Executive Summary
Considerable attention has recently been devoted to the dynamics inherent in competitor price interactions. The aim of this article is to contribute to this literature by considering how pricing interactions differ across various periodicities. This issue has received increased interest in the marketing literature as of late: Pauwels and colleagues (2004, p. 176) note that “as there is increasing evidence that the same relationship need not hold among two variables at different frequencies, various substantive marketing problems may warrant further investigation along that dimension.” Because little is known about the nature of pricing interactions across various planning cycles, this article develops several empirical generalizations about the role of periodicity in pricing.
Pricing interactions may vary across different frequencies because prices observed at the retail level stem from multiple decisions (e.g., regular pricing, discounting) made by multiple decision makers (e.g., retailers, manufacturers). For example, temporary price reductions (high-frequency price changes) can be used to price discriminate in the short run, whereas regular price adjustments (low-frequency price changes) might reflect changes in long-term costs or demand. Time disaggregation (Leone 1995) cannot disentangle these factors, because frequency aggregation exists even when data are analyzed at the lowest possible level of temporal aggregation.
To assess the influence of periodicity on competition, the authors employ a spectral decomposition of price series. Spectral analysis is a technique that creates a (co)variance decomposition of price data into price cycles at different frequencies. This descriptive analysis is intended to show how the magnitude and direction of price interactions depend on a given range of frequencies, or “planning horizons.” In a subsequent step, the authors use regression methods to determine how the magnitude and direction of price interactions depend on various brand- and category-level variables.
Applying this analysis using week–store stockkeeping-unit-level price data in 35 grocery categories, the authors find the following:
- Most of the observed variation in price results not from rapid pricing vacillation but rather from slower, less frequent pricing movements. This is surprising in light of the emphasis on discounts over regular pricing in both academic and corporate research.
- Although pricing interactions are more common with regular price changes, they occur across the entire spectrum. Moreover, short-term pricing interactions can offset long-term pricing interactions. For example, regular prices of beer are positively correlated across brands, whereas discounts are negatively correlated. These interactions cancel out when aggregated across frequencies, suggesting no correlation and, thus, the potential for frequency aggregation bias
- Price cycles and periodicities at which prices interact have a strong category component that is common to the brands in a category.
- Using a structural model of price competition, the authors show that inferences about price responses are confounded when periodicity is ignored. They show how spectral techniques can be used to isolate specific frequencies to infer the nature of competitive response at different planning cycles.
Biography
Bart J. Bronnenberg received his PhD and MSc in Management from INSEAD and an MSc in Industrial Engineering from Twente University (the Netherlands). He is an Associate Professor of Marketing in the Anderson Graduate School of Management at the University of California, Los Angeles, where he teaches courses on market assessment. His most recent research focuses on spatial and multimarket aspects of marketing (e.g., how firms sell products differently when they serve many interrelated [domestic or international] markets as opposed to a single market). Other areas of interest are consideration sets and preference formation and consumer choice. Recent publications have appeared in Marketing Science and Journal of Marketing Research.
Carl F. Mela is a Professor of Marketing at Duke University. He received his BSc from Brown University; an MBA from the University of California, Los Angeles; and an MPhil and a PhD from Columbia University. He has also held various management positions at Hewlett-Packard, Hughes, and Proxima. Professor Mela’s research focuses on the long-term effects of marketing activity. Articles along these lines have appeared in Journal of Marketing Research, Marketing Science, and Journal of Consumer Research, among other journals. His research has received nine best-paper awards from the Marketing Science Institute, INFORMS, the American Marketing Association, and other professional organizations. He serves on the editorial boards of Journal of Marketing, Journal of Marketing Research, Marketing Science, Marketing Letters, Quantitative Marketing and Economics, and Journal of Public Policy & Marketing. For more information, see his home page at http://faculty.fuqua.duke.edu/~mela/bio.
William Boulding is Professor of Business Administration and Associate Dean of the Daytime MBA program at the Fuqua School of Business, Duke University. His research focuses on evaluating how managers make decisions and how consumers respond to those decisions. He publishes his research in marketing and management journals, including Marketing Science, Management Science, Journal of Consumer Research, Harvard Business Review, Journal of Marketing, and Journal of Marketing Research. He is a cowinner of the 1998 William F. O’Dell Award, given by Journal of Marketing Research. He has served on the editorial boards of Journal of Marketing Research, Journal of Consumer Research, and Journal of Service Research. He has also served as a consulting editor for Journal of Marketing, is a past associate editor for Journal of Consumer Research, and is a past area editor for Marketing Science.
J Marketing Research, Volume 43, Number 3, August 2006
View Table of Contents.