Revealed preferences over risk and uncertainty

We develop a nonparametric procedure, called the lattice method, for testing the consistency of contingent consumption data with a broad class of models of choice under risk and under uncertainty. Our method allows for risk loving and elation seeking behavior and can be used to calculate, via Afriat's efficiency index, the magnitude of violations from a particular model of choice. We evaluate the performance of different models (including expected utility, disappointment aversion, rank dependent utility, mean-variance utility, and stochastically monotone utility) in the data collected by Choi et al. (2007), in terms of pass rates, power, and predictive success.