Optimal Managerial Contracts and Equilibrium Security Prices

RECENT DEVELOPMENTS IN ECONOMIC theory have focused attention on efficient methods for providing incentives within a corporation, especially to top management. Jensen-Meckling [12] have examined some implications of incentive (or agency) problems for corporate finance. In this paper we combine a simple model of security market equilibrium with that of an optimal managerial incentive contract, using results of Harris-Raviv [8] and Holmstrom [9]. We provide a rationale for including security prices in the managerial incentive contracts of publicly held firms. In this model we also show that non-systematic, diversifiable risks will be considered in the capital budgeting decisions of a managerial firm operated in the interests of its stockholders. This is in contrast to the implication of the Arbitrage Pricing Theory (a similar result follows from the Capital Asset Pricing Model) which abstracts from incentive problems and suggests that only systematic risks ought to be considered. The optimal-contract model of the corporation takes the following view of the firm. Since decision makers within a firm are not sole owners of the business, but employed agents, the decisions which they make depend on the incentives which the organization provides. These incentives are embodied explicitly in employment contracts, and implicitly through the threat of dismissal or effect on reputation. The contract, whether explicit or not, describes the decision maker's anticipation of the consequences to him conditional on various observable outcomes. These outcomes are used to provide information useful in determining whether he did his job properly. In the simplest optimal contract scenario, there is a single principal and a single agent whose decision concerns the level of effort he expends. The level of effort affects the level of output of the firm, but not unambiguously because the output is also governed by other stochastic elements. As a result, some information about the agent's level of effort can be inferred