Learning in a black box

Abstract

We study behavior in repeated interactions when agents have no information about thestructure of the underlying game and they cannot observe other agents’ actions or payoffs.Theory shows that even when players have no such information, there are simple payoff-based learning rules that lead to Nash equilibrium in many types of games. A key feature ofthese rules is that subjects search differently depending on whether their payoffs increase,stay constant or decrease. This paper analyzes learning behavior in a laboratory setting andfinds strong confirmation for these asymmetric search behaviors in the context of voluntarycontribution games. By varying the amount of information we show that these behaviorsare also present even when subjects have full information about the game.