This paper considers structural nonparametric random utility models for continuouschoice variables. It provides suffcient conditions on random preferences to yield reduced-form systems of nonparametric stochastic demand functions that allow global invertibilitybetween demands and random utility components. Invertibility is essential for globalidentification of structural consumer demand models, for the existence of well-specifiedprobability models of choice and for the nonparametric analysis of revealed stochasticpreference.
Centre for microdata methods and practice