Economists have developed a wide variety of models to predict how economic agents will interact strategically over time. However, many of these models suffer from a lack of predictive power because multiple equilibria are predicted by the standard equilibrium definitions used in game theory. Recent research shows that in some special classes of games, adding a richer set of possible behavior to the game allows reputations to form that dictate a single outcome. This project will apply these ideas to a large and important class of games, those that model bargaining behavior.
The project will place particular emphasis on understanding what happens when bargainers are not fully informed about some of the opponent's goals or payoffs. The goal is to develop stronger and more robust predictions in this central area of economic theory.
Broader implications include the ability to conduct much stronger empirical tests of economic models. Once tested, the models will allow economists to give much better advice to policy-makers and others who use bargaining to reach agreements between nations, interest groups, firms, and individuals.
