Policy Evaluation in Behavioral Models

This paper studies the data requirements to conduct an unambiguous policy evaluation when consumers may exhibit nonstandard preferences. We theoretically analyze how the equilibrium combinations before and after a policy introduction allow to clearly assess different economic indicators. Specifically, we study the data requirements for an unambiguous evaluation of two distinct consumer "cooling off" policies in a model in which consumers may exhibit a projection bias when making consumption decisions. We find that market-level data can help to make simple policy evaluations using baseline statistics and sometimes even discriminate between different classes of consumer preferences. We further discuss the generalizability to other models and the use for ex-ante policy design.