Search Costs , Demand Structure and Long Tail in Electronic Markets : Theory and Evidence

1. Introduction An important advantage that internet channels posit over physical markets is a reduction in search costs for prices. The reduction in search cost has rekindled interests in examining the Law of One Price (Brynjolfsson and Smith 2000). Empirical evidence shows that firms have been able to maintain their pricing power and there is significant price dispersion online (Brynjolfsson and Smith 2000, Baye et al. 2005). These price dispersions are often attributed to the presence of search costs in the online world. In a similar vein, theoretical research has also shown that search costs create imperfect information about sellers' prices among consumers, which leads to equilibrium price dispersion in otherwise homogeneous product markets (Reinganum 1979). However, despite its importance, little empirical evidence exists so far on how search costs actually affect demand structures faced by retailers. To empirically investigate whether search costs are present in online retail markets, our research uses prior theoretical insights that relate the nature of consumer demand structure to the presence of search costs. The presence of search costs also implies that that it takes time for information about price changes to dissipate among consumers in the market. This leads to lower short-term price elasticity and higher long-term price elasticity, especially for information on price decreases. A number of empirical studies have tried to quantify search costs both in online and offline settings.The empirical work on search costs has shown that online search costs, although smaller than search costs in physical markets, are still significant. The objective of our study is therefore to consider how such search costs affect market competition, how they influence consumer demand structure faced by online retailers and how fast price change information dissipates in the online market. An additional contribution of this study is that we do so with only aggregate level data by exploring a unique insight of Stiglitz (1989) that search costs create kinkedness in aggregate demand when firms change prices. The significance of the kink reflects the magnitude of online search costs and the kinked demand function provides information on how search costs affect competition in the online market. One advantage of using aggregate level data is that we can estimate the impact of search cost for an entire market. This approach compliments prior studies that mainly focus on estimating search costs for individual consumers. Further, our approach allows us to estimate the change in consumer surplus …