Stochastic Demand and Revealed Preference

Abstract

This paper develops new techniques for the estimation and testing of stochastic consumer demand models. For general non-additive stochastic demand functions, we demonstrate how preference inequality restrictions can be utilized to improve on the nonparametric estimation and testing of demand responses. We show how bounds on demands can be estimated non- parametrically, and derive their asymptotic properties utilizing recent results on estimation and testing of parameters characterized by moment inequalities. We also devise a test for rationality of the consumers. An empirical application using data from the U.K. Family Expenditure Survey illustrates the usefulness of the methods.

JEL codes