Consumer Welfare under Individual Heterogeneity

Welfare effects of price changes are often estimated with cross-sections; however, such data do not identify demand models with heterogeneous consumers. We develop a method that leverages moments of demand to construct a local approximation to the curvature of the expenditure function, which is essential for welfare analysis. Our method applies to arbitrary cross-sections and delivers a nonparametric approximation to the entire distribution of consumer welfare, robust to unobserved preference heterogeneity. We provide moment- and quantile-based estimation strategies that are simple to implement.