An Empirical Investigation of Online Competitive Price Promotions

Using daily price data collected from a leading price comparison website, this paper empirically examines the role of vertical store differentiation on firms’ equilibrium pricing strategies in a clearinghouse model. By allowing for different quality levels across firms the theoretical model used for estimation generates asymmetric pricing strategies. Our estimates show that observable store effects explain about 30 percent of the variation in prices for four digital cameras, a finding that is supported by hedonic regression results. Furthermore, we find that prices adjusted for quality differences between stores have greater intra-distribution mobility than unadjusted prices: stores move up and down the adjusted price distribution more frequently so that consumers cannot tell the identity of the lowest-priced firm over time. Using duration and Markov transition analysis we find that firms keep their prices constant for on average three weeks. We structurally estimate the model using nonlinear least squares and show nonlinear least squares gives consistent parameter estimates. Our general model explains the real world price data reasonably well. Factors that generate price dispersion are similar across the four digital cameras: the maximal profit margin ranges from 16 to 22 percent and the estimated proportion of consumers who visit the price comparison site is 16 to 26 percent. The cost of updating the prices is about two percent of the maximal profit margin. Finally, our estimates suggest that store managers believe to be competing with four to five stores.