Robust Bayesian choice.

A major concern with Bayesian decision making under uncertainty is the use of a single probability measure to quantify all relevant uncertainty. This paper studies prior robustness as a form of continuity of the value of a decision problem. I show that this notion of robustness is characterized by a form of stable choice over a sequence of perturbed decision problems, in which the available acts are perturbed in a precise fashion. I then introduce a choice-based measure of prior robustness and apply it to models of climate mitigation and portfolio choice. • Belief Robustness and Continuity: Beliefs are "robust" if minor changes in beliefs lead to minor changes in choices. This notion can be tested by evaluating perturbations of feasible bets. • Measure of Robustness: A comparative measure of belief robustness can quantify the sensitivity of optimal choices to parametric assumptions. • Impact on Key Economic Models: The measure shows that heavy-tailed distributions yield more robust choices in portfolio and economy-climate models, highlighting its practical importance. [ABSTRACT FROM AUTHOR]