A nonparametric multiple choice method within the random utility framework

Many researchers use categorical data analysis to recover individual consum ption preferences, but the standard discrete choice models require restrict ive assumptions. To improve the flexibility of discrete choice data analysi s, we propose a nonparametric multiple choice model that applies the penali zed likelihood method within the random utility framework. We show that the deterministic component of the random utility function in the model is a c ubic smoothing spline function. The method subsumes the conventional condit ional legit model (McFadden, 1973, in: Zarembka, P., (Ed.), Frontiers in Ec onometrics) as a special case. In this paper, we present the model, describ e the estimator, provide the computational algorithm of the model, and demo nstrate the model by applying it to nonmarket valuation of recreation sites . (C) 2000 Elsevier Science S.A. All rights reserved. JEL classification. C 14.