Content uploaded by William S. Schulze
Author content
All content in this area was uploaded by William S. Schulze on Feb 19, 2020
Content may be subject to copyright.
HAL Id: hal-02276698
https://hal.archives-ouvertes.fr/hal-02276698
Submitted on 3 Sep 2019
HAL is a multi-disciplinary open access
archive for the deposit and dissemination of sci-
entic research documents, whether they are pub-
lished or not. The documents may come from
teaching and research institutions in France or
abroad, or from public or private research centers.
L’archive ouverte pluridisciplinaire HAL, est
destinée au dépôt et à la diusion de documents
scientiques de niveau recherche, publiés ou non,
émanant des établissements d’enseignement et de
recherche français ou étrangers, des laboratoires
publics ou privés.
Exploring the agency consequences of ownership
dispersion among the directors of private family rms
Michael Lubatkin, William S. Schulze, Richard N. Dino
To cite this version:
Michael Lubatkin, William S. Schulze, Richard N. Dino. Exploring the agency consequences of own-
ership dispersion among the directors of private family rms. Academy of Management Journal, 2003,
pp.179-194 P. �hal-02276698�
Exploring the Agency Consequences of Ownership Dispersion among the Directors of Private
Family Firms
Author(s): William S. Schulze, Michael H. Lubatkin, Richard N. Dino
Source:
The Academy of Management Journal,
Vol. 46, No. 2 (Apr., 2003), pp. 179-194
Published by: Academy of Management
Stable URL: http://www.jstor.org/stable/30040613
Accessed: 10/02/2009 10:45
Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at
http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless
you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you
may use content in the JSTOR archive only for your personal, non-commercial use.
Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at
http://www.jstor.org/action/showPublisher?publisherCode=aom.
Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed
page of such transmission.
JSTOR is a not-for-profit organization founded in 1995 to build trusted digital archives for scholarship. We work with the
scholarly community to preserve their work and the materials they rely upon, and to build a common research platform that
promotes the discovery and use of these resources. For more information about JSTOR, please contact support@jstor.org.
Academy of Management is collaborating with JSTOR to digitize, preserve and extend access to The Academy
of Management Journal.
http://www.jstor.org
c Academy of Management
Journal
2003, Vol. 46, No. 2, 179-194.
EXPLORING
THE AGENCY
CONSEQUENCES
OF
OWNERSHIP
DISPERSION AMONG THE DIRECTORS
OF PRIVATE FAMILY
FIRMS
WILLIAM S. SCHULZE
Case Western Reserve University
MICHAEL H. LUBATKIN
University of Connecticut and Ecole de Management de Lyon
RICHARD N. DINO
University of Connecticut
Using an agency-theoretic
lens and insights drawn from the behavioral economics and
family business literatures, we developed hypotheses concerning the effect of disper-
sion of ownership on the use of debt by private family-owned and family-managed
firms. A field study of 1,464 family firms was conducted. Results suggest that, during
periods of market growth, the relationship between the use of debt and the dispersion
of ownership among directors at family firms can be graphed as a U-shaped curve.
The principal-agent model has had a profound
influence on corporate governance theory (Jensen,
1998). A central premise of this theory is that man-
agement decisions are strongly influenced by the
ownership status of each decision maker who
serves on a corporation's board of directors. The
agency positions of outside owners and owner-
managers differ. Outside owners prefer growth-
oriented risk taking because they benefit solely
from the appreciation of shareholder value. They
are also indifferent to the level of risk that is spe-
cific to any particular
investment made by a given
firm because they can reduce that risk by holding
diversified portfolios. Owners who manage a pri-
vate firm, in contrast, define its value in terms of
utility, and so they will undertake risks that are
commensurate with their preferences for certain
outcomes. These outcomes not only include finan-
cial and nonfinancial benefits, but also include the
utility generated by the ability to exercise author-
ity, dictate strategy,
and choose which investments
the firm will undertake.
Should an owner-manager relinquish equity to
outside owners, the agency theory prediction is
that changes in the incentives facing the owner-
manager will cause the firm's value to decline.
Specifically, because inside owners would now
bear only a fraction of the cost of the benefits they
receive, they have incentive to act opportunisti-
cally and make decisions that promote their self-
regarding interests as opposed to the interests of
outside shareholders (Demsetz, 1973; Jensen &
Meckling, 1976; Fama & Jensen, 1983a). In this
way, fractional ownership creates agency prob-
lems: it gives inside owners incentive to free ride
on outside owners' equity and to favor consump-
tion over investment.
But what if there are no outside shareholders and
firm equity is instead distributed among family
members? Will fractional ownership create agency
problems, as the conventional agency model
implies, or do family relationships promote the
within-group alignment of ownership interests and
encourage investment? This and other questions
about the governance of private family firms has
been largely glossed over in the management liter-
ature, yet family firms account for 40 to 60 percent
of U.S. gross national product and employ upwards
of 80 percent of the workforce (Gomez-Mejia,
Nufiez-Nickel, &
Gutierrez,
2001). Answers to these
questions can enrich corporate governance theory,
which heretofore has focused primarily on public
firms and the challenge of aligning insider goals
with those of outside investors (Morck,
Shleifer, &
Vishny, 1988; Wright, Ferris, Sarin, & Awasthi,
We would
like to thank
Jack
Veiga
and the anonymous
reviewers for their
encouragement
and
critique
of earlier
versions of this article.
Survey
data were
provided
cour-
tesy of the Arthur
Andersen Center
for
Family
Business.
The
support
of the H. R.
Horvitz
Family
Foundation
and
of the Family Business Programs
at the Weatherhead
School of Management,
Case
Western
Reserve
Univer-
sity, and the School of Business at the University
of
Connecticut is also acknowledged.
179
180 Academy of Management
Journal April
1996), while overlooking private
firms and the chal-
lenge of achieving within-group
goal alignment.
In this study, we examine how ownership dis-
persion among family directors influences a firm's
use of debt in 1,464 medium-sized, private, family-
owned and -managed
firms;
the average
firm in our
sample had annual sales of $36 million, had 182
employees, and had been in business for 49 years.
Our thesis is that both market conditions and the
dispersion of ownership influence the agency po-
sition of individual directors in such a way that
they are more willing to use debt and bear the
attendant risk it poses to their individual wealth
when (1) market
growth rates are high and (2) con-
trol rests in the hands of a controlling owner or
with a coalition of minority shareholders, rather
than being more equally dispersed. Consistent with
our hypotheses, our finding is that the relationship
between a family firm's use of debt and the disper-
sion of ownership among its directors forms a U-
shaped function when market growth is high, but
not when market
growth is low.
GOVERNANCE
EFFICIENCIES
IN THE
FAMILY
FIRM
Agency costs arise whenever ownership and con-
trol are separated. Agency theorists, beginning with
Fama and Jensen (1983a, 1985), have long pre-
sumed that family governance minimizes these
costs. For example, the need to monitor family
agent conduct is reduced because familiarity and
the intimate knowledge gained from long associa-
tion facilitate communication and promote cooper-
ation among family owners and family agents.
Fama and Jensen noted that "family
members have
many dimensions of exchange with one another
over a long horizon that lead to advantages
in mon-
itoring and disciplining family-related decision
agents" (1983b: 306). The need to incur bonding
costs is also reduced because family ties link them
to a kinship network that is characterized
by norms
of reciprocity, strong social ties, a shared identity,
and a common history (Ouchi, 1980). Kinship
thus tempers self-interest-and the conflict it can
cause-by fostering loyalty and commitment to the
family and the firm.
Self-interest is further tempered by parental al-
truism. This trait, which economists model as a
utility function in which the welfare of individuals
is positively linked to the welfare of others (Becker,
1981; Lunati, 1997), compels parents to be gener-
ous to their children. It also encourages family
members to be considerate of one another and to
care for each other in time of need, even to the
point of sacrifice. The result, Fama and Jensen con-
cluded, is that "special relations with other deci-
sion agents allow agency problems to be controlled
without separation of the management and control
decisions" (1983b: 306).
Altruism and kinship offset some of the ineffi-
ciencies in risk bearing that otherwise accompany
private ownership. All else being the same, private
ownership limits access to capital, forcing a pri-
vate, family-owned and -managed firm to rely on
internal sources to fund investments. The fact that
most of their wealth is invested in the firm also
tends to make private firm owners reluctant to use
debt. Should the owners be linked to the same
kinship network, however, their individual calcu-
lus for framing
investment decisions changes. Spe-
cifically, altruism and kinship can make them more
willing to use debt and bear the threat it poses to
their individual wealth, because they temper their
self-interest with concern for the welfare of the
family and firm. Altruism and kinship thus make
family directors more willing to use debt to fund
investment and pursue growth than agency theo-
rists would predict, especially when market
condi-
tions are promising. The assumption that owner-
ship will remain within the family also gives family
directors incentive to make investments that will
benefit the next generation of owners. This long-
term perspective, combined with the type of deep
knowledge that family directors acquire from life-
long involvement in the principal industry of their
family firm, makes them better able to evaluate risk
and make strategic investments (Kang, 1999).
It is tempting to conclude that family ownership
and management naturally minimize agency costs
while giving family directors the incentive to make
investment decisions that serve the best interests of
a firm and family. However, this positive portrait
is
at odds with evidence suggesting that these firms
are "plagued by conflicts" that can cause them to
flounder, if not fail (Levinson, 1971: 90) and that
they are vulnerable to a form of inertia that can
paralyze decision making and threaten firm sur-
vival (Meyer & Zucker, 1989).
This positive portrait is also at odds with the
recent study by Schulze, Lubatkin, Dino, and Buch-
holtz (2001), who found indirect support for the
thesis that altruism has a dark side. Although it can
temper self-interest and engender loyalty, commit-
ment, and a long-term perspective, altruism can
also alter the incentive structure of a firm so that
some of the agency benefits gained are offset by free
riding and other agency problems. For example,
altruism can create a sense of entitlement among
family members by encouraging CEOs (usually a
parent and/or head of household of the controlling
family) to use the firm's resources to provide family
2003 Schulze, Lubatkin,
and Dino 181
members with employment, perquisites, and priv-
ileges that they would not otherwise receive. Altru-
ism can also bias CEOs' perceptions of their em-
ployed children, which hampers their ability to
monitor and discipline them. The result, Schulze
and colleagues concluded, is that family-owner
management does not necessarily minimize the
agency cost of fractional ownership and, in some
cases, can exacerbate it.
We expand this argument in the following sec-
tions. Our thesis is that just as the separation of
ownership from control in widely held firms drives
a wedge between the interests of principal and
agent, the dispersion of ownership in family-held
firms drives a wedge between the interests of those
who lead a firm-and often own a controlling in-
terest-and other family owners. We begin by pro-
posing that, contrary
to the tenets of agency theory,
inside ownership and board oversight do not effi-
ciently resolve the agency problems experienced by
private, family-owned and -managed firms. Draw-
ing on behavioral economics theory, we then ex-
plain how private ownership and family manage-
ment can combine to raise the agency costs of
fractional
ownership, and thereby influence family
director conduct and a firm's use of debt.
THE GOVERNANCE
EFFECTS OF PRIVATE
OWNERSHIP AND FAMILY MANAGEMENT
According to the principal-agent model, inside
ownership and board oversight efficiently resolve
the conflicts caused by fractional ownership be-
cause: (1) ownership aligns inside owners' risk
preferences with those of outsiders while increas-
ing communication and cooperation among them;
(2) liquid markets limit the cost of board conflict by
making it possible for disputing parties to buy or
sell shares at a market-determined
price; and (3)
voting generates economically efficient outcomes
since it reflects the proportionate distribution of
risk and reward among a firm's owners (e.g., Al-
chian & Woodward, 1988; Fama & Jensen, 1983a;
Jensen, 1998; Jensen & Smith, 1985). We argue in
this section that none of these three governance
mechanisms operates as theorized when firms are
privately owned and family managed.
First, the agency theory assumption that in-
creased inside ownership aligns owner preferences
implies that individuals are economically rational
wealth maximizers. In contrast, behavioral econo-
mists, like O'Donoghue and Rabin (2000) and
Thaler and Shefrin (1981), have argued that indi-
viduals are motivated by an idiosyncratic set of
preferences-some economic and some noneco-
nomic in character, and some self-regarding (egois-
tic) and some other-regarding
(altruistic)-and are
driven to maximize the utility they gain from each.
Taken together, these assertions suggest that goal
alignment within any board would be difficult to
attain and sustain. Further, they suggest that con-
flicts of interest arise because resource constraints
prevent board members from maximizing their dif-
ferent types of preferences simultaneously. For ex-
ample, actions taken to promote
wealth can prevent
actions taken to promote leisure, while actions mo-
tivated by self-interest can prevent actions taken to
promote the welfare of others.
Unlike public firms, which can rely on external
governance mechanisms to minimize the adverse
effects of these internal conflicts, family firms can-
not do so because private ownership isolates them
from the discipline that external markets provide.
Moreover, altruism hampers the ability of a family
firm's principal owner (who is usually the CEO)
to
use internal governance mechanisms like monitor-
ing to minimize internal conflicts and the agency
threats they engender (Schulze et al., 2001). Field
study findings concur: family-firm CEOs tend to
rely on informal monitoring and control mecha-
nisms (Daily & Dollinger, 1992; Geeraerts, 1984)
and are notorious for avoiding disciplinary issues
that might have repercussions for familial relations
both inside and outside the firms (Meyer
&
Zucker,
1989; Ward, 1987). In sum, whereas the agency
theory assumption is that ownership and monitor-
ing efficiently align shareholder interests in public
firms, behavioral economic perspectives suggest
that ownership can have the opposite effect when
firms are private and family-managed.
Second, agency theory also suggests that market
liquidity, and hence the ability to exit a firm at low
cost, limit the potential cost of settling conflicts of
interests among the directors of public firms, be-
cause those who disagree
with the majority
opinion
can simply sell their shares at the current market
price and exit the firm. Of course, this claim rests
on the assumption that the only transaction cost
that matters is the cost of selling equity. Behavioral
agency theorists, like Wiseman and Gomez-Mejia
(1998), would take exception to this statement. Ac-
cording to their theory, which melds insights from
prospect theory with agency theory, insiders face a
number of noneconomic exit costs, including the
value of the firm-specific knowledge, experience,
and social networks that they accumulated while
employed in the firm's upper-management ranks,
as well as the emotional costs associated with a
change in status, the possible relocation of the fam-
ily, and so on.
Family-member inside directors arguably face
higher exit costs. There are no liquid markets for
182 Academy of Management
Journal April
their stock. Even if there were, exiting the firm
would still mean forgoing certain rights, perqui-
sites, and other privileges that generally come with
being employed by one's family (Schulze et al.,
2001). Moreover,
exiting might not only entail for-
going (or at least reducing) the share one expects to
inherit in the firm and/or the family's estate, but
also forgoing
or reducing benefits that might accrue
from continued close association with the firm and
family (Holtz-Eakin,
Joulfian, & Rosen, 1993). Fi-
nally, and perhaps
most importantly,
leaving a fam-
ily firm entails significant emotional costs associ-
ated with lost intimacy, reduced status, breaking
familial expectations and, in some cases, a severing
of family ties (Gersick, Davis, Hampton, & Lans-
berg, 1997).
Thus, if market
liquidity (along with monitoring
and voting) is necessary for an efficient resolution
of conflicts among members of a board of directors,
then higher exit costs should make board conflict
resolution more costly in family firms. High exit
costs, therefore,
tend to lock insider directors into a
firm, thereby making the conflicts that arise more
persistent and a convergence of interests more dif-
ficult to achieve. Thus, we again infer that a differ-
ent set of incentives is at play among inside family
directors (those who are
both directors and employ-
ees), a situation that affects board conduct in ways
that extant agency models do not predict.
Finally, agency theorists assert that although in-
side and outside owners of public firms may have
differences that defy consensus, voting assures that
the preferences of the risk-neutral majority will
prevail. Of course, this assertion relies on the as-
sumption that board members carry out their fidu-
ciary responsibilities and that the independence of
outsiders is not compromised by the influence of
insiders. This is not always the case (e.g., Finkel-
stein & Hambrick, 1996). CEOs, by virtue of their
professional and political ties as well as the author-
ity of their office, can make both inside and (some)
outside directors beholden to them (Kroll, Wright,
& Theerathorn, 1993). Also, the boards of some
firms, especially those with widely distributed
ownership and without large-block owners, may
appoint outsiders who are not vigilant in monitor-
ing and/or fail to exercise their fiduciary authority
over insiders (Walsh & Seward, 1990). Thus, en-
trenchment threatens the autonomy of such a board
and undermines the effectiveness of its oversight.
Family firms are particularly vulnerable to voting
imperfections and entrenchment. The CEO of a
family firm generally wields power that is dis-
proportionate to his or her share of ownership;
this disproportionate power stems from familial
sources (for instance, status as the head of the fam-
ily), hierarchical sources (such as status as the head
of the firm), and (because the firm is privately held)
freedom from the oversight and discipline pro-
vided by the market for corporate
control and other
sources of external governance. Not surprisingly,
family-firm CEOs tend to be entrenched; their av-
erage tenure of 24 years (Beckhard
&
Dyer, 1983) is
twice that observed in widely held firms (Hambrick
&
Fukutomi, 1991: 736). Further,
family firms tend
to have small boards of directors (there was an
average of four members per board in the sample
used in this study, whereas experts recommend
seven or more), and they tend to appoint directors
who are friends of the CEO and/or happen to have
a fiduciary relationship with the firm (such as their
attorneys and accountants), further compromising
director
autonomy and board vigilance (Ford, 1988;
Gersick et al., 1997; Nash, 1988; Ward & Handy,
1988).
In sum, the combined influence of private own-
ership and family management results in a web of
incentives that undermine a family firm's gover-
nance and raise the agency cost of fractional own-
ership. In the next section, we explain how owner-
ship dispersion influences director conduct and
family firms' use of debt.
THE EFFECTS OF OWNERSHIP
DISPERSION
AT FAMILY
FIRMS
Whereas the boards of public firms consist of
inside and outside directors, a family firm's board
consists of a principal owner (who is usually, but
not always, the founder and CEO) and minority
shareholders (who tend to be members of the nu-
clear and/or extended family and are often, but not
always, employed by the firm). Family-firm
owner-
ship tends to get dispersed in a somewhat episodic
and "stepwise" fashion over a relatively long pe-
riod of time, with shares usually passed from
parent to child around the time of the principal
owner's retirement and/or death.
While patterns of ownership dispersion vary,
ranging from primogeniture (in which leadership
and control of the voting stock passes to the first-
born) to coparcenary (in which offspring receive
relatively equal shares), the tendency in the United
States is to grant the most shares to the chief exec-
utive, and more shares to offspring who are em-
ployed by the firm than to those who are not. (Test-
ing for confirmation, we found the first tendency
held true in all the cases in which we could iden-
tify the occupation of a firm's principal share-
holder, and the second tendency held true for the
preponderance of the 1,464 family firms repre-
sented in this sample.)
2003 Schulze, Lubatkin,
and Dino 183
The ownership of a family firm generally passes
through
three broad stages of dispersion (Gersick
et
al., 1997): controlling owner, in which most shares
are held by the founder, or in the case of later
generations, by a single individual; the sibling part-
nership, in which relatively equal proportions of
ownership are held by members of a single gener-
ation; and the cousin consortium, in which owner-
ship is further fractionalized as it is passed on to
include third and later generations. Although the
conflicts that accompany each stage differ, the
agency model that we describe below explains the
conduct indigenous to each.
Controlling Owner
As we previously noted, in the principal-agent
model, owners who manage a private (family or
nonfamily) firm define its value in terms of their
personal utility. Thus, they have powerful incen-
tives to pursue options that they perceive as best
and to bear the associated risks to the point where
the marginal
benefit received is offset by the threat
the risks pose to their personal wealth (Jensen &
Meckling, 1976). Proponents of behavioral agency
theory (e.g., Wiseman & Gomez-Mejia,
1998) have
refined that insight, arguing
that the amount of risk
that these owners are willing to bear is based, at
least in part, on how they frame
their expectations.
For example, the owner-managers
of private firms
should have incentive to invest when they expect
conditions in their firms' markets
to grow, but not
when they expect market
growth
to decline or slow.
Lacking access to the equity markets, however,
their ability to invest is limited by the availability
of internally generated funds (Casson, 1991)-un-
less, of course they take on or increase debt. It
follows that the owner-managers of private firms
will be more willing to use debt, and more willing
to bear the threat it poses to their individual
wealth, during periods of high market
growth than
they will during periods of decline or low market
growth.
We posit that parental altruism causes owners to
pursue first-best actions when a private firm is fam-
ily owned and is managed by a controlling owner.'
Altruism is a trait that positively links the control-
ling owner's welfare, as head of the family, to that
of other family members (Schulze et al., 2001).
Altruism thus compels the controlling owner to
consider the needs of the firm and each family
member when defining her or his first-best options
(for instance, it may make the controlling owner
more willing to pursue investments with longer-
term payoffs). Over time, however, the economic
incentive to do what maximizes personal utility
can blur the controlling owner's perception of what
is best for the firm or family; self-interest and the
firm's and family's best interests may be viewed as
one and the same in what we might call the "what's
good for GM"
phenomenon. For instance, age may
cause the controlling owner to avoid investments
that other family members favor because he or she
views the investments as too risky or as personally
threatening-in the case, perhaps, of their requiring
the controlling owner to learn new skills. Conflicts
of interest can therefore arise that give family mem-
bers reason to question the extent to which they can
rely on the controlling owner to make decisions
that they deem as being in the family's best inter-
ests. The family members thus have incentives to
monitor the controlling owner and incur other
agency costs in an effort to assure that their best
interests are being served.
Moreover,
family members and controlling own-
ers face different sets of incentives, and thus hold
different views of what investments are best.
For example, like their counterparts in public
firms, family-member employees of family-owned
and -managed firms bear only a fraction of the risk
associated with an investment decision but, unlike
their counterparts, are able to enjoy a dispropor-
tionate share of the benefits owing to their family
status. They are also likely to feel entitled to these
benefits since, as family members, they believe that
they own de facto options in the firms, or a residual
but legitimate claim on them in the form of an
inheritance at a future date (Holtz-Eakin et al.,
1993; Stark & Falk, 1998).
This sense of entitlement has two important
agency consequences. First, it can cause employed
family members and their prospective heirs to be-
come fiscally conservative, if not loss-averse, since
added risk threatens the value of their anticipated
inheritance. Put differently, an endowment effect
(Wiseman & Gomez-Mejia, 1998) engenders loss
aversion by altering their risk-return calculus. Sec-
ond, when the sense of entitlement is coupled with
high exit costs and the perception that the potential
cost of exiting a firm exceeds the expected value of
other opportunities, hopeful family heirs can be-
come locked into a dependent relationship with the
firm. This makes it possible for a state of "double
moral hazard" (Gupta & Romano, 1998) or "owner
opportunism" (Perrow, 1986) to develop in which
' First-best
actions maximize a principal's
expected
utility subject
to the constraint that an agent
receives his
or her reservation
utility (the utility the agent could
receive
by redeploying
resources
to their
best
alternative
use).
184 Academy of Management
Journal April
the controlling owner has the power, and perhaps
the incentive, to unilaterally change his or her es-
tate plans, thereby placing family members' claims
to the firm at risk.2
Although double moral hazard ordinarily gives
agents incentives to invest resources in monitoring
a principal's conduct, family members' ability to
influence a controlling family owner is constrained
by both their minority (or even nonshareholder)
status and the added authority controlling owners
have by virtue of being the heads of the family
households and, in many cases, the founders of the
firms. These constraints, combined with the risk
that the controlling owner may undertake invest-
ments that other family members do not view as
best, gives these family members incentive to prefer
consumption to investment, and to do so (in the
form of pecuniary and nonpecuniary benefits) at
rates that are high relative to their ownership
stakes. Consumption, of course, precludes alterna-
tive uses for the funds that are consumed.
The net result is a pattern
of incentives that is the
reverse of that theorized to exist at widely held
public firms. Whereas fractional ownership at
widely held public firms gives insiders incentives
to free ride on the outside owners equity, we argue
that it gives family insiders (family-member
direc-
tors and employees) incentives to free ride on the
controlling owner's equity. Controlling owners are
likely to recognize that (some) family members are
free riding on their (qua the family's) holdings.
And, although altruism and the repercussions that
disciplinary actions might have for family relations
compels the controlling owners to accept this free
riding, it also gives them incentive to be wary of the
investment decisions that the family insiders might
recommend.
Thus, although we anticipate that the controlling
owners of family firms, and especially founders,
will initially have strong incentive to use debt to
fund investments that they think are the best, we
argue in the next section that their ability (and
willingness to do so) will decline as the percentage
of ownership held by loss-averse, consumption-
oriented family members increases.
Sibling Partnership
The agency dynamics during the sibling partner-
ship stage become more problematic. As in the
controlling owner stage, the principal shareholder
(that is, the largest shareholder) in a sibling part-
nership is likely to serve as the CEO,
to control the
largest single block of ownership, and, by virtue of
his or her office, to continue to wield influence that
is disproportionate
to ownership share. And, like a
founder, the principal shareholder can be expected
to fulfill a quasi-family-leader
role, using the firm's
resources to promote family welfare and to favor
the reinvestment of earnings over the consumption
of those earnings
via dividends and other payments
(Gersick et al., 1997). Yet typically, the principal
shareholder in a sibling partnership is neither the
founder of the family firm nor the biological head
of the family and, lacking that authority and influ-
ence over the siblings, is less able to obtain-
whether by cooperation, co-option, or edict-the
support of the other family directors for making
investments and pursuing the opportunities that he
or she believes to be the best options.
In addition, and in line with both the family
business literature (Gersick et al., 1997) and eco-
nomic theory about altruism, all sibling partners
are also likely to be more concerned about their
own welfare and that of their immediate families
than they will be about each other's welfare. (Ac-
cording to this theory, a parent's concern for her or
his children tends to be stronger
than the children's
concern for the parent [Stark & Falk, 1998], and
altruistic ties among members of a nuclear family
tend to be stronger than those among members of
an extended family [Becker, 1981]). Thus, agency
conditions in sibling partnerships resemble those
in the controlling owner stage, with sibling part-
ners having incentives to use a family-firm's re-
sources to maximize their own utility; acting on
these incentives can, again, engender double moral
hazard problems and conflict between the sibling
partners.
The risk of intrafamily conflict is further exacer-
bated as families age. Siblings who were once able
to forge an effective partnership may find it torn
apart as resource constraints force them to make
hard decisions about dividend payout policy (to
fund college tuitions, for instance) and/or the in-
volvement of their adult children in the firm's op-
erations. The risk of intrafirm conflict also rises if
ownership is distributed somewhat equally among
a principal shareholder and sibling partners. In this
scenario, one loss-averse sibling can prevent others
from putting a firm's resources to their desired
(first-best) use. Consequently, we would expect sib-
ling partners to have the incentive to engage in
various political maneuvers, like vote swapping
and "hostage taking," actions that might cause a
series of compromises, ill-will, and second-best de-
2 Witness the now famous case of the former
Playboy
model, Anna Nicole Smith, who inherited $475 million
(reduced on legal appeal to $88 million) after a brief
marriage
to a septuagenarian
husband, much to his chil-
dren's dismay (Miller, 1999).
2003 Schulze, Lubatkin,
and Dino 185
cisions about growth, investments in new technol-
ogy, and so on.
The result could be a state of paralysis in which
no one sibling is willing to bear added risk or use
debt to pursue opportunities that others believe are
best. Thus, whereas the principal-agent model as-
sumption is that increased ownership aligns the
interests of rival parties, we posit that increased
ownership dispersion among sibling partnerships
will engender misalignment and loss aversion. Put
differently, the increased concern for their own
children, the added pressure from outside family
directors (and in-laws) to sustain or enhance the
rate of dividend payout, and the aging siblings'
increasing reluctance to bear risk can cause a firm
to reduce its use of debt, even when market
condi-
tions are perceived as favorable, and to get bogged
down in the types of conflict that cause many fam-
ily firms to flounder or fail (Levinson, 1971).
Cousin Consortium
Finally, we expect that the agency position of a
principal and the minority owners will become
more aligned in a family firm's cousin consortium
stage. By the time the firm enters this stage, own-
ership has become more dispersed, or fractional-
ized, and it is less likely that a single individual
owns a controlling or majority interest in the firm.
This situation increases the degree of relative influ-
ence that each family director has on the future
value of his or her claim on the firm, thereby mit-
igating the double moral hazard problem that char-
acterized owner control and sibling partnership.
Inside directors, it follows, should be less con-
cerned with consumption and more concerned
about the future value of their estates and how that
value will be affected by any future dilution of
ownership. The end result is an increase in the
alignment of interest that exists among board
mem-
bers and, hence, reduced agency costs.
In the cousin consortium stage of a family firm,
ownership has likely passed to members of the
extended family, the majority of whom are not em-
ployed by the firm. All things being the same, these
outside family members are less "overinvested" in
the firm and, so, they should have risk preferences
that are more akin to those of institutional investors
and others who invest in public firms. We therefore
anticipate that cousin consortiums' managers are
both more willing to use debt to pursue their ob-
jectives and, because of the dispersion of owner-
ship, more able (and more likely) to bear that risk.
It remains true, however, that because these firms
are private (there is no liquid market for their
shares), outside family owners can benefit from
growth in earnings (through the payout of divi-
dends), but not from growth in valuation. Conse-
quently, during this ownership stage, most outside
family shareholders will continue to favor con-
sumption, while insiders (whose combined equity
holdings usually represent the majority) will con-
tinue to favor investment owing to their concern
about the effect of further dilution on the value of
their estates. Thus, the primary
challenge facing the
cousin consortium boards of family firms is to in-
vest in growth while maintaining a dividend level
that satisfies outside family owners (Gersick
et al.,
1997).
In summary, we anticipate that the dispersion of
ownership that characterizes the cousin consor-
tium stage engenders a coalition in which owner-
ship brings the interests of the inside family direc-
tors into alignment. We do not expect that this
alignment will be as stable as it is for widely held
public firms, because ownership is not as dispersed
and the problems of market
liquidity and exit costs
remain. Nevertheless, we posit that for family firms
in the cousin consortium stage, the greater
the own-
ership dispersion (and the smaller the average
shareholding), the more likely their boards will be
to favor growth and, in the absence of the ability to
issue equity or cut dividends, the more likely they
will be to risk the use of debt to fund growth.
Hypotheses
Taken together, our arguments support Morck
and his coauthors' (1988) conjecture that disper-
sion of ownership has a significant influence on
board conduct. By extending their arguments
to the
domain of private family-owned and -managed
firms, we hypothesize that family boards will be
more willing to use debt when ownership is either
concentrated
in the hands of a controlling owner or
dispersed into the hands of many owners (as in a
cousin consortium), and less willing to use debt
when ownership is split into relatively equal
proportions (as in a sibling partnership). Stated
formally:
Hypothesis 1. The relationship between a fam-
ily firm's use of debt and the dispersion of its
ownership can be graphed as a U-shaped
curve.
In line with behavioral agency theory, we antic-
ipate that a board's willingness to use debt varies
with growth conditions in a family firm's market.
All else being the same, we predict that family
firms will increase borrowing during periods of
high growth and, because of their dependence on
internal cash flows and limited access to external
186 Academy of Management
Journal April
capital markets, will reduce it during periods of
low growth (Wright
et al., 1996). Accordingly, we
tested our hypotheses under conditions of both
high and low market
growth rates, positing that the
relationships stated in Hypothesis 1 will be sup-
ported during periods of high growth, but not dur-
ing periods of low growth.
Hypothesis 2. The relationship between a fam-
ily firm's use of debt and the dispersion of its
ownership is moderated by the growth rate of
the firm's market.
METHODS
Sample
Reliable information
on family firms is extremely
difficult to obtain (Wortman, 1994). Public infor-
mation is unreliable because most family firms are
privately held and have no legal obligation to dis-
close information. Government documents and
Dunn and Bradstreet are also of little use because
family-managed firms are not listed as a separate
category of business organization. Finally, it is dif-
ficult for researchers to collect primary data or to
target selected groups of family-managed
firms for
study because there is no reliable way to identify
family firms a priori (Daily &
Dollinger, 1993). Con-
sequently, researchers are forced to rely on self-
reported
data, sample from a broad
population, and
identify family-managed firms ex post (Daily &
Dollinger, 1992, 1993; Handler, 1989).
We field-tested our hypotheses using data from
one of the largest
and most comprehensive surveys
ever conducted on family firms (Gersick et al.,
1997), a 1995 survey of American
family businesses
that was designed and administered by the Arthur
Andersen Center for Family Business. Since all of
the firms in the sample were privately held, and the
data were confidential and proprietary, we were
unable to independently establish the data's reli-
ability. Andersen's statisticians assured us, how-
ever, that they were reliable and representative of
the population.
While the use of secondary data can limit gener-
alizability, Ilgen (1986) and Sackett and Larsen
(1990: 435) pointed out that representativeness is
less of a concern when a sample typifies the rele-
vant population and the research question con-
cerns whether the hypothesized effects can occur,
as opposed to concerning the frequency or strength
of observed effects. The Arthur Andersen data are
well suited to this task because the survey was
designed to obtain "reliable benchmarks" about
American family businesses (Arthur Andersen &
Co., 1995: 3).
Before we mailed the survey to the chief execu-
tives of 37,304 privately held U.S. family busi-
nesses, we had the items in this survey reviewed by
a focus group of family business owners and pilot-
tested it on a hold-out sample. A single mailing
yielded 3,860 responses within one month; this
constitutes a response rate of 10.3 percent, which is
comparable to "the 10-12 percent rate typical for
studies which target
executives in upper echelons"
(Geletkanycz,
1997: 622; Hambrick,
Geletkanycz, &
Fredrickson, 1993; Koch &
McGrath,
1996) or chief
executives in small to midsized firms (MacDougall
& Robinson, 1990).
Because of the a priori selection problems,
Andersen survey respondents ranged from "mom
and pop" proprietorships to large family-managed
corporations. We therefore applied a number of ex
post screening criteria to the data. First, we deleted
334 partnerships and proprietorships because dif-
ferent laws and tax policies influence their gover-
nance. Second, we dropped 1,650 cases because
data about firm ownership and/or board composi-
tion were missing, and we dropped another 209
because some information about the other 13 vari-
ables included in the regression analyses was miss-
ing. (We tested and found no differences in the use
of debt or in the mean values of our model's inde-
pendent variables between cases that included data
about ownership or board composition and those
that did not.) Finally, by deleting 203 firms that had
$5 million or less in sales, we excluded "lifestyle
firms" (small firms that might be operated mainly
for the purpose of "income substitution" [Allen &
Panian, 1982]); firms whose use of debt might be
biased by their receiving subsidies;3 and others for
which growth might not be a strategic objective
(see, for example, Rubenson and Gupta [1996] and
Carland, Hoy, Boulton, and Carland [1984], who
also excluded these types of firms from samples,
arguing
that growth may not be among their strate-
gic objectives). Larger firms are less likely to be
operated in this manner since the demands of man-
aging them mitigate a family's or a family CEO's
primary motive for suppressing growth-to more
easily maintain managerial and ownership control
(Daily & Dollinger, 1992; Whisler, 1988). Thus, the
final sample consisted of 1,464 firms. Our average
firm had annual sales of $36 million, had 182 em-
ployees, and had been in business for 49 years.
3 For example,
95 percent
of all loans guaranteed
by
the Small Business
Administration
go
to firms
with fewer
than 50 employees (Small Business Administration,
2002),
a statistic
that characterizes
the firms
in this sam-
ple that
had less than
$5 million
in annual
sales.
2003 Schulze, Lubatkin,
and Dino 187
Variables
Dependent variable. We used debt as the depen-
dent variable. As we previously discussed, the
ownership structure of privately owned firms, un-
like that of their public counterparts, does not al-
low for unrestricted risk bearing via the issue of
common stock. Capital investment is thus limited
to that which can be supported by internally gen-
erated funds and the shareholders' willingness to
bear the risk that debt poses to their individual
wealth (Casson, 1999). Although private ownership
thus engenders fiscal conservatism, all else being
the same, we deduce from a behavioral agency view
that directors are willing to incur debt to pursue the
investments that they perceive to be the best ones,
particularly when the directors expect their firm's
market to grow.
Our variable measuring debt is a six-level indi-
cator of a firm's debt-to-equity ratio; (from 1 to 6,
codings were "no debt," "1-25%," "26-50%,"'
"50-100%," "101-200%," and "over 200%"). The
mean debt-to-equity
ratio among our sample group
of mature firms was 2.57, which interpolates to
about 18 percent. The measure was self-reported
and, as is the case with virtually all privately held
firms, objective measures were not available. How-
ever, performance
measures reported
by executives
have been shown to be reliable (Nayyar, 1992; Tan
& Litschert, 1994), particularly when reported on
anonymous surveys (Dillman, 1978; Nunnally,
1978). The impact of common method bias, which
arises when a common method (such as a survey) is
used to gather data about both independent and
dependent variables, should also be less here than
it might be for other types of studies because social
desirability and other sources of bias are dimin-
ished when variables are demographic,
descriptive,
and/or nonaffective, as are most of our variables
(Crampton
& Wagner, 1994).
Covariates. We included nine covariates to re-
duce variance that would be extraneous to the re-
search question or that might confound interpreta-
tion: firm size, firm age, multiple family ownership,
number of family employees, exports as a percent-
age of sales, information technology intensity, CEO
tenure, average board tenure, board size, family
ownership goal, and ownership held by the board.
Each covariate is described and its use is justified
in the Appendix.
Independent variable. The mean percentage of
shares controlled by the boards of the firms in our
sample was 90 percent, and the largest shareholder
controlled an average 52 percent of the votes (the
largest shareholder was a sole owner in 18 percent
of the cases). In contrast, the second through the
fifth largest shareholders on the family-firm
boards
controlled, on the average, 25, 8, 3, and 1 percent of
the votes. Interestingly, the remaining members
about whom we had information (that is, the sixth
through the eighth largest shareholders) together
controlled barely half of 1 percent of the votes.
Moreover, extreme variance in the shares held by
these board members was indicated by standard
deviations that ranged
to values up to ten times the
size of the mean. This distribution is not a surprise,
given that only 166 firms in the sample had six or
more board members, and 107 had seven or more.
Following Tabachnick and Fidell (1989), we there-
fore dropped these observations in our primary
analyses of Hypotheses 1 and 2 and based our cal-
culation of the independent variable, balance of
voting power, on the total shares held by the five
largest shareholders who served on a board. We
also tested the sensitivity of our results by adding
the shares held by the sixth largest shareholder,
then the seventh, and then the eighth to the calcu-
lation and repeating our tests; results are reported
below.
Balance of voting power was calculated as the
sum of the squares of the minority board members'
percentage share of votes divided by the square of
the largest shareholder's percentage share of the
votes. The sum of squares is used here like the
Herfindal index, which economists use to describe
the distribution of market share among industry
participants:
the sum of squares captures
the effects
of different distributions of ownership. Higher val-
ues are associated with increased power held by
individual shareholders, and lower values, with a
more equal, and/or more diffuse dispersion of
power on a board.4 The balance of voting power
variable therefore captures the variance associated
with changes in the dispersion of ownership that
would not appear if the ratio were computed using
a simple sum of each director's
shareholdings. Val-
ues of 1:1 or less indicated the distribution of own-
ership favored the largest shareholder, and values
higher than this indicated that the dispersion of
ownership favored the minority shareholders. The
dispersion of ownership favored the largest share-
holder in 73 percent of our cases and favored mi-
nority shareholders in the remaining 27 percent of
the cases.
Moderator. We used a dummy variable, industry
sales growth, to test the proposition that investor
expectations influence a family firm's use of debt.
4 For example, a change in minority ownership disper-
sion from 25:25 to 20:20:10 will cause the value of the
numerator
to fall from
1,250
to 900.
188 Academy of Management
Journal April
TABLE 1
Descriptive Statistics and Correlationsa
Mean s.d. 1 2 3 4 5 6 7 8 9 10 11 12 13
1. Debt 2.56 1.45
2. Firm size 37.44 121.71 .18
3. Firm age 48.98 27.00 -.02 .10
4. Multiple family ownership 0.91 0.29 .05 -.03 .01
5. Number of family-member employees 3.43 2.03 .02 .09 .01 -.07
6. Exports as a percentage of sales 1.51 0.82 .01 .07 .01 .00 -.03
7. Information technology intensity 3.07 0.92 .08 .17 .06 .03 .00 .06
8. Industry sales growth 0.52 0.50 .04 -.02 .08 -.01 .02 .24 -.02
9. Ownership held by board 89.62 19.61 .01 -.15 -.16 .08 -.07 -.09 -.02 -.06
10. CEO tenure 2.45 1.13 -.08 -.07 -.10 .02 .09 -.03 -.05 .00 .01
11. Average board tenure 17.52 7.75 -.11 -.06 .13 -.03 .06 -.10 -.02 .00 .08 .13
12. Board size 3.98 1.75 .04 .21 .22 -.05 .25 .08 .04 .08 -.23 -.04 -.20
13. Family ownership goal 0.68 0.47 -.03 -.03 -.04 .07 -.04 -.05 .04 -.09 .12 .01 .03 -.09
14. Balance of voting power 0.70 0.81 -.01 .02 .08 -.14 .27 -.03 .01 .02 .00 -.11 .06 .23 -.03
a n = 1,464. Correlations larger than .04 are significant at p - .05.
Since the industry categories identified in the
Andersen survey do not correspond directly to SIC-
based industry classifications, we coded industry
sales growth 1 if the mean of the reported growth in
sales for the industry category
was greater
than the
median ratio for all industry categories, and we
coded the variable 0 if mean growth was below the
median. Although coarse-grained,
this measure dis-
tinguishes industries that enjoyed high levels of
sales growth during this period (for instance, man-
ufacturing and telecommunications) from those
that did not. Further, we were unable to employ
financial statistics derived from SIC-based data for
control purposes since such data include informa-
tion from large, widely held businesses whose mar-
kets and capital structures differ markedly from
those of family firms.
RESULTS
Table 1 reports descriptive statistics (unstand-
ardized) and Pearson
correlations,
and Tables 2 and
3 report the results for all regression analyses. The
change in explained variance (F) associated with
the covariate set ranges from 3.52 (p s .001) to 7.17
(p d .000), and the F-statistic associated with the
set of hypothesized variables, after hierarchically
adjusting for covariates, ranges from 6.54 (p .01)
to 10.99 (p 5 .000).5 Hypothesis 1 was tested in
both the full sample (Table 2) and in the two in-
dustry (high and low industry growth) subsamples
(Table 3). We used only the full sample to test
Hypothesis 2.
We used moderated hierarchical polynomial re-
gression analysis to confirm that industry sales
growth does, indeed, influence a family-owned
and -managed
firm's use of debt. In model 1 (Table
2), centered variables (and their product terms)
were entered hierarchically for both the balance of
voting power and its square, and then, in the next
step, the products of the independent variable and
its square with industry sales growth were entered.
As one would expect if the hypothesized relation-
ships were nonlinear and moderated, the product
of balance of voting power and industry sales
growth (p 5 .009) was negatively associated with
the use of debt, while the product of balance of
voting power squared with industry sales growth
was positively associated with the use of debt (p -<
.001). The significance of the product terms indi-
cates support for Hypothesis 2. In addition, and
consistent with Hypothesis 1, the negative value of
the first coefficient, combined with the positive
sign of its square, suggests that the relationship is
positively U-shaped over the relevant range.
We then tested the sensitivity of these results by
using the three alternative calculations of balance
of voting power previously mentioned to test Hy-
5
Overall,
the regression
models explain
from 5 to 6
percent
of the variance in our
dependent
variable. These
small effect sizes are likely the result
in part
(1) of the
high heterogeneity
the sample
contained
by virtue
of its
very
large
and diverse
representation
of firms
and
indus-
tries and (2) of the categorical
nature
of the dependent
variable. Cohen and Cohen (1980) noted, however, that
an advantage
of large samples
is that
they give analysts
the ability
to detect small effects
with a high degree
of
confidence. Kraemer and Thiemann (1987: 105) esti-
mated
that the reliability
or power
of a sample
to detect
observed effect sizes is .80.
2003 Schulze, Lubatkin,
and Dino 189
TABLE 2
Results of Full-Sample Regression Analyses for Debt
Variables
Covariates
Firm size 0.17*** 0.17*** 0.17*** 0.17***
Firm age -0.03 -0.03 -0.03 -0.03
Multiple family ownership 0.06* 0.05* 0.05* 0.05*
Number of family-member employees 0.02 0.02 0.02 0.03
Exports
as a percentage
of sales -0.02 -0.02 -0.02 -0.02
Information
technology intensity 0.05* 0.05* 0.05* 0.05*
Ownership held by board 0.04 0.04 0.04 0.04
CEO tenure -0.06* -0.07* -0.07* -0.06*
Average
board tenure -0.09*** -0.09*** -0.09*** -0.09***
Board size -0.01 -0.01 -0.01 -0.01
Family ownership goal -0.03 -0.03 -0.03 -0.03
Industry
sales growth 0.06* 0.06* 0.06* -0.01
Predictors
Balance of voting power -0.04 -0.04 0.05
Balance of voting power squared 0.03 0.03 0.08
Interactions
Balance of voting power x industry sales growth -0.01 -0.14**
Balance of voting power squared x industry sales growth 0.02***
R2 .05
Adjusted R' .05 .05 .06
F 7.17*** 6.23*** 5.84*** 6.19***
AF 0.36 0.10 10.99***
n 1,464 1,464 1,464 1,464
*p - .05
** ps .01
S** p - .001
potheses 1 and 2. All results (available on request)
were correctly signed, and although
their
significance
levels weakened incrementally
as shareholders were
added to the calculation, balance of voting power
(p d .08)-but not its square-became marginally
insignificant
when calculated
using information from
all eight shareholders. Given the extreme variance
associated
with the addition of this information to the
computation of the variable (previously discussed
and reported), we concluded the results were not
highly sensitive to its computation.
Results of the subgroup analysis (Table 3, models
2 and 3) lend further support to the hypotheses.
The balance of voting power variable and its square
are significant (p d .03 and p df .01, respectively)
and correctly signed when industry sales growth is
high, and are they insignificant when industry
sales growth is low.
DISCUSSION AND CONCLUSION
Family firms constitute over 80 percent of all
business organizations in the United States and are
the dominant form of economic enterprise
through-
out the world (La Porta, Lopez-de-Silanes, &
Shleifer, 1999), yet over two-thirds of first-genera-
tion family firms do not survive to a second gener-
ation of family ownership (Gersick et al., 1997).
Understanding how the agency positions of the
controlling owner and the minority shareholders
influence the conduct of family firms is a small step
toward understanding why many fail. In this arti-
cle, we drew from behavioral and economic theo-
ries to argue
that the incentives facing the directors
of privately held, family-managed
firms are differ-
ent from those facing the directors of widely held
public firms. Whereas ownership is expected to
align incentives in public firms, we found that how
ownership is dispersed among the various family
owners of a privately held firm affects such deci-
sions as the use of debt. Our findings are therefore
not only consistent with Morck and his colleagues'
(1988) conjecture, but also suggest that the princi-
pal-agent model requires modification before being
applied to family firms. Furthermore,
our findings
about the moderating influence of industry growth
are also consistent with the views of behavioral
agency theorists like Wiseman and Gomez-Mejia
(1998), who pointed out that the amount of risk that
190 Academy of Management
Journal April
TABLE 3
Results of Subsample Regression Analyses for Debt
Variable Model 2: High Industry Growth Model 3: Low Industry Growth
Covariates
Firm size 0.17*** 0.17*** 0.17*** 0.17*** 0.17*** 0.17***
Firm age -0.05 -0.05 -0.05 -0.01 -0.01 -0.01
Multiple family ownership 0.08* 0.08* 0.07* 0.04 0.03 0.04
Number of family-member employees 0.02 0.02 0.03 0.02 0.02 0.02
Exports as a percentage of sales -0.03 -0.03 -0.03 -0.01 -0.01 -0.01
Information technology intensity 0.08* 0.08* 0.08* 0.02 0.02 0.02
Ownership held by board 0.02 0.02 0.02 0.06 0.06 0.06
CEO tenure -0.06 -0.06 -0.06 -0.07 -0.07 -0.07
Average board tenure -0.10** -0.09** -0.09** -0.08* -0.08* -0.08*
Board size -0.03 -0.03 -0.05 0.01 0.01 0.01
Family ownership goal -0.05 -0.05 -0.05 0.01 -0.01 -0.01
Predictors
Balance of voting power -0.02 -0.12* -0.01 0.03
Balance of voting power squared 0.13** -0.05
R2 .05 .04
Adjusted R2 .06 .06 .04 .04
F 4.48*** 4.13*** 4.31*** 3.52*** 3.26*** 3.05***
AF 0.25 6.54** 0.77 0.35
n 764 764 764 700 700 700
* p - .05
** p .01
*** p .001
owners will bear is a function of how they frame
their expectations in terms of opportunities.
Our story identifies two interesting de facto re-
versals of incentives, patterns that diverge from
predictions based on the conventional agency
model. First, we argue that, whereas "blockhold-
ing" in widely held public firms reduces the risk
that insiders will free ride on outside owners' eq-
uity, controlling ownership in a family firm (the
counterpart of blockholding in public firms) can
give family-member firm employees and directors
the incentive to free ride on the controlling owner's
equity. Second, we argued that, whereas the out-
side shareholders in widely held public firms have
the incentive to promote investment and growth-
oriented risk taking, dispersion of ownership can
give outside shareholders at private family firms
the incentive to favor consumption. This story is
also interesting from a theoretical perspective,
since it identifies at least one population in which
information about ownership dispersion, as well as
ownership concentration, is needed to predict di-
rector (and board) conduct.
Results from our field study of 1,464 family firms
support our hypotheses that their use of debt has a
curvilinear (U-shaped) relationship to the disper-
sion of ownership among voting members of their
boards of directors, particularly during periods of
market
expansion. The nonlinear relationship sug-
gests that family firms are most vulnerable to con-
flict, and least willing to bear added risk, when
ownership is split in relatively equal proportions.
Interestingly, the fact that this distribution ap-
peared in only 22 percent of our sample firms sug-
gests that most family firm owners may take such
risk into consideration when making their estate
plans. This speculation is consistent with the views
of Gersick and his coauthors (1997), who noted that
successful sibling partnerships are rare because
they are so difficult to manage and recommended
that founders and controlling owners settle their
estates in a manner that prevents the development
of sibling partnerships.
The purpose of our empirical tests, however, was
to lend credibility to our theory, not to confirm its
validity or determine the strengths of its effects.
Indeed, we cannot claim that our tests were confir-
matory since we used cross-sectional data and re-
lied upon survey data gathered for other purposes.
However, we think these tests lend credibility to
our theory since the firms represented by the
Andersen survey are typical of a population of
firms that is rarely studied and whose data are
difficult (and quite expensive) to obtain (Gersick
et
al., 1997: 25; Sackett & Larsen, 1990). The size of
our sample also gave us sufficient power to detect
small effects and yet conclude with a high degree of
confidence that these results are not the product of
2003 Schulze, Lubatkin,
and Dino 191
chance. Although some of the measures are coarser
than we would have liked (for example, balance of
voting power was computed from a simple sort of
shareholdings by size, which captured only the
average
effects of ownership dispersion and its spe-
cific effects), that coarseness also lends a conserva-
tive bias to the analysis, since coarse measures de-
flate variance and the likelihood of obtaining
significant results (Hunter & Schmidt, 1990). The
fact that we obtained significant results, despite the
type of measures used and the presence of nine
control variables, suggests that these results are
robust. Future research must, however, address
these weaknesses through use of more appropriate
survey research methodology and finer-grained
measures.
The effect of ownership dispersion on goal align-
ment within family-firm
boards is a complex issue,
and this study investigates only one of its aspects.
Future studies might examine the relationship be-
tween minority shareholder influence and the dis-
persion of ownership among minority members.
We suspect that a relatively equal distribution of
ownership among fewer minority owners promotes
a more stable coalition that would be more able to
influence firm conduct. Future studies might also
examine whether the severity of the double moral
hazard problem varies with ownership stage. We
suspect that this problem is more likely to manifest
itself under a controlling owner than under a sib-
ling partnership.
We also suspect that this problem
may be more common when family firms are
owned by more than one family than when differ-
ent branches of the same family compete on the
boards and for the firms' resources.
Future studies might examine whether our hy-
potheses apply without modification to family-
controlled public firms. At what level of ownership
and control might the controlling owners' concern
for family welfare start to generate agency costs for
outside shareholders?
Are double agency problems,
or the allegiance of a CEO to both stockholders
and family, more problematic in public owner-
controlled firms (like Microsoft), sibling partner-
ships (such as Wal-Mart), or cousin consortiums
(like the Ford Motor Company)? And what is the
effect of family ownership on the outside owners'
agency costs? (The recent proxy fight at Hewlett-
Packard over a proposed merger with Compaq
Computer is one case germane to this question.)
These and other interesting questions that can
enrich corporate governance theory remain. Our
study represents an early attempt to pinpoint the
dynamics of ownership and control in family firms.
By showing that the dispersion of ownership influ-
ences family firms' use of debt, we provided a long
overdue response to Morck and colleagues' (1988)
call for research, while at the same time revealing
information about an economically important
pop-
ulation of firms that has been largely neglected by
researchers.
REFERENCES
Alchian, A. A., &
Woodward, S. 1988. The firm is dead:
Long live the firm: A review of Oliver E. William-
son's The economic institutions of capitalism. Jour-
nal of Economic Literature, 26: 65-79.
Allen, M. P., &
Panian, S. K. 1982. Power, performance,
and succession in the large corporation.
Administra-
tive Science Quarterly, 27: 538-547.
Arthur Andersen & Co. 1995. 1995 American family
business survey. St. Charles, IL: Arthur Andersen
Center for Family Business.
Becker,
G. S. 1981. A treatise on the family. Cambridge,
MA: Harvard
University Press.
Beckhard,
R., &
Dyer, W. G., Jr.
1983. Managing
continu-
ity in the family owned business. Organizational
Dynamics, 12 (1): 5-12.
Carland,
J. W., Hoy, F., Boulton, W., &
Carland,
J. A. C.
1984. Differentiating
entrepreneurs
from small busi-
ness owners: A conceptualization. Academy of
Management Review, 9: 354-359.
Casson, M. 1999. The economics of the family firm.
Scandanavian Economic History Review, 47(1):
10-23.
Cohen, J., & Cohen, P. 1983. Applied multiple regres-
sion/correlation analysis for the behavioral sci-
ences. Hillsdale, NJ:
Erlbaum.
Crampton,
S. M., & Wagner,
J. A. 1994. Percept-percept
inflation in microorganizational
research: An inves-
tigation of prevalence and effect. Journal of Applied
Psychology, 79: 67-76.
Daily, C. M., &
Dollinger,
M. J. 1992. An empirical exam-
ination of ownership structure in family and profes-
sionally-managed firms. Family Business Review,
5(2): 117-136.
Daily, C. M., &
Dollinger, M. J. 1993. Alternative
methods
for identifying family vs. nonfamily managed small
businesses. Journal of Small Business Manage-
ment, 31(2): 79-90.
Demsetz, H. 1973. Industry
structure,
market
rivalry,
and
public policy. Journal of Law and Economics, 16:
1-10.
Dillman, D. A. 1978. Mail and telephone surveys: The
total design method. New York:
Wiley.
Fama, E., & Jensen, M. C. 1983a. Agency problems and
residual claims. Journal of Law and Economics, 26:
325-344.
Fama, E., &
Jensen,
M. C. 1983b. Separation
of ownership
192 Academy of Management
Journal April
and control. Journal of Law and Economics, 26:
301-325.
Fama,
E., &
Jensen,
M. C. 1985. Organizational
forms and
investment decisions. Journal of Financial Econom-
ics, 14: 101-119.
Finkelstein, S., &
Hambrick,
D. C. 1990. Top management
team tenure and organizational
outcomes: The mod-
erating role of managerial discretion. Administra-
tive Science Quarterly, 35: 484-503.
Finkelstein, S., & Hambrick,
D. C. 1996. Strategic lead-
ership: Top executives and their effects on organi-
zations. St Paul: West.
Ford, R. H. 1988. Outside directors and privately-held
firms: Are they necessary? Entrepreneurship The-
ory and Practice, 13(1): 49-57.
Geeraerts,
G. 1984. The effect of ownership on the or-
ganization structure in small firms. Administrative
Science Quarterly, 29: 232-237.
Geletkanycz,
M. A. 1997. The salience of "culture's con-
sequences": The effect of cultural values on top ex-
ecutive commitment to the status quo. Strategic
Management Journal, 18: 615-634.
Gersick,
K. E., Davis, J. A., Hampton,
M. M., &
Lansberg,
I. 1997. Generation to generation: Life cycles of the
family business. Boston: Harvard
Business School
Press.
Gomez-Mejia,
L., Nufiez-Nickel, M., &
Gutierrez,
I. 2001.
The role of family ties in agency contracts.
Academy
of Management Journal, 44: 81-95.
Gupta, S., &
Romano, R. E. 1998. Monitoring
the princi-
pal with multiple agents. Rand Journal of Econom-
ics, 29: 427-442.
Hambrick,
D. C., &
Fukutomi, G. D. S. 1991. The seasons
of a CEO's tenure. Academy of Management Re-
view, 16: 719-742.
Hambrick,
D. C., Geletkanycz,
M. A., &
Fredrickson,
J.
A.
1993. Top executive commitment to the status quo:
Some tests of its determinants. Strategic Manage-
ment Journal, 14: 401-418.
Handler,
W. C. 1989. Methodological issues and consid-
erations in studying family business. Family Busi-
ness Review, 1(4): 257-276.
Holtz-Eakin,
D., Joulfian, D., & Rosen, H. S. 1993. The
Carnegie conjecture: Some empirical evidence.
Quarterly Journal of Economics, 108: 413-435.
Hunter, J.
E., & Schmidt, F. L. 1990. Methods of meta-
analysis: Connecting error and bias in research
findings. Newbury Park,
CA: Sage.
Ilgen, D. R. 1986. Laboratory
research: A question of
when, not if. In E. A. Locke (Ed.),
Generalizing from
laboratory to field settings: 257-267. Lexington,
MA: Heath.
Jensen, M. C. 1998. Self-interest, altruism, incentives,
and agency. In M. C. Jensen (Ed.), Foundations of
organizational strategy: 39-50. Cambridge, MA:
Harvard
University Press.
Jensen, M. C., & Meckling, W. H. 1976. Theory of the
firm:
Managerial
behavior, agency costs, and owner-
ship structure.
Journal of Financial Economics, 3:
305-360.
Jensen, M. C., &
Smith, C. L. 1985. Stockholder,
manager
and creditor interests: Applications of agency the-
ory. In E. I. Altman & M. G. Subrahmanyam
(Eds.),
Recent advances in corporate finance: 95-131.
Homewood, IL:
Irwin.
Kang,
D. 1999. Ownership structure and corporate div-
idend policy: How large block family owners in-
crease dividend payout and achieve superior firm
performance. Harvard Business School working pa-
per number 99-032, Boston.
Koch, M. J., & McGrath,
R. G. 1996. Improving labor
productivity:
Human resource management policies
do matter.
Strategic Management Journal, 17: 335-
354.
Kraemer,
H. C., & Theimann, S. 1987. How many sub-
jects: Statistical power analysis in research. New-
bury Park,
CA: Sage.
Kroll, M., Wright, P., & Theerathorn, P. 1993. Whose
interests do top managers pursue? An examination
of acquisition performance
and CEO
rewards. Jour-
nal of Business Research, 26: 133-148.
La Porta, R., Lopez-de-Silanes, F., & Shleifer, A. 1999.
Corporate ownership around the world. Journal of
Finance, 54: 471-517.
Levinson, H. 1971. Conflicts that plague family business.
Harvard Business Review, 49(2): 90-98.
Lunati, M. T. 1997. Ethical issues in economics:
From altruism to cooperation to equity. London:
Macmillan.
MacDougall, P., & Robinson, R. B. 1990. New venture
strategies: An empirical identification of eight "ar-
chetypes" of competitive strategies
for entry. Strate-
gic Management Journal, 11: 447-468.
Meyer, M., & Zucker, L. G. 1989. Permanently failing
organizations. Newbury Park,
CA: Sage.
Morck, R., Shleifer, A., & Vishny, R. W. 1988. Manage-
ment ownership and market valuation:
An empirical
analysis. Journal of Financial Economics, 20: 293-
315.
Nash, J. M. 1988. Boards of privately-held companies:
Their responsibility and structure.
Family Business
Review, 1(3): 263-369.
Nayyar,
P. R. 1992. On the measurement of service diver-
sification strategy:
Evidence from large U.S. service
firms. Strategic Management Journal, 13: 219-235.
Nunnally, J. C. 1978. Psychometric theory. New York:
McGraw-Hill.
O'Donoghue, T., & Rabin, M. 2000. The economics of
2003 Schulze, Lubatkin,
and Dino 193
immediate gratification.
Journal of Behavioral De-
cision Making, 13: 233-250.
Ouchi, W. G. 1980. Markets, bureaucracies, and clans.
Administrative Science Quarterly, 25: 129-141.
Perrow, C. 1986. Complex organizations. New York:
Random House.
Rubenson,
G. C., &
Gupta,
A. K. 1996. The initial succes-
sion: A contingency model of founder tenure.
Entre-
preneurship Theory and Practice, 21(2): 21-35.
Sackett, P. R., Larsen,
J. R., Jr. 1990. Research strategies
and tactics in industrial organization
psychology. In
M. D. Dunnette &
L. M. Hough (Eds.),
Handbook of
industrial and organization psychology: 419-490.
Palo Alto, CA: Consulting Psychologists Press.
Schulze, W. S., Lubatkin,
M. H., Dino, R. N., &
Buchholtz,
A. K. 2001. Agency relationships in family firms:
Theory and evidence. Organization Science, 12:
99-116.
Small Business Administration. 2002. SBA loan pro-
grams. Document number 507. http://asbdc.ualr.
edu/fod/fo7.htm.
Stark,
O., &
Falk, I. 1998. Transfers,
empathy formation,
and reverse transfers. American Economic Review,
88: 271-276.
Tabachnick,
B. G., & Fidell, L. S. 1989. Using multivar-
iate statistics. Cambridge,
MA: Harper
& Row.
Tan, J. J., & Litschert, R. J. 1994. Environment-strategy
relationship and its performance implications: An
empirical study of the Chinese electronics industry.
Strategic Management Journal, 15: 1-20.
Thaler,
R. H., &
Shefrin, H. M. 1981. An economic theory
of self-control. Journal of Political Economy, 89:
392-406.
Walsh, J. P., & Seward, J. K. 1990. On the efficiency of
internal and external control mechanisms. Academy
of Management Review, 15: 421-458.
Ward,
J. L. 1987. Keeping the family business healthy:
How to plan for continuous growth, profitability,
and family leadership. San Francisco:
Jossey-Bass.
Ward, J. L., & Handy, J. L. 1988. A survey of board
practices. Family Business Review, 1(3): 298-308.
Whisler, T. L. 1988. The role of the board in threshold
firms. Family Business Review, 1(3): 309-321.
Wiseman,
R. M., &
Gomez-Mejia,
L. R. 1998. A behavioral
agency model of managerial
risk taking. Academy of
Management Review, 23: 133-153.
Wortman, M. S., Jr. 1994. Theoretical foundations for
family-owned business: A conceptual and research-
based paradigm. Family Business Review, 7(1):
3-27.
Wright, P., Ferris, S. P., Sarin, A., & Awasthi, V. 1996.
Impact of corporate insider, blockholder, and insti-
tutional equity ownership on firm risk taking.
Acad-
emy of Management Journal, 39: 441-465.
APPENDIX
Nine covariates
were included to reduce extraneous or
confounding variance. Firm size was total firm sales log-
arithmically
transformed to correct for its skewed distri-
bution. Firm age was calibrated in years. Age may be
linked to performance via a self-selection bias; older
firms existed in this sample simply because they were
successful. Multiple family ownership was a dummy
variable (0/1) that adjusted for the influence of multiple
families owning at least 15 percent of a firm's stock. We
also controlled for the number
offamily-member
employ-
ees since employment risk rises as families become de-
pendent on a firm for their livelihood. The mean number
of family employees for our sample was 3.51; the range
was 1-24. Exports as a percentage of sales was a five-
level indicator ranging from "zero" through "over 50
percent." In general, firms with export sales report a
higher level of indebtedness because they use bank let-
ters of credit and other types of debt instruments to
facilitate payment from international customers. Vari-
ance in performance and agency conditions linked to
information technology intensity was controlled by using
this item: "How important are investments in informa-
tion technology for the accomplishment of your future
goals?" (1 = "not important," 4 = "very important").
We also controlled for CEO
tenure, average
board ten-
ure, and board size. A large body of managerial
research
indicates that long CEO
tenure is generally detrimental to
firm performance
(Finkelstein & Hambrick,
1990). Indi-
vidual risk tolerance falls with age, and cognitive pro-
cesses rigidify. For example, Hambrick and Fukutomi
(1991) observed that as managers age, they tend to re-
ceive narrower and more filtered information, acquire
task knowledge more slowly, lose interest in routine
tasks as repetition leads to tedium, and increase their
commitment to the status quo (Hambrick
et al., 1993).
The negative effects of age and tenure on both cognitive
diversity and risk tolerance are exacerbated by group
processes (Finkelstein & Hambrick, 1996: 124-130). The
Andersen survey from which we drew data measured
CEO tenure with a five-level variable with responses
ranging from "11 or more years until retirement" to
"semiretired."
We found that this indirect measure of
tenure correlated with other indirect measures of tenure
available from the survey. For example, the bivariate
correlation between CEO tenure and CEO
age was high
(r = .62, p - .001), particularly given that scaling differ-
ences naturally deflated the correlation between the two
variables. Also, the mean age of the CEOs
(54 years) and
the mean age of the heirs-apparent at the time of this
designation (38 years) differed, as we expected, by about
one generation. Like Finkelstein and Hambrick (1990),
we measured average board tenure as the average of the
years members had served and controlled for variance in
the number of board members across firms with a count
indicating board size. The mean board size in this sample
was 3.98, and the standard deviation was 1.74. An item
that asked the respondents to rate the likelihood that
194 Academy of Management
Journal April
their families would retain control of the sampled firms
in the foreseeable future, family ownership goal, con-
trolled for variance in the strategic directions of these
family firms. Lastly, we measured ownership held by
board as the percentage
of a firm's shares held by mem-
bers of its board of directors.
William S. Schulze (Schulze@po.cwru.edu)
is an assis-
tant professor of entrepreneurship
and the H. R. Horvitz
Professor
in Family Business at the Weatherhead
School
of Management, Case Western Reserve University. He
earned his Ph.D. in strategic management
from the Uni-
versity of Colorado at Boulder. His research focuses on
entrepreneurship,
firm governance, risk, and managerial
applications of economic theory.
Michael Lubatkin is the John and Bette Wolff Family
Professor
of Strategic Entrepreneurship
at the University
of Connecticut and a professor at the Ecole de Manage-
ment de Lyon. He earned his Ph.D. in business adminis-
tration from the University of Tennessee. His research
focuses on firm governance, mergers and acquisitions,
corporate
diversification, international
management,
and
managerial applications of economic theory.
Richard N. Dino is an associate professor of manage-
ment, an associate dean of the School of Business, and
the chair of the Family Business Program
at the Univer-
sity of Connecticut. He earned his Ph.D. in economics
from the State University of New York at Buffalo.
His teaching and research focus on the management
and performance
of small to medium-sized, closely held
businesses.
