Econometrica: Jul, 2024, Volume 92, Issue 4
Nonparametric Identification of Differentiated Products Demand Using Micro Data
https://doi.org/10.3982/ECTA20731
p. 1135-1162
Steven T. Berry, Philip A. Haile
We examine identification of differentiated products demand when one has “micro data” linking the characteristics and choices of individual consumers. Our model nests standard specifications featuring rich observed and unobserved consumer heterogeneity as well as product/market‐level unobservables that introduce the problem of econometric endogeneity. Previous work establishes identification of such models using market‐level data and instruments for all prices and quantities. Micro data provides a panel structure that facilitates richer demand specifications and reduces requirements on both the number and types of instrumental variables. We address identification of demand in the standard case in which nonprice product characteristics are assumed exogenous, but also cover identification of demand elasticities and other key features when these product characteristics are endogenous and not instrumented. We discuss implications of these results for applied work.
Supplemental Material
Supplement to "Nonparametric Identification of Differentiated Products Demand Using Micro Data"
Steven T. Berry and Philip A. Haile
In this Appendix we explore variations on our baseline model and associated identification conditions. We show that identification is robust in the sense that a relaxation of one condition assumed in the text can often be accommodated by strengthening another. An understanding of such trade-offs can be helpful to both producers and consumers of research relying on demand estimates. Although a full exploration of these trade-offs describes an entire research agenda, we illustrate some possibilities that relax key restrictions of our model, allow demand systems outside discrete choice settings, enlarge the set of potential instruments, reduce the number of required instruments, eliminate the need for continuous consumer-level observables, or reduce the required dimension of those observables. For simplicity, we focus here on the traditional case in which X
t is exogenous, recalling that in this case we have
h(X
t Ξ
t)= Ξ
t .
View pdf