Title:Tests of Bayesian Rationality

Abstract:What are the testable restrictions imposed on the dynamics of an agent's belief by the hypothesis of Bayesian rationality, which do not rely on the additional assumption that the agent has an objectively correct prior? In this paper, I argue that there are essentially no such restrictions. I consider an agent who chooses a sequence of actions and an econometrician who observes the agent's actions but not her signals and is interested in testing the hypothesis that the agent is Bayesian. I argue that -- absent a priori knowledge on the part of the econometrician on the set of models considered by the agent -- there are almost no observations that would lead the econometrician to conclude that the agent is not Bayesian. This result holds even if the set of actions is sufficiently rich that the agent's action fully reveals her belief about the payoff-relevant state and even if the econometrician observes a large number of identical agents facing the same sequence of decision problems.