Ch. 8: Conglomerate Firms and Internal Capital Markets 443
likely outcomes of bargaining over the surplus and make investment choices accord-
ingly. It is in the top management’s interest to allocate initial investment funds in ways
that induce the divisional managers to choose projects that maximize the firm’s value.
Hence, in RSZ top management uses the initial allocation of investment to divisions
as a commitment device to substitute for its inability to commit to a distribution of
surplus. This form commitment is clearly not as efficient as a first-best case in which
top management can commit to the distribution of profits that the divisions realize.
Empirically, the capital expenditures of conglomerates might seem, and would be, less
efficient than those of single-segment firms. However, they are value maximizing given
the constraints that top managers face.
RSZ make specific assumptions about the way the bargaining between divisions
works and obtain predictions about the distortions that arise. Specifically they assume
that each division can choose to invest in two types of investment projects. “Efficient”
projects are value maximizing. “Defensive” projects produce less value, but the value
generated can be better defended against redistribution to other divisions. The top man-
agement’s problem is to allocate the right amount of capital to each division and to
motivate the divisional management to invest the capital in the efficient project.
The divisional manager’s incentive to choose a defensive project is higher when the
surplus generated by the efficient project, which he has to partially give up in ex-post
bargaining with other divisions, is high relative to the manager’s share of the other
divisions’ surplus that he expects to gain in bargaining. Under plausible assumptions,
this occurs when the manager’s division has better investment opportunities than the
other divisions. As a result, perverse investment incentives are more likely to occur in
firms with divisions facing diverse investment opportunities.
The RSZ model predicts that the value of diversified firms is inversely related to the
diversity in their investment opportunities. The model also predicts that capital transfers
will occur from large high-value divisions to small low-value divisions. Both of these
predictions are testable. We discuss these tests later.
A central feature of most theoretical models of the conglomerate firm is that they
are partial equilibrium in the sense that they do not analyze the firm’s internal allo-
cation of capital in the context of the market for whole firms and partial-firm assets.
AsMaksimovic and Phillips (2001)show, there is a large market for assets in which
conglomerates are important players. Thus, as an alternative to distorting the firm’s
investment expenditures, a firm facing the problems modeled by RSZ might trade
divisions to obtain a diversified portfolio of assets that faces comparable investment
opportunities. Thus, a generalized RSZ framework might suggest that the firm can op-
erate on an alternative margin, yielding the prediction that at times when the market for
firms’ assets is active, firms are less likely to distort investment flows.
3.5. Neoclassical modelof conglomerates and resource allocation
The case in which firms maximize value and there are no unresolved agency prob-
lems provides a benchmark for an analysis of conglomerate growth and diversification.