440 V. Maksimovic and G. Phillips
Diversification may also yield concrete career benefits, because experience running
a complex diversified firm might provide experience that the increases the value of the
manager’s future employment prospects (Gibbons and Murphy, 1992). On the other side
of the coin, diversification may entrench the manager because it may be harder to find
a replacement who has a demonstrated ability in managing the firm’s particular mix of
businesses (Shleifer and Vishny, 1989).
Taken together, the literature on agency makes a powerful prima fasciae case that
agency conflicts may drive unprofitable diversification. An issue in determining the
extent to which this is the case is that most of the contributions are set in a partial
equilibrium framework. Thus, it is not clear why the incentives are not set in ways
that penalize unprofitable diversification. Moreover, it is not clear why diversification is
inefficient. A rational empire-building CEO of a diversified firm can in principle decen-
tralize its operations and provide incentives to the managers running the firm’s divisions
so that firm value is not destroyed. Thus, it must either be the case that increasing the
firm’s scope the firm itself destroys value or that managers of firms that diversify are
irrational and have a hubristic belief in their ability to run acquired businesses (Roll,
1986 ).
There have been only a few attempts to analyze the manager’s incentive to diversify
in a more general model of the trade-offs. This is in part because the incentives of,
and the constraints faced by, the board of directors, the party that formally employs the
manager, are not well understood.^19
Aggarwal and Samwick (2003)model the diversification process by assuming that
the board maximizes the value of the firm. The key assumption is that diversification,
which is assumed to be value destroying, is not contractible and cannot be forbidden
by the board. The CEO benefits from diversification, because it enables him to di-
versify his risk and because he has private gains from diversification. The board can
attempt to motivate the manager to work harder by tying his compensation to firm
value. However, this type of compensation has a byproduct of increasing the manager’s
risk exposure, making value destroying diversification more attractive. In equilibrium,
managerial compensation is set as a result of contracting in a standard principal agent
problem where managerial effort is costly.
InAggarwal and Samwick (2003)the manager’s compensationwis given byw=
w 0 +απ+γn, whereαandγare constants chosen by the firm,nis the amount of
diversification andπis the firm value. Firm value is given byπ =x−n+ε(n),
wherexis the costly managerial effort andε(n)is a normally distributed shock to firm
value with zero mean and varianceσ^2 /n. For the manager, diversification has three
consequences. First, it affects the value of the corporation and thereby the manager’s
compensation through theαterm. Second, it enables the manager to diversify risks since
it reduces the risk of the corporationσ^2 /n. Third, diversification enters directly in to the
(^19) SeeHermalin and Weisbach (1998)for a theoretical model in which the relationship between the board of
directors and the CEO evolves over time.