Ulric B. and Evelyn L. Bray Social Sciences Seminar
Mixed Logit and Pure Characteristics Models
Abstract: Mixed logit or random coefficients logit models are used extensively in empirical work while pure characteristic models feature in much of theoretical work. We provide a theoretical analysis comparing and contrasting the two classes of models. First, we show an approximation theorem that precisely characterizes the extent to which mixed logit models can approximate pure characteristic models. In the process, we introduce a general class of models that corresponds exactly to the closure of logit models. We then present two conditions that highlight behavioral differences between mixed logit and pure characteristic models. Both pertain to choice patterns relating to product differentiation. The first is a substitutability condition that is satisfied by many pure characteristic models (including the Hotelling model of horizontal differentiation) but is violated by almost all mixed logit models. The second is a continuity condition that is satisfied by all pure characteristic models but is violated by all mixed logit models.
Professors Lu and Saito will be joined by guests Yusufcan Masatlioglu, Giovanni Compiani, Simone Cerreia Vioglio, and Tomasz Strzalecki.
Contact: Letty Diaz firstname.lastname@example.org