Intertemporal Consumption with Risk: A Revealed Preference Analysis

We run an experiment designed to elicit preferences over state contingent, timed payouts. We analyze the data using a new revealed preference method (building on Nishimura, Ok, and Quah (2017)) that can test for consistency with utility functions that increase with a given preorder. Using this approach, we find strong evidence of correlation averse behavior, a property ruled out by discounted expected utility. We also find evidence in favor of stochastic impatience.