Essays in Economic TheoryPublic Deposited
This dissertation comprises three chapters, each focusing on a different question in economic theory. The first two chapters focus on repeated games and reputations, while the third is about large games. In "Cooperation and Community Responsibility", I study whether cooperation can be sustained between communities where members interact repeatedly but with different people in the community. When communities are large and players change rivals over time, players may not recognize each other or may have limited information about past play. Can players cooperate in such anonymous transactions? I analyze an infinitely repeated random matching game, where payers' identities are unobservable and players only observe their own matches. Players may send an unverifiable message (a name) before playing each game. I show that for any such game, all feasible individually rational payoffs can be sustained in equilibrium if players are sufficiently patient. In "Observability and Sorting in a Market for Names" I study the value of reputations. I ask whether firm names can be tradeable assets when changes in name ownership are observable? Earlier literature suggests that non-observability is critical to tradeable names. Yet, casual empiricism suggests that shifts in name ownership are often observable. I show how firm names can be traded under full observability. In equilibrium, even when consumers see a reputed name being divested they continue trusting it and so, these names are tradeable. I also demonstrate an appealing "sorting" property. Competent firms separate themselves by buying valuable names, or only incompetent firms use worthless names. In "Large Games of Limited Individual Impact" (co-authored with Ehud Kalai), I study robustness properties of equilibria. Bayesian Nash equilibria that are not ex-post stable are a poor modeling tool for many applications. Earlier literature showed that Bayesian equilibria are ex-post stable in games with a large number of anonymous players, with finite types and actions and continuous payoff functions. These assumptions limit the applicability of the results in important games like market games, location games etc. We identify a broad class of large games that satisfies ex-post stability, without requiring finiteness or anonymity. We show that one regularity condition on payoff functions (a version of Lipschitz continuity) can guarantee ex-post stability.