AN ANALYSIS OF THE PRINCIPAL-AGENT PROBLEM
IT HAS BEEN RECOGNIZED for some time that, in the presence of moral hazard, market allocations under uncertainty will not be unconstrained Pareto optimal (see Arrow , Pauly ). It is only relatively recently, however, that economists have begun to undertake a systematic analysis of the properties of the secondbest allocations which will arise under these conditions. Much of this analysis has been concerned with what has become known as the principal-agent problem. Consider two individuals who operate in an uncertain environment and for whom risk sharing is desirable. Suppose that one of the individuals (known as the agent) is to take an action which the other individual (known as the principal) cannot observe. Assume that this action affects the total amount of consumption or money which is available to be divided between the two individuals. In general, the action which is optimal for the agent will depend on the extent of risk sharing between the principal and the agent. The question is: What is the optimal degree of risk sharing, given this dependence?