Contracting with Third Parties

odels of bilateral contracting, such as the canonical holdup model, typically assume that third parties are not available. Given this assumption, the equilibrium outcome is not first-best efficient if contracts can be renegotiated. To be more specific, suppose an architect and a builder must cooperate to build a building. The quality of the building will depend on three things: the quality of the architect's design, the builder's skill, and a stochastic shock. We will refer to these three variables as the state of the world. The architect and the builder know the true state but no one else does. Thus, the state is " observable but not verifiable. " (An outsider may be able to judge the quality of the building after it is built, but he cannot disentangle the various contributions to it.) Suppose the contract specifies ex post transfers as a function of announcements made by the architect and the builder (a " message game "). If both report the state truthfully, the transfers will reflect the contributions of each party and provide correct incentives to invest in the transaction. In order to support an equilibrium where both tell the truth, however, it may be necessary to punish both of them if they disagree. In the absence of a third party, this punishment typically involves an ex post inefficiency (say, the destruction of the building). If such outcomes can be renegotiated, then it may be impossible to implement the first best. This holdup problem underlies recent work on the " foundations of incomplete contracts " (see Yeon-Koo Che and Donald B. Hausch 1999 and Ilya Segal 1999). If we introduce a third party (who may not know the true state), ex post renegotia-tion becomes less of a problem. The punishment can now consist of fines paid to the third party who acts as a " budget breaker " (as suggested by Bengt R. Holmström 1982). Since a fine is simply a transfer from one person to another, renegotiation is not an issue. This suggests that models that rely on renegotiation to generate holdup In bilateral holdup and moral hazard in teams models, introducing a third party allows implementation of the first best, even if renego-tiation is possible. Fines paid to the third party provide incentives for truth-telling and investment. This result holds even if the third party is corruptible, as long as the grand coalition has access to …