Efficiency and Renegotiation in Repeated Games.

This paper introduces a general framework for the discussion of renegotiation in repeated games, provides a new concept of "renegotiation proof" equilibrium, and shows how this model clarifies and unifies existing work in this area. The procedure involves restricting axiomatically the class of agreements which may be considered admissible, thus taking incentives into consideration at the beginning of the game when an agreement is selected. We define a preference ordering on agreements, so that given the set of agreements which satisfy the axioms imposed, we may select a "best" agreement. We use this framework to introduce a new concept of renegotiation proofness -- recursive efficiency -- and apply it to the efficiency wage model.