Linear Models: A Simple Measure. Review
of Economics and Statistics, 79(2), 348-352.
Smith, C. J., and Watts, R. L., 1992. The investment opportunity set and corporate financing,
dividend, and compensation policies. Journal of Financial Economics, 32(3), 263-292.
Smith, C. W., and Stulz, R. M., 1985. The Determinants of Firms' Hedging Policies. The Journal
of Financial and Quantitative Analysis, 20(4), 391-405.
Starik, M., 1993. Is the environment an organizational stakeholder? Naturally. In Proceedings of
the Fourth Annual Meeting of the International Association for Business and Society (Vol.
466, p. 471).
Sundaram, A. K., and Inkpen, A. C., 2004. The Corporate Objective Revisited. Organization
Science, 15(3), 350-363.
Tian, Y. S., 2004. Too much of a good incentive? The case of executive stock options. Journal of
Banking and Finance, 28(6), 1225-1245.
Titman, S, 1984. The Effect of Capital Structure on a Firms Liquidation Decision. Journal of
Financial Economics, 13, 137-151.
Titman, S., and Wessels, R., 1988. The Determinants of Capital Structure Choice. Journal of
Finance, 43, 1-19.
Tong, Z., 2008. Deviations from optimal CEO ownership and firm value. Journal of Banking and
Finance, 32(11), 2462-2470.
Waddock, S. A., and Graves, S. B., 1997. The Corporate Social Performance-Financial
Performance Link. Strategic Management Journal 18(4), 303-319.
Wooldridge, J., 2002. Econometric Analysis of Cross Section and Panel Data. The MIT Press,
Cambridge.
Wright, P. et al., 1996. Impact of corporate insider, blockholder, and institutional equity
ownership on firm risk taking. Academy of Management Journal, 39(2), 441-463.
Wright,P., Kroll, M. and Elenkov, D., 2002. Acquisition returns, increase in firm size, and chief
executive officer compensation: the moderating role of monitoring. Academy of
Management Journal, 45(3), 599-608.
Yermack, D., 1996. Higher market valuation of companies with a small board of directors.
Journal of Financial Economics, 40(2), 185-211.
Yermack, D., 2006. Flights of fancy: Corporate jets, CEO perquisites, and inferior shareholder
returns. Journal of Financial Economics, 80(1), 211-242.

56

Table 1: Descriptive statistics and correlations
This table displays descriptive statistics and correlations between key variables for the sample. Variables include
deviation from expected stakeholder management (DESM) estimated using the residual from regression equation (1),
industry adjusted stakeholder management (IASM), and stakeholder management (SM), the dollar change in CEO
portfolio wealth for a 1% change in firm value scaled by total annual compensation (DELTA), the number of
directors serving on the board during the year (BRDSIZE), the percentage of outside directors serving on the board
(PIND), whether or not the CEO holds the title of CEO and COB (DUALITY), the percentage of shares held by
institutions during the year (INSTOWN), cash to total assets (CASH), total debt to total assets (TD/TA), the log of
sales (LN SALES), earnings before interest and taxes to total assets (ROA), 1year total return to shareholders
(TRS), Tobins Q (Q), the ratio of research and development expenses to total assets with missing values set equal to
zero (R&D/TA), the standard deviation of stock returns calculated over 36 months (VOLATILITY), net property,
plant, and equipment to total assets (CAP), the age of the CEO (AGE), the length of time the CEO has been in the
current position (TENURE), the salary and bonus compensation paid to the CEO (TCC), and the log of one plus the
ratio of monthly volume to number of shares outstanding in the month prior to the institutional ownership observation
(TURNOVER). Data are obtained from the KLD, COMPUSTAT, EXECUCOMP, IRRC, and 13F databases and
consist of an unbalanced panel of 9,051firm-year observations for 1,631 firms from 1995 to 2008. Panel A displays
descriptive statistics. Panel B displays correlations between key variables. Superscript * indicates a 5% level of
significance.

Panel A: Descriptive statistics
Variable N Mean S.d. Min. 25% Median 75% Max.
DESM 9051 0.02 0.35 0.88 0.20 0.02 0.23 0.93
IASM 9051 0.04 0.40 1.01 0.20 0.05 0.27 1.14
SM 9051 0.05 0.42 1.12 0.33 0.00 0.17 1.16
DELTA 9051 446.49 1,533.59 0.00 46.37 85.41 181.84 11,372.59
BRDSIZE 9051 9.66 2.44 5.00 8.00 9.00 11.00 17.00
PIND (%) 9051 29.82 15.13 8.33 18.18 28.57 40.00 72.73
DUALITY 9051 0.66 0.47 0.00 0.00 1.00 1.00 1.00
INSTOWN (%) 9051 75.41 16.87 32.45 64.15 76.70 88.43 100.00
CASH/TA 9051 0.09 0.10 0.00 0.02 0.05 0.13 0.47
TD/TA 9051 21.07 16.42 0.00 6.80 19.99 31.61 71.07
SALES ($Mil) 9051 6,669.43 12,396.53 94.06 862.17 2,142.88 6,220.01 81,303.00
ROA 9051 10.96 7.75 11.11 6.23 10.29 15.08 35.40
TRS 9051 15.65 39.65 65.35 9.67 11.43 35.38 161.26
Q 9051 2.14 1.31 0.87 1.31 1.71 2.48 8.19
R&D/TA 9051 151.70 481.73 0.00 0.00 2.46 69.00 3,513.00
VOLATILITY 9051 0.37 0.18 0.14 0.25 0.33 0.45 1.03
CAP 9051 27.04 21.49 0.32 10.49 21.33 38.49 88.85
AGE 9051 55.99 6.99 40.00 51.00 56.00 61.00 75.00
TENURE

9051 8.49 7.41 1.00 3.00 6.00 11.00 37.00
TCC ($k) 9051 1,557.68 1,338.74 110.00 732.31 1,136.57 1,917.06 8,300.27
TURNOVER

9051 0.84 0.50 0.09 0.46 0.75 1.14 2.37
57

Panel B: Correlations
DESM IASM SM DELTA BRDSIZE BRDIND DUALITY INSTOWN CASH/TA TD/TA LNSALES ROA TRS
DESM 1.00
IASM 0.88* 1.00
SM 0.81* 0.93* 1.00
LNDELTA 0.01 0.03* 0.06* 1.00
BRDSIZE 0.09* 0.08* 0.13* 0.05* 1.00
PIND 0.07* 0.08* 0.01 0.22* 0.01 1.00
DUALITY 0.03* 0.04* 0.01 0.11* 0.12* 0.07* 1.00
INSTOWN 0.01 0.02 0.13* 0.16* 0.26* 0.26* 0.08* 1.00
CASH/TA 0.01 0.04* 0.05* 0.08* 0.30* 0.01 0.12* 0.11* 1.00
TD/TA 0.01 0.06* 0.06* 0.15* 0.23* 0.06* 0.10* 0.02* 0.33* 1.00
LNSALES 0.01 0.02 0.05* 0.01 0.57* 0.11* 0.20* 0.17* 0.31* 0.24* 1.00
ROA 0.01 0.10* 0.11* 0.18* 0.00 0.05* 0.00 0.00 0.07* 0.15* 0.06* 1.00
TRS 0.00 0.00 0.00 0.12* 0.02* 0.03* 0.01 0.03* 0.11* 0.08* 0.04* 0.15* 1.00
58

Table 2: Fixed-effects (FE) regression and multinomial logit regression results of deviation from expected and conditional
stakeholder management
This table reports results of fixed-effects (FE) and multinomial regression estimates of equations (3) and (4). The dependent variable is deviation from expected
stakeholder management (DESM) estimated using the residual from regression equation (1), firms in the lowest two quintiles relative to the middle quintile of
DESM (Under-investment vs. Expected investment), firms in the highest two quintiles relative to the middle quintile of ESM (Over-investment vs. Expected
investment), industry adjusted stakeholder management (IASM), and stakeholder management (SM). Independent variables are the log of the dollar change in CEO
portfolio wealth for a 1% change in firm value scaled by total annual compensation (LNDELTA), the number of directors serving on the board during the year
(BRDSIZE), the percentage of outside directors serving on the board (PIND) , the ratio of independent directors to insider and affiliated directors multiplied by the
inverse of board size (BRDIND), whether or not the CEO holds the title of CEO and COB (DUALITY), the percentage of shares held by institutions during the year
(INSTOWN), a proxy for less financially constrained firms (OVER), cash to total assets (CASH), total debt to total assets (TD/TA), the log of sales (LN SALES),
ROA: measured as earnings before interest and taxes to total assets, 1year total return to shareholders (TRS). Data are obtained from the KLD, COMPUSTAT,
EXECUCOMP, IRRC, and 13F databases and consist of an unbalanced panel of 9,051 firm-year observations for 1,631 firms from 1995 to 2008. All regressions
include dummy variables year (The coefficients for year are not reported). The tstatistics are reported using robust standard errors clustered at the firm level (see
Rogers, 1993; Wooldridge, 2002). Robust tstatistics are given in parentheses. Superscripts */**/*** indicate levels of significance of 10%, 5%, and 1%,
respectively.
Method FE FE Multinomial Logit FE FE FE FE




Dependent Variable




Pred.




DESM




DESM
Under-
investment
vs.
Expected
investment
Over-
investment
vs.
Expected
investment
Under-
investment
vs.
Expected
investment
Over-
investment
vs.
Expected
investment




IASM




IASM




Pred.




SM




SM
Independent Variable Sign (1) (2) (3) (4) (5) (6) (7) (8) Sign (9) (10)
LNDELTA (+/) 0.00810** 0.00821** 0.0210 0.0259 0.0238 0.0225 0.0131*** 0.0134*** 0.0170 0.0178*
(2.18) (2.20) (0.79) (0.98) (0.90) (0.85) (2.99) (3.04) (1.63) (1.69)
BRDSIZE (+) 0.0106*** 0.0753*** 0.0457** 0.0139*** 0.00226
(3.01) (3.77) (2.25) (3.79) (0.31)
PIND () 0.000143 0.00255 0.00632** 0.000362 0.00117
(0.30) (0.93) (2.40) (0.71) (1.03)
BDIND () 0.0293 1.303* 2.008*** 0.179 0.443
(0.27) (1.90) (2.68) (1.46) (1.35)
DUALITY (+/) 0.0156 0.0170 0.133 0.00383 0.141* 0.00701 0.0266** 0.0283** 0.0475* 0.0457
(1.36) (1.48) (1.63) (0.05) (1.73) (0.09) (2.04) (2.13) (1.66) (1.58)
INSTOWN () 0.0000654 0.0000243 0.000538 0.00252 0.00102 0.00329 0.000382 0.000304 0.00161 0.00142
(0.14) (0.05) (0.20) (0.94) (0.39) (1.23) (0.75) (0.59) (1.45) (1.28)
DELTA x OVER (+/) 0.00762 0.00842
(0.54) (0.58)
BRDSIZE x OVER (+) 0.0304***
(2.63)
PIND x OVER () 0.00263
(1.60)
BRDIND x OVER 0.956**
(2.20)
59

DUALITY x OVER (+/) 0.122*** 0.123***
(2.62) (2.58)
INSTOWN x OVER () 0.00313* 0.00269*
(1.95) (1.67)
OVER 0.354* 0.0487
(1.77) (0.32)
CASH/TA 0.0104 0.0179 0.195*** 0.109*** 0.142*** 0.138*** 0.00848 0.000285
(0.52) (0.89) (5.44) (3.16) (4.37) (4.59) (0.35) (0.01)
TD/TA 0.000920* 0.000958* 0.00130 0.000370 0.000578 0.000735 0.0000772 0.000132
(1.77) (1.83) (0.49) (0.15) (0.22) (0.29) (0.14) (0.24)
LNSALES 0.0331 0.0349 0.328 0.500 0.453 0.445 0.0155 0.0212 0.00380 0.00480
(0.56) (0.59) (0.77) (1.22) (1.07) (1.08) (0.26) (0.35) (0.16) (0.20)
ROA 0.00207** 0.00217** 0.00568 0.00358 0.00551 0.00366 0.000266 0.000360 0.0000800 0.000127
(2.06) (2.13) (1.13) (0.75) (1.09) (0.76) (0.25) (0.33) (0.08) (0.12)
TRS
0.0000172 0.0000160 0.00131 0.00216*** 0.00138* 0.00211*** 0.000182** 0.000186** 0.0000883 0.0000845
(0.22) (0.21) (1.60) (2.63) (1.69) (2.58) (2.32) (2.36) (1.14) (1.09)

1
+
6
= 0 3.854** 3.962**

2
+
7
= 0 11.85*** 4.734***

3
+
8
= 0 1.762

4
+
9
= 0 4.566** 4.558**

5
+
10
= 0 2.131 1.488

Year Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes

n 1631 1631 1631 1631 1631 1631 1631 1631
N 9051 9051 9051 9051 9051 9051 9051 9051
Within R
2
0.0271 0.0244 0.0414 0.0377 0.0598 0.0546
Between R
2
0.000672 0.000716 0.00363 0.000425 0.0549 0.0551
Overall R
2
0.00556 0.00104 0.0109 0.00499 0.0507 0.0461
Pseudo R
2
0.0151 0.0114
60

Table 3: First difference threestage least squares (3SLS) regression results of deviation
from expected stakeholder management
This table reports results of first-difference threestage least squares (3SLS) regression estimates of equations (5) (8).
The endogenous variables are deviation from expected stakeholder management (DESM) estimated using the residual
from regression equation (1), the log of the dollar change in CEO portfolio wealth for a 1% change in firm value scaled
by total annual compensation (LNDELTA), the ratio of independent directors to insider and affiliated directors
multiplied by the inverse of board size (BRDIND), and the percentage of shares held by institutions during the year
(INSTOWN). Exogenous variables include whether or not the CEO holds the title of CEO and COB (DUALITY), cash
to total assets (CASH), total debt to total assets (TD/TA), the log of sales (LN SALES), earnings before interest and
taxes to total assets (ROA), and 1year total return to shareholders (TRS). Instrumental variables include Tobins Q
(Q), the ratio of research and development expenses to total assets with missing values set equal to zero (R&D/TA), the
standard deviation of stock returns calculated over 36 months (VOLATILTY), net property, plant, and equipment to
total assets (CAP), the salary and bonus compensation paid to the CEO (TCC), the length of time the CEO has been in
the current position (TENURE), the age of the CEO (AGE), and the log of one plus the ratio of monthly volume to
number of shares outstanding in the month prior to the institutional ownership observation (TURNOVER). Data are
obtained from the KLD, COMPUSTAT, EXECUCOMP, IRRC, and 13F databases and consist of an unbalanced panel
of 9,051 firm-year observations for 1,631 firms from 1995 to 2008. The zstatistics are given in parentheses.
Superscripts */**/*** indicate levels of significance of 10%, 5%, and 1%, respectively.

Dependent Variable Pred. DESM LNDELTA BRDIND INSTOWN
Independent Variable Sign (1) (2) (4) (5)
DESM 9.121*** 0.180*** 9.329
(2.64) (4.83) (0.82)
LNDELTA (+/) 0.00993 0.00131
(0.97) (0.69)
BRDIND () 5.060***
(6.26)
DUALITY (+/) 0.0266*** 0.00385**
(2.73) (1.98)
INSTOWN () 0.0000689
(0.07)
CASH/TA 0.149*** 0.0608 0.0282*** 1.422
(2.63) (0.15) (2.76) (0.74)
TD/TA 0.000667 0.00522 0.000118 0.0234
(1.39) (1.29) (1.27) (1.28)
LNSALES 0.0402** 0.0990 0.00741* 0.235
(1.97) (0.63) (1.95) (0.32)
ROA 0.00134 0.00834 0.000233 0.0987***
(1.60) (1.11) (1.40) (2.66)
TRS 0.00000572 0.00432*** 0.00000160 0.0196***
(0.08) (5.79) (0.12) (8.79)
Q 0.0373 0.000197
(0.66) (0.50)
RD/TA 0.000133 3.14e08
(1.05) (0.04)
VOLATILITY 0.163 0.00252
(0.72) (1.30)
CAP 0.0141***
(3.01)
LNTCC 0.327***
(7.13)
TENURE 0.0780***
(13.00)
AGE 0.0000478
(1.52)
TURNOVER 2.676***
(6.09)
CONSTANT 0.0235*** 0.00363 0.00475*** 2.031***
(3.26) (0.11) (6.47) (14.36)
61


n 1373
N 7033
62

Table 4: First difference threestage least squares (3SLS) regression results of deviation from expected stakeholder
management split by industry
This table reports results of first-difference threestage least squares (3SLS) regression estimates of equations (5) (8). The endogenous variables are deviation
from expected stakeholder management (DESM), deviation from expected stakeholder management strengths (DESMSTR), deviation from expected stakeholder
management concerns (DESMCON), and deviation from expected social issue participation (DESIP) estimated using the residual from regression equation (1), the
log of the dollar change in CEO portfolio wealth for a 1% change in firm value scaled by total annual compensation (LNDELTA), the ratio of independent directors
to insider and affiliated directors multiplied by the inverse of board size (BRDIND), and the percentage of shares held by institutions during the year (INSTOWN).
Exogenous variables include whether or not the CEO holds the title of CEO and COB (DUALITY), cash to total assets (CASH), total debt to total assets (TD/TA),
the log of sales (LN SALES), earnings before interest and taxes to total assets (ROA), and 1year total return to shareholders (TRS). Instrumental variables include
Tobins Q (Q), the ratio of research and development expenses to total assets with missing values set equal to zero (R&D/TA), the standard deviation of stock
returns calculated over 36 months (VOLATILTY), net property, plant, and equipment to total assets (CAP), the salary and bonus compensation paid to the CEO
(TCC), the length of time the CEO has been in the current position (TENURE), the age of the CEO (AGE), and the log of one plus the ratio of monthly volume to
number of shares outstanding in the month prior to the institutional ownership observation (TURNOVER). Data are obtained from the KLD, COMPUSTAT,
EXECUCOMP, IRRC, and 13F databases and consist of an unbalanced panel of 9,051 firm-year observations for 1,631 firms from 1995 to 2008. The zstatistics
are given in parentheses. Superscripts */**/*** indicate levels of significance of 10%, 5%, and 1%, respectively.
Sample Consumer Industrial Consumer Industrial Consumer Industrial Consumer Industrial
Dependent Variable Pred. DESM DESM DESMSTR DESMSTR DESMCON DESMCON DESIP DESIP
Independent Variable Sign (1) (2) (3) (4) (5) (6) (7) (8)
LNDELTA (+/) 0.00388 0.00861 0.000633 0.00593 0.00523 0.00252 0.000783 0.00680
(0.33) (0.35) (0.09) (0.35) (0.85) (0.19) (0.05) (0.31)
BRDIND () 4.838*** 5.010 2.793*** 9.853*** 0.166 0.437 6.620*** 2.848
(4.79) (1.19) (3.66) (3.76) (0.25) (0.15) (11.25) (0.66)
DUALITY (+/) 0.0186 0.0233 0.00273 0.0115 0.0216*** 0.0135** 0.00587 0.0256
(1.31) (1.01) (0.39) (0.77) (3.50) (2.11) (0.30) (1.32)
INSTOWN () 0.000801 0.00141 0.00557*** 0.00251 0.00478*** 0.00476*** 0.00136 0.00819***
(0.33) (0.48) (3.12) (1.04) (3.76) (3.67) (0.98) (3.54)
CASH/TA 0.0830 0.374* 0.0117 0.180 0.0686* 0.345*** 0.121 0.292
(1.41) (1.92) (0.31) (1.38) (1.71) (2.78) (1.61) (1.52)
TD/TA 0.000589 0.000629 0.000679** 0.000319 0.0000377 0.000371 0.00131* 0.000107
(1.11) (0.49) (2.01) (0.36) (0.10) (0.46) (1.88) (0.09)
LNSALES 0.0677*** 0.0173 0.0776*** 0.0685** 0.0255 0.0892*** 0.0867*** 0.0307
(2.76) (0.42) (4.92) (2.33) (1.52) (3.49) (2.76) (0.76)
ROA 0.00000295 0.00521*** 0.00210*** 0.000620 0.00219*** 0.00473*** 0.000384 0.00568***
(0.00) (2.59) (3.23) (0.43) (3.30) (3.83) (0.32) (2.93)
TRS 0.0000575 0.000129 0.000180*** 0.000274 0.0000642 0.00000211 0.0000782 0.0000762
(0.70) (0.52) (3.32) (1.54) (1.26) (0.01) (0.78) (0.32)
CONSTANT 0.0232** 0.0215 0.00838 0.0482** 0.0169*** 0.000640 0.0338*** 0.00248
(2.54) (0.79) (1.29) (2.56) (3.10) (0.03) (4.22) (0.09)

n 929 444 929 444 929 444 929 444
N 4602 2431 4602 2431 4602 2431 4602 2431

63

Table 5: First difference threestage least squares (3SLS) regression results of deviation from expected stakeholder
management dimensions split by industry
This table reports results of first-difference threestage least squares (3SLS) regression estimates of equations (5) (8). The endogenous variables are deviations
from expected investment in the 5 stakeholder dimension categories of employee relations (DEEMP), diversity issues (DEDIV), product issues (DEPRO),
community relations (DECOM), and environmental issues (DEENV) estimated using the residual from regression equation (1), the log of the dollar change in
CEO portfolio wealth for a 1% change in firm value scaled by total annual compensation (LNDELTA), the ratio of independent directors to insider and affiliated
directors multiplied by the inverse of board size (BRDIND), and the percentage of shares held by institutions during the year (INSTOWN). Exogenous variables
include whether or not the CEO holds the title of CEO and COB (DUALITY), cash to total assets (CASH), total debt to total assets (TD/TA), the log of sales
(LN SALES), earnings before interest and taxes to total assets (ROA), and 1year total return to shareholders (TRS). Instrumental variables include Tobins Q
(Q), the ratio of research and development expenses to total assets with missing values set equal to zero (R&D/TA), the standard deviation of stock returns
calculated over 36 months (VOLATILTY), net property, plant, and equipment to total assets (CAP), the salary and bonus compensation paid to the CEO (TCC),
the length of time the CEO has been in the current position (TENURE), the age of the CEO (AGE), and the log of one plus the ratio of monthly volume to
number of shares outstanding in the month prior to the institutional ownership observation (TURNOVER). Data are obtained from the KLD, COMPUSTAT,
EXECUCOMP, IRRC, and 13F databases and consist of an unbalanced panel of 9,051 firm-year observations for 1,631 firms from 1995 to 2008. The zstatistics
are given in parentheses. Superscripts */**/*** indicate levels of significance of 10%, 5%, and 1%, respectively.
Sample Consumer Industrial Consumer Industrial Consumer Industrial Consumer Industrial Consumer Industrial
Dependent Variable Pred. DEEMP DEEMP DEDIV DEDIV DEPRO DEPRO DECOM DECOM DEENV DEENV
Independent Variable Sign (1) (2) (3) (4) (5) (6) (7) (8) (9) (10)
LNDELTA (+/) 0.00420 0.0194 0.00247 0.0167 0.00159 0.0140** 0.00286 0.00277 0.000554 0.00152
(0.84) (1.49) (0.46) (1.26) (0.75) (1.97) (1.07) (0.45) (0.36) (0.23)
BRDIND () 2.480*** 3.719*** 0.742 7.592*** 1.061*** 3.404*** 1.073*** 1.455 0.211 2.685
(5.47) (6.00) (1.17) (3.28) (4.61) (8.26) (4.99) (1.00) (1.41) (1.51)
DUALITY (+/) 0.0198*** 0.0109 0.000415 0.00917 0.00971*** 0.00583 0.00734** 0.000604 0.00201 0.00557*
(3.70) (0.82) (0.06) (0.82) (3.87) (0.79) (2.29) (0.11) (1.24) (1.87)
INSTOWN () 0.00167 0.000163 0.00687*** 0.00192 0.00147*** 0.000941*** 0.00169*** 0.000792 0.000813*** 0.00152
(1.54) (0.67) (4.43) (1.54) (3.18) (3.31) (2.74) (0.79) (2.64) (1.53)
CASH 0.0630** 0.189** 0.0114 0.113 0.0462** 0.00986 0.0163 0.0852 0.0223* 0.185***
(2.33) (2.14) (0.36) (1.07) (2.31) (0.17) (1.10) (1.61) (1.86) (2.92)
TD/TA 0.000352 0.0000411 0.00000820 0.000803 0.000126 0.000683* 0.000263* 0.000234 0.000140 0.0000311
(1.44) (0.06) (0.03) (1.17) (0.68) (1.65) (1.95) (0.75) (1.27) (0.08)
LNSALES 0.00595 0.0165 0.0523*** 0.0452** 0.00779 0.0226 0.00906 0.00852 0.0121** 0.0220**
(0.53) (0.72) (4.01) (2.03) (0.93) (1.52) (1.46) (0.94) (2.42) (2.08)
ROA 0.000881* 0.00178 0.000194 0.000119 0.000281 0.00128* 0.000752*** 0.00125*** 0.00000656 0.00183***
(1.94) (1.62) (0.37) (0.11) (0.84) (1.80) (3.01) (2.82) (0.03) (3.54)
TRS 0.0000852** 0.000236** 0.0000742* 0.0000935 0.0000481* 0.000165*** 0.000000686 0.0000999 0.0000282* 0.000146*
(2.34) (2.46) (1.67) (0.71) (1.93) (2.75) (0.03) (1.45) (1.87) (1.78)
CONSTANT 0.0113*** 0.0238*** 0.0115* 0.0312** 0.00429** 0.0198*** 0.00155 0.00829 0.000562 0.0184*
(2.76) (4.20) (1.95) (2.17) (2.23) (5.53) (0.70) (0.97) (0.40) (1.76)

n 929 444 929 444 929 444 929 444 929 444
N 4602 2431 4602 2431 4602 2431 4602 2431 4602 2431

A-1

[The Appendices are not a formal part of the paper. They have been
included for use by the referees and will be made available to readers
upon request.]
Appendix A: Calculating stakeholder management
To construct measures for SM, we follow the procedure used in prior research in the area
(Waddock and Graves, 1997; Hillman and Keim, 2001; Coombs and Gilley, 2005), and construct
a measure of the firms stakeholder management (SM) performance using the KLD categories of
employee relations (EMP), diversity issues (DIV), product issues (PRO), community relations
(COM), and environmental issues (ENV). These five categories parallel the primary stakeholder
groups with regard to employees (including diversity initiatives), customers (product
safety/quality), the natural environment, the community, and suppliers (to the extent that certain
diversity initiatives are directed toward suppliers). Of note, the adapted KLD measure for SM
excludes issues outside of the primary stakeholder domains, but included in the domain of
corporate social performance. Following (Hillman and Keim, 2001), we include these excluded
issues when creating the variable social issue participation (SIP). The SIP variable includes the
KLD categories of human rights and controversial business issues (i.e. alcohol/
tobacco/gambling exclusionary screens, military exclusionary screens, and nuclear power
exclusionary screens). The seven stakeholder and social issue areas, as well as the strength and
concern indicators by dimension are listed in Table A1.
-----Insert Table A1 About Here-----
To measure SM and SIP we sum the number of positive (strength) and negative (concern)
indicators assigned to a company in a given year by stakeholder and social issue dimension. The
strength (concern) indicators are binary variables. Each is assigned a value of 1 when the
company has strengths (concerns) in each dimension. However, the number of KLD indicators
used to measure strengths and concerns varies from year to year. Therefore, in each year, we
create dimension (i.e. community relations, diversity, etc.) strength (concern) scores by taking
the sum of strengths (concerns) for each dimension and scaling by the total number of strength
(concern) indicators by dimension for each year.
1

We then create strength (concern) scores for SM and SIP by summing the relative
dimension strength (concern) scores for each year for each measure.

Finally, we create an
aggregate SM and SIP score variable by taking the SM (SIP) strength variable minus the SM
(SIP) concern variable for each year.


1
For example, KLD uses 6 indicators for community strengths in 1995 and 7 in 2002. Therefore, we sum up the
total strength indicators in 1995 and divide by 6 to obtain a Community strength score for 1995. We sum up the total
strength indicators in 2002 and divide by 7 to obtain a Community strength score for 2002. If a company scores 3 in
1995 and 2002, its community strength scores would be 3/6 = .5 and 3/7 = .4286, respectively. The community
strength score in 2002 would be slightly lower than 1995, due to the additional strength indicator used in 2002.
A-2

Table A1: KLD ratings indicators
This table provides a description of the KLD rating indicators by category.
Community
Strengths Concerns
Charitable Giving
Innovative Giving
Non-US Charitable Giving
Support for Education
Support for Housing
Volunteer Programs
Other Strengths
Investment Controversies
Negative Economic Impact
Tax Disputes
Other Concerns
Diversity
Strengths Concerns
Board of Directors
CEO
Employment of the Disabled
Gay and Lesbian Policies
Promotion
Women and Minority Contracting
Work/Life Benefits
Gay and Lesbian Policies
Other Strengths
Controversies
Non-Representation
Other Concerns
Employee Relations
Strengths Concerns
Health and Safety
Retirement Benefits
Union Relations
Cash Profit Sharing
Union Relations
Health and Safety
Retirement Benefits
Workforce Reductions
Other Concerns
Environment
Strengths Concerns
Beneficial Products and Services
Clean Energy
Management Systems
Pollution Prevention
Recycling
Other Strengths
Agricultural Chemicals
Climate Change
Hazardous Waste
Ozone Depleting Chemicals
Regulatory Problems
Substantial Emissions
Other Concerns
Human Rights
Strengths Concerns
Labor Rights
Relations with Indigenous Peoples
Other Strengths
Labor Rights
Relations with Indigenous Peoples
Burma
Other Concerns

A-3

Product
Strengths Concerns
Benefits the Economically Disadvantaged
Quality
R&D/Innovation
Other Strengths
Antitrust
Marketing/Contracting Controversy
Safety
Other Concerns
Controversial Business Issues
Alcohol Firearms
Licensing
Manufacture
Manufacturer of Products Necessary for
Alcoholic Beverages
Retailer
Ownership by an Alcohol Company
Ownership of an Alcohol Company
Manufacturer
Retailer
Ownership by a Firearms Company
Ownership of a Firearms Company
Gambling Military Weapons
Licensing
Manufacturer
Owner and Operator
Supporting Products or Services
Ownership by a Gambling Company
Ownership of a Gambling Company
Manufacturer of Weapons or Weapons Systems
Manufacturer of Components for Weapons or
Weapons Systems
Ownership by a Military Company
Ownership of a Military Company
Nuclear Power Tobacco
Ownership of Nuclear Power Plants
Construction and Design of Nuclear Power
Plants
Nuclear Power Fuel and Key Parts
Nuclear Power Service Provider
Ownership by a Nuclear Power Company
Ownership of a Nuclear Power Company
Licensing
Manufacturer
Manufacturer of Products Necessary for Tobacco
Products
Retailer
Ownership by a Tobacco Company
Ownership of a Tobacco Company
B-1

Appendix B: Calculating CEO portfolio delta
We use delta as a proxy for CEO equity-based compensation and ownership incentives.
Delta (DELTA) is defined as the dollar change in CEO portfolio wealth for a 1% change in firm
value scaled by total annual compensation:
EIIA =
_j# sborcs o stk. owncJ + (c
-dm
(N(
1
))(# sborcs o stk. opt. grontcJ in oworJ))

N
=1
[ (u .u1)(P)]
IC1

(B1)
where: d equals the dividend yield, m equals time to maturity, P equals the fiscal year closing
stock price, N(.) is the standard normal cumulative density function,

1
=
Iog[
P
X
+_]-d+
o
2
2
](m)
cm
, X equals the exercise price, rf equals the risk-free rate, equals
volatility, and TDC1 equals total direct compensation from the EXECUCOMP database.
We use data from the EXECUCOMP database to construct each CEOs aggregate stock
holdings at the end of each fiscal year. EXECUCOMP reports the aggregate number of shares,
which includes both contractually unrestricted and restricted stock. We estimate each CEOs
portfolio of stock options as in Core and Guay (2002). The procedure involves two steps,
determining the exercise price and time-to-maturity of new options, and estimating the exercise
price and time-to-maturity of previous option grants. We obtain the data for new grants from
EXECUCOMP, which reports the number of options granted for each grant series, the exercise
price for each series, the expiration date of each option, and the closing price of the firms stock
for the fiscal year. We obtain the fair value inputs used by each firm in calculating the fair value
of option grants post FASB 123R (volatility and dividend yield) from COMPUSTAT. When
these values are missing, we estimate volatility using the annualized standard deviation of
monthly stock returns calculated over 36 months from the CRSP database and the firms average
dividend yield over the past three years from the COMPUSTAT database. Missing values for
volatility or dividend yield are set equal to the sample median each year and all volatility and
dividend yield values are winsorized annually at the upper and lower 1%. These six inputs, along
with the risk free rate of interest, allow for the calculation of the incentive effect of executive
stock option grants each year
1
We estimate the exercise price of the unexercisable options by
finding the difference between the fiscal year end stock price and value of unexercisable in-the-
money options, excluding the value of new option grants, to the number of unexercisable
options, excluding the number of new grants.
2
To estimate the exercise price of exercisable

1
We use the risk-free interest rate that is the approximate average yield that could have been earned in a particular
year by investing in a U.S. Treasury bond carrying a term equal to the term of each option grant after taking a 30%
haircut. Other researchers have employed a risk-free rate equal to the rate of interest on a ten-year constant-maturity
Treasury bond (e.g. Palia, 2001; Brick, Palia, and Wang, 2005). We believe that matching the term on the risk free
rate to the early exercise behavior of executives (typically seven years) is more appropriate (see Carpenter, 1998;
Huddart and Lang, 1997; and Bizjak, Bettis, and Lemmon, 2003).
2
When the number of new options granted exceeds the number of unexercisable options, such as in the case of
immediate vesting of some or all new grants, the excess realizable value and number of options is deducted from the
number and realizable value of exercisable options.
B-2

options, we take the difference between the fiscal year end stock price and the value of
exercisable in-the-money options to the number of unexercised exercisable in-the-money
options. These calculations yield average estimates for the exercise price of exercisable and
unexercisable options. We set the time-to-maturity of the unexercisable options equal to one
minus the time-to-maturity of the current years new option grants. The time-to-maturity of
exercisable options is set to three minus the time-to-maturity of the unexercisable options. When
no grant is made in the current year, the time-to-maturity of unexercisable and exercisable
options is set to nine and six years respectively, as in Core and Guay (2002).

C-1

Appendix C: Additional tables and robustness tests
Table C1: GICS industry group and Fama-French 49 industry classification breakdowns
Panel A: GICS Industry Group Classification
GICS Industry Group Description 91 92 93 94 95 96 97 98 99 00 01 02 03 04 05 06 07 08 Total
0 Missing 2 2 3 2 0 2 1 1 1 0 2 1 10 5 4 3 6 0 45
1010 Energy 31 33 33 31 30 30 31 29 26 26 57 51 130 133 153 177 189 194 1,384
1510 Materials 65 63 63 62 65 64 66 67 62 58 63 61 112 126 131 135 137 135 1,535
2010 Capital Goods 75 76 76 74 71 69 65 65 63 61 71 76 200 213 232 239 237 256 2,219
2020 Commercial and Professional Services 27 26 26 25 26 27 27 27 27 27 36 35 90 91 97 101 103 100 918
2030 Transportation 19 19 19 20 21 20 19 17 18 17 24 24 45 54 54 61 62 64 577
2510 Automobiles and Components 14 14 14 15 15 15 14 13 14 16 18 19 34 36 31 29 31 34 376
2520 Consumer Durables and Apparel 39 39 38 38 39 39 40 42 41 38 43 47 105 101 105 102 103 99 1,098
2530 Consumer Services 16 17 17 17 18 20 19 18 19 17 31 38 97 107 110 115 107 106 889
2540 Media 24 24 27 26 28 25 25 25 25 25 53 53 89 93 94 86 86 72 880
2550 Retailing 38 38 37 37 34 33 32 31 33 34 51 62 150 150 149 140 137 136 1,322
3010 Food and Staples Retailing 16 17 17 17 16 16 17 17 17 15 19 18 32 30 29 31 27 26 377
3020 Food, Beverage and Tobacco 25 26 26 26 26 26 26 26 28 27 31 35 61 61 61 60 63 64 698
3030 Household and Personal Products 13 13 12 11 10 8 10 10 10 10 11 12 23 24 25 28 25 25 280
3510 Health Care Equipment and Services 21 23 23 25 25 26 29 29 30 30 54 61 195 206 197 188 180 192 1,534
3520 Pharm., Biotechnology & Life Sciences 21 21 21 21 19 19 19 19 21 23 59 47 181 203 180 175 178 179 1,406
4010 Banks 34 30 30 31 33 30 31 34 32 33 66 83 348 278 293 272 235 256 2,149
4020 Diversified Financials 11 11 12 14 16 16 18 22 23 25 43 44 70 93 99 104 119 100 840
4030 Insurance 24 25 25 24 25 25 26 25 22 22 42 51 95 103 106 105 122 112 979
4040 Real Estate 3 3 3 3 3 3 2 2 2 2 27 40 141 157 167 147 140 143 988
4510 Software and Services 12 14 15 13 14 15 17 19 21 28 78 65 233 244 218 210 217 205 1,638
4520 Technology Hardware and Equipment 27 27 24 24 26 33 34 36 38 40 81 64 188 198 176 172 174 175 1,537
4530 Semiconductor Equipment 8 8 8 7 8 9 9 10 10 17 40 36 103 112 106 105 99 105 800
5010 Telecommunication Services 14 15 14 15 16 17 15 16 15 13 28 16 44 47 47 47 49 47 475
5510 Utilities 47 48 51 52 51 51 50 51 57 52 69 63 91 90 91 89 88 94 1,185
Total 626 632 634 630 635 638 642 651 655 656 1,097 1,102 2,867 2,955 2,955 2,921 2,914 2,919 26,129


C-2

Panel B: Fama-French 49 Industry Classification
Industry 91 92 93 94 95 96 97 98 99 00 01 02 03 04 05 06 07 08 Total
Agriculture 1 1 1 2 2 2 2 2 2 0 2 3 6 7 6 6 6 7 58
Food Products 18 19 19 18 18 18 18 18 19 18 19 22 40 37 40 39 40 42 462
Candy and Soda 1 1 1 1 1 1 1 1 2 2 3 3 5 6 6 7 7 6 55
Beer and Liquor 5 5 5 5 5 5 5 5 5 5 6 6 8 8 7 7 7 6 105
Tobacco Products 2 2 2 2 2 2 2 2 2 2 3 3 5 5 5 4 4 6 55
Recreation 5 5 5 5 5 4 4 4 4 4 4 4 13 12 13 13 12 11 127
Entertainment 4 4 4 3 3 3 3 3 4 2 13 14 36 39 37 39 37 34 282
Printing and Publishing 17 16 17 16 18 17 17 17 17 15 19 18 27 30 29 29 25 23 367
Consumer Goods 21 21 21 20 19 18 18 18 18 19 20 22 39 37 35 35 34 36 451
Apparel 10 10 10 10 10 11 11 11 11 11 14 14 33 32 34 35 36 34 337
Healthcare 4 4 5 5 4 4 5 4 4 4 13 16 39 47 49 43 39 45 334
Medical Equipment 8 9 9 10 11 12 13 13 12 12 16 16 81 87 80 76 77 81 623
Pharmaceutical Products 19 20 20 19 18 18 19 19 22 23 51 39 166 189 162 160 162 161 1,287
Chemicals 26 25 26 25 24 24 24 23 20 20 25 25 44 48 54 62 62 57 614
Rubber and Plastic Products 3 3 3 3 3 4 4 4 3 4 3 4 14 15 13 15 14 14 126
Textiles 3 3 3 3 4 4 4 5 5 4 3 3 6 4 5 4 4 5 72
Construction Materials 14 14 13 13 13 13 15 15 14 13 15 16 37 38 40 38 37 35 393
Construction 7 8 7 8 8 6 6 6 6 5 9 11 32 34 34 36 37 36 296
Steel Works Etc 13 13 13 12 12 12 13 14 13 11 11 10 27 32 36 34 36 31 343
Fabricated Products 0 0 0 0 0 0 0 0 0 0 1 1 8 6 5 7 5 6 39
Machinery 29 28 29 28 27 26 24 24 24 25 33 36 79 84 83 84 84 90 837
Electrical Equipment 9 9 9 9 10 10 9 9 8 8 10 9 30 36 36 34 32 40 317
Automobiles and Trucks 17 17 17 18 18 18 17 16 17 16 17 20 39 42 39 36 37 38 439
Aircraft 7 7 7 6 6 6 5 5 5 5 7 7 13 13 13 16 15 16 159
Shipbuilding, Railroad Equip. 4 4 4 3 3 3 3 3 3 3 4 3 6 6 7 8 9 9 85
Defense 1 2 2 2 1 1 1 1 1 1 1 2 5 6 6 6 7 7 53
Precious Metals 5 5 5 6 7 7 6 6 5 5 3 1 3 3 3 3 4 3 80
Non-Met. and Ind. Metal Mining 2 1 1 1 2 2 3 3 4 4 5 5 9 9 9 9 11 13 93
Coal 1 1 1 0 0 0 0 0 0 0 4 2 6 5 8 11 10 12 61
Petroleum and Natural Gas 28 30 29 28 27 27 28 26 23 22 40 38 90 93 107 120 123 134 1,013
Utilities 45 46 50 51 51 52 51 52 57 53 70 66 94 94 96 94 92 96 1,210
Communication 21 23 23 24 25 25 23 24 22 19 51 38 84 90 88 86 97 88 851
Personal Services 7 8 8 8 8 8 7 7 7 7 9 12 28 31 33 37 38 40 303
Business Services 8 7 9 9 10 11 12 13 13 16 41 44 140 143 151 154 145 143 1,069
Computer Hardware 13 13 13 13 14 17 17 18 18 19 34 22 58 60 51 48 58 57 543
Computer Software 14 16 14 12 13 15 17 19 20 25 68 51 212 221 195 182 191 183 1,468
Electronic Equipment 22 22 22 21 21 23 23 25 25 32 71 63 166 176 167 170 154 166 1,369
Measuring and Control Equip. 7 7 6 6 6 6 6 7 7 10 17 17 48 55 55 51 45 51 407
Business Supplies 27 27 26 26 26 23 24 25 24 22 22 21 32 33 33 32 31 33 487
Shipping Containers 3 3 2 3 3 3 4 4 4 4 4 6 10 11 10 9 9 10 102
Transport 19 19 19 20 21 20 19 17 19 18 28 28 56 64 66 71 83 79 666
Wholesale 17 17 16 16 15 14 15 15 18 18 23 24 69 74 78 78 78 73 658
Retail 46 47 47 46 43 43 44 43 45 44 68 76 169 168 163 159 152 150 1,553
Restaurants, Hotels, Motels 10 10 10 10 10 12 11 11 11 11 15 19 38 40 44 46 43 41 392
Banking 41 37 38 42 45 41 42 47 46 47 81 99 375 315 330 309 275 291 2,501
Insurance 26 27 27 27 29 31 33 32 29 29 57 66 121 125 125 125 143 133 1,185
Real Estate 0 0 0 0 0 0 0 0 0 0 1 2 8 11 11 11 16 16 76
Trading 8 8 8 8 9 11 10 10 11 13 53 65 186 209 231 216 215 202 1,473
C-3

Almost Nothing 9 9 9 8 6 6 5 5 6 6 10 10 28 26 28 29 36 29 265
Total 627 633 635 631 636 639 643 651 655 656 1,097 1,102 2,868 2,956 2,956 2,923 2,914 2,919 26,141
C-4

Table C2: RE (GLS) results from Table 2
Dependent Variable Pred. DESM DESM IASM IASM Pred. SM SM
Independent Variable Sign (1) (2) (3) (4) Sign (5) (6)
LNDELTA (+/) 0.00566* 0.00553* 0.0119*** 0.0121*** 0.0169* 0.0179*
(1.78) (1.73) (3.12) (3.12) (1.74) (1.82)
BRDSIZE (+) 0.0136*** 0.0161*** 0.0000925
(4.76) (5.24) (0.01)
PIND () 0.000613 0.000822* 0.000466
(1.58) (1.93) (0.46)
BDIND () 0.184** 0.314*** 0.187
(2.05) (2.99) (0.65)
DUALITY (+/) 0.0148 0.0161* 0.0270** 0.0286** 0.0449* 0.0416
(1.53) (1.65) (2.41) (2.50) (1.72) (1.59)
INSTOWN () 0.000197 0.000305 0.0000482 0.0000933 0.00184* 0.00169*
(0.56) (0.86) (0.12) (0.22) (1.88) (1.72)
DELTA x OVER (+/) 0.00800 0.00968
(0.60) (0.72)
BRDSIZE x OVER (+) 0.0293***
(2.94)
PIND x OVER () 0.00212
(1.43)
BRDIND x OVER 0.767**
(2.00)
DUALITY x OVER (+/) 0.118*** 0.116***
(2.85) (2.75)
INSTOWN x OVER () 0.00286** 0.00239*
(2.03) (1.69)
OVER 0.298* 0.000897
(1.66) (0.01)
CASH/TA 0.00656 0.00353 0.00156 0.00945 0.00308 0.00970
(1.02) (0.58) (0.19) (1.18) (0.38) (1.23)
TD/TA 0.000481 0.000530 0.000341 0.000282
(1.23) (1.33) (0.78) (0.64)
LNSALES 0.0400 0.0429 0.0410 0.0474
(0.82) (0.87) (0.80) (0.92)
ROA 0.00146* 0.00160** 0.000677 0.000534 0.000904 0.000790
(1.91) (2.07) (0.80) (0.62) (1.12) (0.96)
TRS
0.00000982 0.00000786 0.000207*** 0.000206*** 0.000119 0.000110
(0.13) (0.11) (2.74) (2.73) (1.60) (1.48)

1
+
6
= 0 9.300*** 9.010**

2
+
7
= 0 39.31*** 18.05***

3
+
8
= 0 6.786**

4
+
9
= 0 12.05*** 11.69***

5
+
10
= 0 4.134 3.057

Industry Yes Yes Yes Yes Yes Yes
Year Yes Yes Yes Yes Yes Yes

n 1631 1631 1631 1631 1631 1631
N 9051 9051 9051 9051 9051 9051
Within R
2
0.0257 0.0231 0.0400 0.0361 0.0585 0.0532
Between R
2
0.0147 0.00630 0.0208 0.0143 0.127 0.119
Overall R
2
0.0139 0.00533 0.0217 0.0138 0.136 0.126


C-5

Table C3: FE balanced panel results from Table 2
Dependent Variable Pred. DESM DESM IASM IASM Pred. SM SM
Independent Variable Sign (1) (2) (3) (4) Sign (5) (6)
LNDELTA (+/) 0.0127 0.0133 0.0274* 0.0293** 0.0711** 0.0752**
(1.23) (1.27) (1.95) (2.09) (2.44) (2.51)
BRDSIZE (+) 0.0128* 0.0159** 0.0185
(1.73) (2.11) (1.21)
PIND () 0.000787 0.00193 0.000994
(0.66) (1.60) (0.35)
BDIND () 0.356 0.976*** 0.262
(1.29) (3.04) (0.28)
DUALITY (+/) 0.00583 0.00408 0.00820 0.00734 0.0224 0.0176
(0.24) (0.17) (0.28) (0.25) (0.33) (0.26)
INSTOWN () 0.000692 0.000738 0.00204 0.00218 0.00679** 0.00728**
(0.57) (0.61) (1.46) (1.57) (2.01) (2.20)
DELTA x OVER (+/) 0.0689* 0.0729*
(1.77) (1.82)
BRDSIZE x OVER (+) 0.000372
(0.01)
PIND x OVER () 0.00504
(1.11)
BRDIND x OVER 0.953
(0.70)
DUALITY x OVER (+/) 0.0125 0.00917
(0.10) (0.07)
INSTOWN x OVER () 0.00876 0.00948*
(1.58) (1.76)
OVER 0.243 0.381
(0.43) (0.86)
CASH/TA 0.0484 0.0429 0.0995 0.0971 0.103* 0.0990*
(0.99) (0.89) (1.54) (1.55) (1.76) (1.73)
TD/TA 0.00262** 0.00263** 0.00270* 0.00273*
(2.26) (2.24) (1.91) (1.95)
LNSALES 0.0749 0.0746 0.0819 0.104
(0.35) (0.34) (0.38) (0.47)
ROA 0.00103 0.00117 0.000882 0.000651 0.00151 0.00128
(0.31) (0.35) (0.26) (0.19) (0.52) (0.43)
TRS
0.000399* 0.000404** 0.000271 0.000272 0.0000725 0.0000856
(1.96) (1.99) (1.11) (1.13) (0.32) (0.38)

1
+
6
= 0 3.463** 3.735**

2
+
7
= 0 3.379** 5.255***

3
+
8
= 0 1.965

4
+
9
= 0 0.193 0.120

5
+
10
= 0 2.246 2.605*

Year Yes Yes Yes Yes Yes Yes

n 153 153 153 153 153 153
N 1989 1989 1989 1989 1989 1989
Within R
2
0.0472 0.0452 0.0609 0.0658 0.0905 0.0902
Between R
2
0.00582 0.00329 0.00929 0.0128 0.000203 0.000964
Overall R
2
0.0207 0.0174 0.0216 0.0260 0.0129 0.0151

C-6

Table C4: First-stage regression results from Table 4
Consumer Industrial Consumer Industrial Consumer Industrial Consumer Industrial
DESM DESM DESMSTR DESMSTR DESMCON DESMCON DESIP DESIP
(1) (2) (3) (4) (5) (6) (7) (8)
Eq. = Main
LNDELTA 0.00388 0.00861 0.000633 0.00593 0.00523 0.00252 0.000783 0.00680
(0.33) (0.35) (0.09) (0.35) (0.85) (0.19) (0.05) (0.31)
BRDIND 4.838*** 5.010 2.793*** 9.853*** 0.166 0.437 6.620*** 2.848
(4.79) (1.19) (3.66) (3.76) (0.25) (0.15) (11.25) (0.66)
DUALITY 0.0186 0.0233 0.00273 0.0115 0.0216*** 0.0135** 0.00587 0.0256
(1.31) (1.01) (0.39) (0.77) (3.50) (2.11) (0.30) (1.32)
INSTOWN 0.000801 0.00141 0.00557*** 0.00251 0.00478*** 0.00476*** 0.00136 0.00819***
(0.33) (0.48) (3.12) (1.04) (3.76) (3.67) (0.98) (3.54)
CASH/TA 0.0830 0.374* 0.0117 0.180 0.0686* 0.345*** 0.121 0.292
(1.41) (1.92) (0.31) (1.38) (1.71) (2.78) (1.61) (1.52)
TD/TA 0.000589 0.000629 0.000679** 0.000319 0.0000377 0.000371 0.00131* 0.000107
(1.11) (0.49) (2.01) (0.36) (0.10) (0.46) (1.88) (0.09)
LNSALES 0.0677*** 0.0173 0.0776*** 0.0685** 0.0255 0.0892*** 0.0867*** 0.0307
(2.76) (0.42) (4.92) (2.33) (1.52) (3.49) (2.76) (0.76)
ROA 0.00000295 0.00521*** 0.00210*** 0.000620 0.00219*** 0.00473*** 0.000384 0.00568***
(0.00) (2.59) (3.23) (0.43) (3.30) (3.83) (0.32) (2.93)
TRS 0.0000575 0.000129 0.000180*** 0.000274 0.0000642 0.00000211 0.0000782 0.0000762
(0.70) (0.52) (3.32) (1.54) (1.26) (0.01) (0.78) (0.32)
CONSTANT 0.0232** 0.0215 0.00838 0.0482** 0.0169*** 0.000640 0.0338*** 0.00248
(2.54) (0.79) (1.29) (2.56) (3.10) (0.03) (4.22) (0.09)
Eq. = LNDELTA
DESM 3.190 5.342
(1.09) (1.53)
DESMSTR 30.89*** 17.50**
(4.56) (2.32)
DESMCON 4.537 22.95***
(0.51) (5.47)
DESIP 1.516 6.000*
(0.74) (1.89)
CASH/TA 0.385 0.877 1.690*** 2.368* 0.574 7.373*** 0.393 1.149*
(1.34) (1.04) (2.73) (1.67) (0.59) (4.46) (1.47) (1.72)
TD/TA 0.000566 0.00861 0.00994* 0.0282** 0.00181 0.0128 0.000327 0.00205
(0.19) (1.61) (1.72) (2.18) (0.26) (1.63) (0.10) (0.56)
LNSALES 0.0260 0.172 2.012*** 1.319** 0.0429 2.034*** 0.0227 0.124
(0.15) (1.13) (3.70) (1.97) (0.10) (4.26) (0.15) (0.73)
ROA 0.00609 0.0287* 0.0549*** 0.00504 0.0125 0.108*** 0.00685 0.0288
(1.29) (1.74) (3.63) (0.34) (0.64) (4.37) (1.51) (1.46)
TRS 0.00345*** 0.00272*** 0.00933*** 0.00545*** 0.00281*** 0.00404*** 0.00320*** 0.00197***
(4.71) (3.40) (5.46) (2.64) (2.97) (3.67) (4.60) (3.69)
C-7

Q 0.0313 0.171 0.394*** 0.129 0.105* 0.315*** 0.0460 0.394***
(0.61) (1.30) (3.04) (0.46) (1.79) (4.15) (0.86) (5.03)
RD/TA 0.00000281 0.000770** 0.000711** 0.000393 0.000357 0.0000426 0.00000700 0.0000558
(0.02) (2.33) (2.03) (0.87) (1.13) (0.15) (0.05) (0.13)
VOLATILITY 0.213 0.101 0.614 0.734 0.702 1.306*** 0.195 0.728**
(0.86) (0.33) (1.62) (1.22) (1.23) (3.41) (0.79) (2.27)
CAP 0.00772 0.0222*** 0.00159 0.0369*** 0.000820 0.0115 0.00918* 0.0130**
(1.14) (2.71) (0.23) (2.77) (0.06) (1.30) (1.73) (2.00)
LNTCC 0.288*** 0.198*** 0.256*** 0.191** 0.396*** 0.0554 0.279*** 0.0446
(7.86) (2.70) (4.34) (2.38) (4.22) (1.02) (7.75) (0.83)
TENURE 0.0882*** 0.0767*** 0.0764*** 0.0520*** 0.0783*** 0.0582*** 0.0900*** 0.0836***
(15.40) (13.09) (8.91) (3.51) (7.03) (5.73) (16.90) (13.65)
CONSTANT 0.0282 0.0285 0.363*** 0.153 0.0361 0.139** 0.0392 0.0573**
(0.94) (0.98) (3.55) (1.49) (0.45) (2.19) (1.55) (2.31)
Eq. = BRDIND
DESM 0.154*** 0.108***
(3.31) (3.10)
DESMSTR 0.125 0.0644
(0.92) (1.00)
DESMCON 0.213*** 0.295***
(2.86) (2.72)
DESIP 0.134*** 0.166***
(5.65) (3.56)
LNDELTA 0.000278 0.0000351 0.000728 0.000261 0.000522 0.00362 0.0000289 0.000338
(0.13) (0.01) (0.39) (0.11) (0.25) (1.05) (0.01) (0.12)
DUALITY 0.0000916 0.00172 0.00196 0.000618 0.00615** 0.000483 0.000251 0.000887
(0.04) (0.71) (0.77) (0.26) (2.57) (0.20) (0.09) (0.36)
CASH/TA 0.0159* 0.0533*** 0.00938 0.0226 0.0271*** 0.122*** 0.0172 0.0585***
(1.69) (3.14) (0.94) (1.41) (2.64) (3.14) (1.52) (3.13)
TD/TA 0.000101 0.0000155 0.0000822 0.0000103 0.0000535 0.000241* 0.000183* 0.0000742
(1.12) (0.13) (0.91) (0.07) (0.62) (1.75) (1.68) (0.61)
LNSALES 0.0118*** 0.00243 0.0144 0.00390 0.000716 0.0275** 0.0118** 0.00625
(2.67) (0.60) (1.38) (0.61) (0.16) (2.49) (2.42) (1.32)
ROA 0.0000399 0.000604** 0.000247 0.0000910 0.000268 0.00155*** 0.0000888 0.00106***
(0.26) (2.46) (0.88) (0.47) (1.26) (2.59) (0.49) (2.98)
TRS 0.00000665 0.0000294* 0.00000131 0.0000310 0.0000239* 0.0000708*** 0.0000115 0.0000377**
(0.41) (1.71) (0.04) (1.40) (1.75) (2.86) (0.71) (2.09)
Q 0.000896 0.000221 0.00156 0.00212 0.00312*** 0.000216 0.000309 0.000963
(0.94) (0.14) (0.62) (0.90) (3.35) (0.12) (0.60) (0.84)
RD/TA 0.00000105 0.00000250 0.00000369 0.000000905 0.000000396 0.00000618 0.000000794 0.00000908
(0.43) (0.65) (0.53) (0.35) (0.10) (0.91) (0.73) (1.36)
VOLATILITY 0.000185 0.00193 0.000567 0.000700 0.000243 0.0101 0.00388** 0.00837
(0.04) (0.34) (0.08) (0.17) (0.03) (1.10) (2.25) (1.24)
AGE 0.0000656 0.000229* 0.000127 0.000106 0.0000262 0.000240 0.0000358 0.0000624
(0.58) (1.96) (1.17) (1.01) (0.17) (1.02) (0.87) (0.41)
C-8

CONSTANT 0.00481*** 0.00524*** 0.00438** 0.00487*** 0.00710*** 0.00782*** 0.00473*** 0.00526***
(5.97) (6.70) (2.32) (4.35) (8.06) (6.41) (5.27) (6.25)
Eq. = INSTOWN
DESM 9.006 14.38
(0.86) (1.16)
DESMSTR 40.19*** 3.060
(3.97) (0.22)
DESMCON 61.04*** 37.41*
(5.39) (1.85)
DESIP 7.234 33.71***
(0.98) (2.65)
CASH/TA 2.496 5.994 4.184* 2.984 5.990** 14.64* 2.427 9.059*
(1.24) (1.19) (1.92) (0.71) (2.54) (1.87) (1.21) (1.69)
TD/TA 0.0242 0.0236 0.0399* 0.000968 0.0162 0.00749 0.0284 0.0188
(1.24) (0.69) (1.94) (0.03) (0.76) (0.22) (1.38) (0.53)
LNSALES 0.0585 0.534 2.144* 0.474 1.312 2.702 0.0667 0.325
(0.06) (0.48) (1.93) (0.31) (1.30) (1.25) (0.07) (0.25)
ROA 0.133*** 0.0191 0.170*** 0.0566 0.238*** 0.132 0.135*** 0.150
(3.95) (0.23) (4.85) (1.12) (5.89) (1.11) (4.12) (1.49)
TRS 0.0173*** 0.0265*** 0.0204*** 0.0259*** 0.0166*** 0.0224*** 0.0173*** 0.0235***
(6.73) (6.77) (7.54) (6.53) (6.26) (4.82) (6.88) (5.11)
TURNOVER 2.589*** 3.559*** 2.263*** 3.259*** 1.584*** 2.989*** 2.611*** 2.961***
(5.13) (4.65) (4.65) (4.54) (3.50) (4.10) (5.20) (3.68)
CONSTANT 2.091*** 1.870*** 2.567*** 1.811*** 2.431*** 1.556*** 2.085*** 1.916***
(12.04) (8.59) (12.22) (7.05) (12.88) (5.29) (12.46) (7.68)
N 4597 2429 4597 2429 4597 2429 4597 2429

C-9

Table C5: FirstStage Regression Results for Table 5
Consumer Industrial Consumer Industrial Consumer Industrial Consumer Industrial Consumer Industrial
DEEMP DEEMP DEDIV DEDIV DEPRO DEPRO DECOM DECOM DEENV DEENV
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10)
Eq. = Main
LNDELTA 0.00420 0.0194 0.00247 0.0167 0.00159 0.0140** 0.00286 0.00277 0.000554 0.00152
(0.84) (1.49) (0.46) (1.26) (0.75) (1.97) (1.07) (0.45) (0.36) (0.23)
BRDIND 2.480*** 3.719*** 0.742 7.592*** 1.061*** 3.404*** 1.073*** 1.455 0.211 2.685
(5.47) (6.00) (1.17) (3.28) (4.61) (8.26) (4.99) (1.00) (1.41) (1.51)
DUALITY 0.0198*** 0.0109 0.000415 0.00917 0.00971*** 0.00583 0.00734** 0.000604 0.00201 0.00557*
(3.70) (0.82) (0.06) (0.82) (3.87) (0.79) (2.29) (0.11) (1.24) (1.87)
INSTOWN 0.00167 0.000163 0.00687*** 0.00192 0.00147*** 0.000941*** 0.00169*** 0.000792 0.000813*** 0.00152
(1.54) (0.67) (4.43) (1.54) (3.18) (3.31) (2.74) (0.79) (2.64) (1.53)
CASH 0.0630** 0.189** 0.0114 0.113 0.0462** 0.00986 0.0163 0.0852 0.0223* 0.185***
(2.33) (2.14) (0.36) (1.07) (2.31) (0.17) (1.10) (1.61) (1.86) (2.92)
TD/TA 0.000352 0.0000411 0.00000820 0.000803 0.000126 0.000683* 0.000263* 0.000234 0.000140 0.0000311
(1.44) (0.06) (0.03) (1.17) (0.68) (1.65) (1.95) (0.75) (1.27) (0.08)
LNSALES 0.00595 0.0165 0.0523*** 0.0452** 0.00779 0.0226 0.00906 0.00852 0.0121** 0.0220**
(0.53) (0.72) (4.01) (2.03) (0.93) (1.52) (1.46) (0.94) (2.42) (2.08)
ROA 0.000881* 0.00178 0.000194 0.000119 0.000281 0.00128* 0.000752*** 0.00125*** 0.00000656 0.00183***
(1.94) (1.62) (0.37) (0.11) (0.84) (1.80) (3.01) (2.82) (0.03) (3.54)
TRS 0.0000852** 0.000236** 0.0000742* 0.0000935 0.0000481* 0.000165*** 0.000000686 0.0000999 0.0000282* 0.000146*
(2.34) (2.46) (1.67) (0.71) (1.93) (2.75) (0.03) (1.45) (1.87) (1.78)
CONSTANT 0.0113*** 0.0238*** 0.0115* 0.0312** 0.00429** 0.0198*** 0.00155 0.00829 0.000562 0.0184*
(2.76) (4.20) (1.95) (2.17) (2.23) (5.53) (0.70) (0.97) (0.40) (1.76)
Eq. = LNDELTA
DEEMP 28.34* 6.026
(1.83) (0.61)
DEDIV 0.382 13.68***
(0.08) (2.93)
DEPRO 21.03* 9.309
(1.79) (0.95)
DECOM 30.66*** 4.391
(4.70) (0.68)
DEENV 107.4*** 34.46***
(3.65) (3.05)
CASH/TA 1.330 0.189 0.429 1.403* 1.099** 1.078 0.348 0.00655 2.761** 3.348**
(1.55) (0.25) (1.54) (1.74) (1.98) (1.06) (0.93) (0.01) (2.46) (2.17)
TD/TA 0.00830 0.00580 0.00189 0.00261 0.00236 0.00155 0.00313 0.00279 0.0163* 0.0135
(1.17) (1.03) (0.76) (0.59) (0.54) (0.39) (0.86) (0.89) (1.68) (1.64)
LNSALES 0.155 0.186 0.0740 0.777*** 0.221 0.160 0.00177 0.115 1.260** 0.740**
(0.55) (0.88) (0.28) (2.72) (0.85) (0.58) (0.01) (0.98) (2.39) (2.09)
ROA 0.0395* 0.0149 0.00639 0.0174* 0.00542 0.00747 0.0231*** 0.0100 0.0159 0.0586***
(1.77) (0.91) (0.97) (1.90) (0.78) (0.46) (3.20) (1.22) (1.09) (2.69)
TRS 0.000701 0.00164** 0.00277*** 0.000730 0.00169*** 0.00181*** 0.00339*** 0.00196*** 0.000283 0.00116
(0.37) (2.05) (5.35) (0.60) (2.65) (3.52) (6.55) (4.11) (0.20) (1.10)
Q 0.343*** 0.314*** 0.0796** 0.501*** 0.156*** 0.348*** 0.0551 0.284*** 0.197*** 0.220***
(2.61) (4.18) (2.49) (4.34) (4.64) (4.82) (1.23) (3.41) (2.85) (2.60)
RD/TA 0.000133 0.000779 0.0000116 0.000599** 0.000141 0.000468* 0.0000104 0.000537** 0.0000845 0.000615**
(0.70) (1.44) (0.06) (2.34) (0.72) (1.86) (0.08) (2.57) (0.71) (2.14)
C-10

VOLATILITY 0.507 0.152 0.193 0.0919 0.395 0.687* 0.0649 0.134 0.380 0.336
(1.43) (0.53) (0.70) (0.26) (1.42) (1.92) (0.27) (0.52) (0.91) (0.74)
CAP 0.0107 0.0221** 0.00823 0.0100 0.00479 0.0223*** 0.00776 0.0179*** 0.0133* 0.0277**
(0.90) (2.05) (1.12) (1.61) (0.80) (3.50) (1.29) (3.27) (1.67) (2.52)
LNTCC 0.179*** 0.165* 0.261*** 0.0684 0.209*** 0.131** 0.322*** 0.117*** 0.213*** 0.130***
(2.90) (1.65) (7.18) (0.76) (4.61) (1.98) (8.07) (2.88) (3.30) (2.82)
TENURE 0.103*** 0.0815*** 0.0928*** 0.0793*** 0.0969*** 0.0798*** 0.0855*** 0.0791*** 0.0679*** 0.0500***
(10.16) (10.64) (22.44) (11.81) (15.58) (15.90) (15.00) (16.12) (4.11) (4.31)
CONSTANT 0.0453 0.0704 0.0486 0.133*** 0.0258 0.0482** 0.0123 0.0405* 0.0566 0.0590
(0.91) (1.53) (1.19) (3.12) (0.75) (2.11) (0.40) (1.76) (0.78) (1.12)
Eq. = BRDIND
DEEMP 0.428*** 0.236***
(2.69) (8.12)
DEDIV 0.0844 0.150***
(0.79) (3.46)
DEPRO 0.391 0.183***
(1.03) (3.15)
DECOM 1.330*** 0.236
(3.83) (1.38)
DEENV 2.727*** 0.0398
(3.83) (0.18)
LNDELTA 0.00113 0.00463* 0.0000753 0.00323 0.00244 0.00279 0.000330 0.0000565 0.00625 0.00171
(0.54) (1.68) (0.03) (1.20) (0.85) (1.13) (0.14) (0.02) (1.38) (0.49)
DUALITY 0.00679*** 0.00204 0.00263 0.000123 0.0101*** 0.00155 0.0106** 0.000804 0.00506 0.000538
(2.69) (0.74) (1.19) (0.05) (2.94) (0.61) (2.46) (0.35) (1.30) (0.22)
CASH/TA 0.0280*** 0.0482*** 0.0116 0.0118 0.0275 0.0130 0.0179* 0.0397** 0.0450* 0.0336
(2.64) (2.68) (1.30) (0.78) (1.17) (0.70) (1.68) (2.42) (1.91) (1.14)
TD/TA 0.000129 0.0000206 0.0000468 0.000114 0.0000637 0.000172 0.000259** 0.000124 0.000496** 0.000100
(1.36) (0.16) (0.59) (1.07) (0.33) (1.32) (2.27) (1.16) (2.46) (0.77)
LNSALES 0.00288 0.00404 0.00945 0.00695 0.00236 0.00374 0.0113** 0.00261 0.0398*** 0.00177
(0.71) (0.85) (1.47) (1.56) (0.23) (0.76) (2.41) (0.66) (3.51) (0.29)
ROA 0.000473* 0.000429* 0.000132 0.0000174 0.000205 0.000228 0.000758*** 0.000335 0.000319 0.000140
(1.84) (1.88) (0.78) (0.09) (0.61) (0.95) (2.88) (1.39) (1.13) (0.35)
TRS 0.0000261 0.0000596*** 0.0000204 0.0000170 0.0000182 0.0000376** 0.00000428 0.0000408** 0.0000749*** 0.0000313*
(1.15) (3.08) (1.52) (0.92) (0.71) (2.01) (0.26) (2.38) (2.82) (1.76)
Q 0.000432 0.000263 0.00310*** 0.000631 0.000603 0.0000482 0.00263 0.00109 0.00710*** 0.00139
(0.31) (1.01) (3.22) (0.44) (0.61) (0.16) (1.47) (0.42) (3.46) (0.60)
RD/TA 0.00000108 0.00000151 0.00000207 0.000000764 0.00000518 0.00000205 0.00000428 0.00000134 8.78e08 0.000000460
(0.43) (1.18) (0.44) (0.20) (1.14) (1.62) (1.45) (0.21) (0.02) (0.06)
VOLATILITY 0.000775 0.00200 0.00190 0.00265 0.0109* 0.00422* 0.00751 0.00256 0.0150 0.0000725
(0.15) (1.62) (0.26) (0.48) (1.70) (1.94) (1.16) (0.31) (1.18) (0.01)
AGE 0.0000760 0.0000193 0.000165 0.0000680 0.000220 0.000128** 0.000490** 0.000205 0.0000789 0.000386
(0.86) (0.45) (0.91) (0.43) (1.09) (2.41) (2.20) (0.87) (0.41) (1.60)
CONSTANT 0.00607*** 0.00623*** 0.00525*** 0.00445*** 0.00665*** 0.00532*** 0.00452*** 0.00528*** 0.00378*** 0.00566***
(8.17) (6.91) (5.07) (5.65) (4.23) (5.91) (4.94) (6.65) (2.72) (7.65)
Eq. = INSTOWN
DEEMP 38.61 1.007
(1.54) (0.04)
DEDIV 52.94*** 2.484
(3.16) (0.15)
C-11

DEPRO 195.4*** 135.1***
(4.59) (2.63)
DECOM 90.41*** 29.29
(2.87) (0.57)
DEENV 290.0*** 20.55
(6.51) (0.45)
CASH/TA 1.293 2.793 0.390 2.454 4.374 13.88* 2.313 4.052 3.791 0.372
(0.56) (0.62) (0.18) (0.56) (1.17) (1.92) (1.06) (0.87) (1.40) (0.06)
TD/TA 0.0306 0.00620 0.0127 0.00626 0.0125 0.0369 0.0377* 0.00417 0.0248 0.000616
(1.43) (0.20) (0.64) (0.21) (0.38) (0.83) (1.77) (0.14) (0.99) (0.02)
LNSALES 0.724 0.706 2.763** 0.834 3.104** 2.629 0.110 0.545 4.236*** 1.084
(0.78) (0.63) (2.40) (0.61) (1.99) (1.32) (0.12) (0.49) (3.49) (0.80)
ROA 0.0764 0.0543 0.108*** 0.0573 0.110** 0.143 0.168*** 0.0210 0.0967** 0.0904
(1.39) (0.86) (3.03) (1.13) (1.96) (1.33) (4.64) (0.26) (2.33) (0.99)
TRS 0.0184*** 0.0262*** 0.0139*** 0.0265*** 0.0204*** 0.0254*** 0.0151*** 0.0255*** 0.0202*** 0.0269***
(6.44) (6.56) (5.40) (6.10) (4.94) (4.64) (5.85) (6.36) (6.63) (6.65)
TURNOVER 2.468*** 3.366*** 1.680*** 3.285*** 1.613*** 2.739*** 2.384*** 3.663*** 1.434*** 3.109***
(4.88) (4.82) (3.52) (4.69) (3.37) (3.15) (4.83) (4.11) (3.25) (3.84)
CONSTANT 2.017*** 1.835*** 1.660*** 1.861*** 1.839*** 1.801*** 2.144*** 1.856*** 1.779*** 1.861***
(11.81) (8.28) (8.07) (7.99) (6.74) (5.94) (12.33) (8.76) (8.68) (8.75)
N 4597 2429 4597 2429 4597 2429 4597 2429 4597 2429