A Model of Search with Two Stages of Information Acquisition and Additive Learning

This paper presents a model of choice with two stages of information acquisition. In this model, the choice problem can be interpreted as a variant of a more general multiarmed bandit problem. We assume that information acquisition takes a simple “additive form”—the value of an alternative is the sum of two components, which the decision maker can learn by undertaking two stages of information acquisition. This assumption yields a model that is tractable for the purposes of structural estimation. One possible application of the model is to online purchasing on e-commerce sites. For a consumer on an e-commerce website, there are potentially two stages of information acquisition: the consumer can obtain information about an alternative from (i) browsing the search results page and (ii) clicking on the alternative. By way of contrast, in much of the literature on structural econometric models of online purchasing, there is typically only one stage of information acquisition. Our paper may, therefore, provide a more realistic theory for modeling search, at least for those types of search—such as online purchasing—that involve two stages of information acquisition. This paper was accepted by Manel Baucells, behavioral economics and decision analysis.

Book Reviews

Benjamin Williams of George Washington University reviews “Structural Econometric Models”, by Eugene Choo and Matthew Shum. The Econlit abstract of this book begins: “Twelve papers explore recent developments in the use of structural econometric models in empirical economics. Papers discuss Euler equations for the estimation of dynamic discrete choice structural models; approximating high-dimensional dynamic models—sieve value function iteration; identifying dynamic games with serially correlated unobservables; partial identification in two-sided matching models; identification of matching complementarities—a geometric viewpoint; comparative static and computational methods for an empirical one-to-one transferable utility matching model; a test for monotone comparative statics; estimating supermodular games using rationalizable strategies; estimation of the loan spread equation with endogenous bank-firm matching; the collective marriage matching model— identification, estimation, and testing; deflation in durable goods markets—an empirical model of the Tokyo condominium market; and a dynamic analysis of the U.S. cigarette market and antismoking policies. Choo is with the Department of Economics at the University of Calgary. Shum is with the Division of Humanities and Social Sciences at the California Institute of Technology.”