Identification of Average Demand Models

• Departmental Seminar

Date and Location

Monday, January 23, 2017, 10:30 AM - 11:50 AM
ARE Library Conference Room, 4101 Social Sciences and Humanities

Abstract

This paper studies the nonparametric identification of a model of average demand for multiple goods, once unobservable heterogeneity has been integrated out. The model can be used for bundles, decisions under uncertainty, stochastic choice, and other examples. Optimizing behavior implies an analogue of Slutsky symmetry, which we exploit to show nonparametric identification of the model. Our main results do not rely on special regressors or identification at infinity. As a special case we provide new conditions for identification of additive random utility models (ARUM). These conditions also apply to a stochastic choice model allowing bounded rationality. In an illustrative application, we refute ARUM in favor of this more general model.