Munich Personal RePEc Archive

A simple axiomatics of dynamic play in repeated games

Mathevet, Laurent (2012): A simple axiomatics of dynamic play in repeated games.

[img]
Preview
PDF
MPRA_paper_36031.pdf

Download (556Kb) | Preview

Abstract

This paper proposes an axiomatic approach to study two-player infinitely repeated games. A solution is a correspondence that maps the set of stage games into the set of infinite sequences of action profiles. We suggest that a solution should satisfy two simple axioms: individual rationality and collective intelligence. The paper has three main results. First, we provide a classification of all repeated games into families, based on the strength of the requirement imposed by the axiom of collective intelligence. Second, we characterize our solution as well as the solution payoffs in all repeated games. We illustrate our characterizations on several games for which we compare our solution payoffs to the equilibrium payoff set of Abreu and Rubinstein (1988). At last, we develop two models of players' behavior that satisfy our axioms. The first model is a refinement of subgame-perfection, known as renegotiation proofness, and the second is an aspiration-based learning model.

UB_LMU-Logo
MPRA is a RePEc service hosted by
the Munich University Library in Germany.