The players in most economically important games are agents, not principals. For example, almost all significant firms are run by professional managers who are not owners of the firms that they direct, or who own only some small fraction of the outstanding shares of the corporations. Bargaining games, too, typically involve agents as players. The bargaining on the selling side is conducted by an agent, e.g., an automobile salesman. Moreover, most transactions in intermediate goods markets involve agents on both sides simultaneously, e.g., the purchasing agent of one company bargaining with the salesman of a supplier. Game theory has been widely applied to economic problems under the assumption that the games are played among principals. This project examines a central, and often more realistic, class of two-stage games in which some or all of the principals hire agents in the first stage and these agents (and any principals who do not hire agents) then play a second-stage game. It is well understood that if: (1) only one principal has access to the use of an agent; (2) the agency contract is observable to the other players in the game; and (3) it is common knowledge that the contract cannot be renegotiated, then the contract can serve as a form of binding precommitment that induces the agent to act as a Stackelberg leader in the second-stage game. The effects of agency when any of these three conditions is not met are much less well understood. The proposed research is intended to shed light on the effects of agency when these three conditions are not satisfied. A variety of applications will be considered, including oligopoly, vertical restraints, international trade policy, and the interrelationship between financial structure and product-market competition.
