Behavioral Foundations of Nested Stochastic Choice and Nested Logit

Abstract

We provide the first behavioral characterization of nested logit, a foundational and widely applied model in discrete choice. We take a revealed preference approach to identify the underlying notion of similarity and use it to characterize a non-parametric version of nested logit we call Nested Stochastic Choice(NSC). We characterize NSC with a single axiom that weakens Independence of Irrelevant Alternatives(IIA) based on revealed similarity to allow for the similarity effect. Nested logit is then characterized by one additional axiom that imposes a weak form of menu-independence. We show that a widely applied generalization, cross-nested logit, lacks testable implications.

Keywords: Nested Stochastic Choice, Luce Model, Nested Logit, IIA, Similarity Effect, Regularity, Revealed Similarity

JEL Classification: D01, D81, D9