Identification of random coefficient latent utility models

This paper provides nonparametric identification results for random coefficient distributions in perturbed utility models. We cover discrete choice and models of multiple purchases. We establish identification using variation in mean quantities. The results apply even when an analyst observes only aggregate demands but not whether goods are chosen together. The identification results use exclusion restrictions on covariates and independence between random slope coefficients and random intercepts. We do not require regressors to have large supports or parametric assumptions. We use an illustrative application to interpret the general identification lessons in a specific context.