Logit demand estimation under competitive pricing behavior: An equilibrium framework

David Besanko*, Sachin Gupta, Dipak Jain

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

172 Scopus citations

Abstract

Discrete choice models of demand have typically been estimated assuming that prices are exogenous. Since unobservable (to the researcher) product attributes, such as coupon availability, may impact consumer utility as well as price setting by firms, we treat prices as endogenous. Specifically, prices are assumed to be the equilibrium outcomes of Nash competition among manufacturers and retailers. To empirically validate the assumptions, we estimate logit demand systems jointly with equilibrium pricing equations for two product categories using retail scanner data and cost data on factor prices. In each category, we find statistical evidence of price endogeneity. We also find that the estimates of the price response parameter and the brand-specific constants are generally biased downward when the endogeneity of prices is ignored. Our framework provides explicit estimates of the value created by a brand, i.e., the difference between consumers' willingness to pay for a brand and its cost of production. We develop theoretical propositions about the relationship between value creation and competitive advantage for logit demand systems and use our empirical results to illustrate how firms use alternative value creation strategies to accomplish competitive advantage.

Original languageEnglish (US)
Pages (from-to)1533-1547
Number of pages15
JournalManagement Science
Volume44
Issue number11 PART 1
StatePublished - Nov 1 1998

Keywords

  • Competitive Strategy
  • Demand Estimation
  • Endogeneity
  • Logit

ASJC Scopus subject areas

  • Strategy and Management
  • Management Science and Operations Research

Fingerprint Dive into the research topics of 'Logit demand estimation under competitive pricing behavior: An equilibrium framework'. Together they form a unique fingerprint.

Cite this