From unobserved to observed preference heterogeneity: a revealed preference methodology

Abstract

We present an easy-to-apply non-parametric revealed preference method to identify observed preference heterogeneity from cross-sectional data. Building on the partitioning approach that was developed by Crawford and Pendakur (Economic Journal, 2013, 123(567), 77–95) and Cosaert (Computational Economics, 2019, 53(2), 533–54), it quantifies the contribution of observable consumer characteristics to describing the identified preference heterogeneity. We demonstrate the practical usefulness of our method through an application to newly gathered experimental data on consumer choice behaviour in two types of decision situations: the allocation of money (choosing between two products) and the allocation of time (choosing between leisure and work). We investigate whether the same consumer characteristics drive the observed variation in choice behaviour in these two settings.