Estimating Price Elasticities with Theory-Based Priors
Author(s): Alan L. Montgomery and Peter E. Rossi
Source: Journal of Marketing Research, Vol. 36, No. 4, (Nov., 1999), pp. 413-423 Published by: American Marketing Association
Stable URL: http://www.jstor.org/stable/3151997
Accessed: 22/07/2008 16:25
Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use. Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at http://www.jstor.org/action/showPublisher?publisherCode=ama. Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission.
JSTOR is a not-for-profit organization founded in 1995 to build trusted digital archives for scholarship. We work with the scholarly community to preserve their work and the materials they rely upon, and to build a common research platform that promotes the discovery and use of these resources. For more information about JSTOR, please contact firstname.lastname@example.org.
ALAN MONTGOMERY PETER ROSSI*
The authors show how price elasticity estimates can be improved in demand systems that involve multiple brands and stores. They treat these demand models in a hierarchical Bayesian framework. Unlike in more standard Bayesian hierarchical treatments, the authors use prior information based on the restrictions imposed by additive utilitymodels. In an additive utilitymodel approach, price elasticities are driven by a general substitution parameter as well as brand-specific expenditure elasticities. The authors employ a differential shrinkage approach in which price elasticities are held closely to the restrictions of the additive utility theory and store-to-store variation is accommodated through differences in expenditure elasticities. Application of these new methods to simulated and real store scanner data shows significant improvements over existing Bayesian and non-Bayesian methods.
Detailed scanner data on sales response and pricing conditions are now available for virtually every type of consumer packaged good and all major retail formats. The availability of these data permits a systematic approachto
studying marketstructureor brandcompetitionpatterns(cf.
Allenby 1989), elasticity-based approachesto optimal pricIn
ing, and evaluation of promotionalprofitability. addition,
there is increased interest in using disaggregatedata for the purpose of micropricing, in which individual stores or
groups of stores charge different prices to exploit differences in consumerprice sensitivity (Montgomery1997). All these analyses rely on the estimation of a demand system
and associated price elasticities.
The promise of a quantitativedemand-basedapproachto
pricing issues is far from fully being realized because of the difficulty of obtaining reasonableprice elasticity estimates. Unrestrictedleast squaresestimates of own- and cross-price
elasticities are often of an incorrectsign and unreasonable
magnitude,particularlyif the analysis is performedat a relatively low level of aggregation,such as the accountor store level. One possible solution is to engage in a lengthy specification analysis and data cleanup, which is not guaranteed to remove the problematicestimates.If demandanalyses are
to be performedfor more than one or two units of aggregation or for large and complicatedgroups of products,it may be impracticalto devote a great deal of time to fine-tuning
References: G. (1989),"A Unified
Berndt, E. and A. Silk (1993), "Consistency Requirements and
the Specification of Asymmetric Attraction Models of Aggregate MarketShare,"working paper, MassachusettsInstituteof
and K. Wisniewski (1989), "Price-InducedPatterns of
Competition,"MarketingScience, 5 (4), 112-24.
Boatwright,P. and P. Rossi (1997), "Account Level Modeling for
Trade Promotion:An Application of a ConstrainedParameter
Christen,M., S. Gupta,J. Porter,R. Staelin, and D. Wittink(1997),
"Using Market-LevelData to UnderstandPromotionEffects in
Cooper, L.G., D. Klapper, and A. Inoue (1996), "CompetitiveComponent Analysis: A New Approach to CalibratingAsymmetric Market-Share
Models," Journal of MarketingResearch,
Deaton, A. and J. Muellbauer(1980), Economics and Consumer
Gelfand, A. et al. (1990), "Illustrationof Bayesian Inference in
Normal Data Models Using Gibbs Sampling," Journal of the
Kamakura,W. and G. Russell (1989), "A Probabilistic Choice
Model for MarketSegmentationand Elasticity Structure,"
Montgomery, A. (1997), "Creating Micro-MarketingStrategies
Scanner Data," MarketingScience, 16 (4),
Newton, M. and A. Raftery(1994), "ApproximateBayesian Inference with the Weighted Likelihood Bootstrap,"Journal of the
Royal Statistical Society, Series B, 56, 3-48.
Villas-Boas, M. and R. Winer (1996), "Endogeneity in Brand
Choice Models," working paper, University of California,
Please join StudyMode to read the full document