An, Yonghong Baye, Michael R. Hu, Yingyao Morgan, John Shum, Matt
Year of Publication:
Working Paper 564
This paper (1) presents a general model of online price competition, (2) shows how to structurally estimate the underlying parameters of the model when the number of competing firms is unknown or in dispute, (3) estimates these parameters based on UK data for personal digital assistants, and (4) uses these estimates to simulate the competitive effects of horizontal mergers. Our results suggest that competitive effects in this online market are more closely aligned with the simple homogeneous product Bertrand model than might be expected given the observed price dispersion and number of firms. Our estimates indicate that so long as two firms remain in the market post merger, the average transaction price is roughly unaffected by horizontal mergers. However, there are potential distributional effects; our estimates indicate that a three-to-two merger raises the average transaction price paid by price sensitive 'shoppers' by 2.88 percent, while lowering the average transaction price paid by consumers loyal to a particular firm by 1.37 percent.