Endogeneity and Individual Consumer Choice
Published 12/1/2008
Author: Dmitri Kuksov and J. Miguel Villas-Boas
View this content
Executive Summary
This article presents a test for the endogeneity bias in panel data when unobserved consumer heterogeneity is also potentially present. It also presents a parsimonious and robust method of unbiased estimation of the market response in the presence of both endogeneity and unobserved consumer heterogeneity. For example, the issue of endogeneity arises if the following two conditions are present in the marketplace: (1) when the variation of consumer utility of a product is correlated across consumers (i.e., when consumers receive common demand shocks from time to time) and (2) when firms observe, at least partially, some of these shocks and use these observations when deciding on prices and other variables that may affect demand. The first condition may hold, for example, because of common taste trends, advertising or coupon availability, changes of prices of competing goods, or changes in the economic outlook or weather. If these shocks are not observed by the researcher or cannot be easily incorporated into the econometric estimation, some of the effects of these demand shocks may be misinterpreted as the effects of the variables set by the firms. Thus, when the endogeneity issue is present, it may bias the estimation of the parameters of the market response to the endogenous variables. In addition to these two conditions, if individual consumer preferences are stable (i.e., they change less through time than from an individual at one time to a different individual at a different time) and the researcher cannot fully account for this stability through observed household characteristics, there is also a question whether such unobserved consumer heterogeneity will bias the estimation and how to correct for the possible endogeneity bias in the presence of this unobserved consumer heterogeneity.
The estimation method presented in this article is a two-step method based on the method of simulated moments, which allows for a simultaneous correction of the endogeneity bias in several marketing variables and can be realistically applied to choice sets with a relatively large number of alternatives. The authors theoretically show that when endogeneity is corrected for through this method, unobserved heterogeneity does not asymptotically bias the parameter estimates, though it may reduce the precision of the estimates; however, the random coefficients that result in the same variance between observations must still be allowed (i.e., it may be important to include the time-varying random coefficient terms). The authors then test the final sample properties of the estimation method through a Monte Carlo simulation study and apply the method to a panel set with three potentially endogenous marketing variables (price, display, and feature); they consider seven choice alternatives that each consumer faces at each purchase occasion.
Biography
Dmitri Kuksov is Assistant Professor of Marketing in the Olin Business School at Washington University in Saint Louis. He holds a PhD in Business Administration from the Haas School of Business at the University of California, Berkeley, a PhD in Mathematics from Brigham Young University, and a BS equivalent in Mathematics from Moscow State University. His research interests include competitive strategy, incomplete information and consumer search, and product design and branding. His research on these topics ahs appeared in Marketing Science, Management Science, and Journal of Economic Theory. He is on the editorial board of Marketing Science and Review of Marketing Science.
J. Miguel Villas-Boas is the director of the PhD Program and J. Gary Shansby Professor of Marketing Strategy in the Haas School of Business at the University of California, Berkeley. He holds a Licenciatura in Economics from Universidade Católica Portuguesa, an MBA from INSEAD, an MSc in Economics from the New University of Lisbon, and a PhD in Applied Economics from the Massachusetts Institute of Technology. His research interests include competitive strategy, design of marketing organizations, customer relationship management, customer recognition, and product line design. His research has appeared in several journals, including Marketing Science, Management Science, Journal of Economic Theory, RAND Journal of Economics, Journal of Marketing Research, and Journal of Economics and Management Strategy. He is an area editor at Marketing Science and an associate editor at Quantitative Marketing and Economics and serves on the editorial boards of Journal of Marketing Research and International Journal of Marketing Research.
J Marketing Research, Volume 45, Number 6, December 2008
View Table of Contents.