Carnegie Mellon University

File(s) stored somewhere else

Please note: Linked content is NOT stored on Carnegie Mellon University and we can't guarantee its availability, quality, security or accept any liability.

Static and Dynamic Pricing in Online Markets: Evidence from Books

journal contribution
posted on 2004-08-02, 00:00 authored by Karen ClayKaren Clay, Michael SmithMichael Smith
Commentators have observed that the ease of monitoring competitors on the Internet may allow Internet retailers to engage in non-competitive pricing. Using data on the daily prices of 399 books at 26 online bookstores between August 1999 and January 2000, we investigate firm pricing behavior in the online book market. Although sales in the Internet channel were growing very rapidly at the time, we find that relative prices differed across the three categories of bookstores (big three, active fringe, and inactive fringe), and these differences were remarkably stable over time. We present a simple model in which cross-channel competition and differentially informed consumers lead to the observed (static) pricing patterns. Although relative prices were stable, prices did change and were on average increasing over time. We document the dynamic strategic interaction across firm categories and across individual firms. Ten pairs of firms involving seven individual firms changed prices in the same direction on the same book within three days in an (one standard deviation) above average number of cases and respond more than 25 percent of the time to competitors’ price changes. Given the behavior of these firms and the large market shares held by the top three firms, we formally test a number of oligopoly and non-oligopoly explanations for the observed price changes. We find that the observed patterns were not consistent with the predictions of oligopoly pricing in the Haltiwanger and Harrington (1991) model or with other explanations such as customer loyalty, inventory considerations or changes in elasticity of demand associated with holidays. We conjecture that the observed cases of parallel pricing were largely attributable to experimentation on the part of the initiator and learning or competitive response on the part of the responder.




Usage metrics


    Ref. manager