Essays on Platform Mergers and Dynamic Price Competition

Abstract

This dissertation covers two innovative topics related to horizontal mergers. One chapter investigates the welfare effect of mergers in the platform markets; another one studies the impact of mergers where firms conduct dynamic price competition with asymmetric information.

In Chapter 1, I analyze the welfare effects of lowering the costs to buyers of searching and multihoming in a setting with multiple two-sided platforms. The analysis is motivated by observed changes following the 2017 acquisition of IronPlanet, which is an online auction marketplace for used heavy equipment by Ritchie Brothers Auctioneers, which operates the largest offline auction marketplace. As is quite common after platform mergers, RBA maintained both platforms but made it easier for buyers of equipment to search across the platforms (multihoming), which has the potential to render the allocation of equipment more efficient, benefitting both buyers and sellers. These efficiencies could offset the market power created by the merger. I use pre- and post-merger transaction data to estimate a new model of search and auction entry by buyers and quantify the increase in welfare effects of the observed changes. Depending on the specification, the proportion of multihoming buyers increases substantially (by 50% in the baseline specification), and the total surplus can increase by more than 8%, although heterogeneity exists in the welfare impact on different market participants. I also consider several additional counterfactuals involving changes in commission and changes in equipment allocation across the marketplaces.

In Chapter 2, my coauthors and I model differentiated product pricing by firms that possess private information about serially correlated state variables, such as their marginal costs, and can use prices to signal information to rivals. In a dynamic game, signaling can raise prices significantly above static complete information Nash levels even when the privately observed state variables are restricted to lie in narrow ranges. We calibrate our model using data from the beer industry, and we show that our model can explain changes in price levels and price dynamics after the 2008 MillerCoors joint venture.