Efficiency and renegotiation in repeated games

In this paper we present a general framework for the discussion of renegotiation in repeated games, provide a new concept of “renegotiation-proof“ equilibrium, and show how the framework clarifies and unifies existing work. We define an agreement as a sequence of payoffs sets, one for each history in the game. We adopt an axiomatic approach to restricting agreements and show how different restrictions an agreements lead to different concepts of renegotiation-proofness. This provides a synthesis of the existing literature. Also, in this framework we introduce a new concept of renegotiation-proofness called recursive efficiency.  Journal of Economic Literature Classification Numbers: C70, C72, D70.