| This dissertation is a theoretical study of the role of beliefs in both individual and group behavior. In Chapter 1, we consider a decision-maker who chooses from a set of options after receiving some private information. This information however is unobserved by an analyst, so from the latter's perspective, choice is probabilistic or random. We provide a theory in which information can be fully identified from random choice. In addition, the analyst can perform the following inferences even when information is unobservable: (1) directly compute ex-ante valuations of option sets from random choice and vice-versa, (2) assess which decision-maker has better information by using choice dispersion as a measure of informativeness, (3) determine if the decision-maker's beliefs about information are dynamically consistent, and (4) test to see if these beliefs are well-calibrated or rational.;In Chapter 2, we dispense with the standard assumption of expected utility maximization and introduce a theory of stochastic ambiguity aversion. In the individual interpretation of this theory, choice is random due to unobservable shocks to the individual's ambiguity aversion. In the group interpretation of this theory, choice is random due to unobservable heterogeneity in ambiguity aversion within the group. A one-parameter distribution characterizing stochastic ambiguity aversion can be fully identified from random choice. From a technical standpoint, we offer decision theoretic foundations for relaxing linearity under random utility maximization.;In Chapter 3, we study a model of competitive trading where agents have heterogeneous beliefs about the persistence of states. This model addresses a robust finding in behavioral finance known as the disposition effect where agents over-purchase stocks after prices fall and over-sell stocks after prices rise. We show that agents who believe in the least persistence exhibit the disposition effect while those who believe in the most persistence engage in momentum trading. Moreover, agents can be ordered by how much of a disposition effect they exhibit if and only if their beliefs can be ordered by a single parameter measuring persistence. This allows for identification of beliefs even when equilibrium considerations restrict the observable choice data. |
