Abstract:
Platforms may give preferential treatment to their own products in search results. Whether and how to regulate this self-preferencing behavior is an intensely debated antitrust issue. This paper identifies self-preferencing and quantifies its equilibrium welfare effects in Apple App Store. I start by examining the effect of a change in the platform's search algorithm that dropped several Apple's apps from top positions. I find that the search algorithm change leads to significantly higher installations and update frequencies of independent apps that compete with Apple's apps in the same categories. Then I develop an empirical model of consumer search and update competition allowing for potential self-preferencing. The model is estimated with aggregate data on consumer search and purchase, search ranking, and app characteristics. Estimation results point to self-preferencing: Apple's apps are more likely to be ranked higher than independent apps conditional on app quality, price, ratings, and title match with search terms. Based on counterfactual simulations, I find that eliminating the identified self-preferencing modestly increases the quality of independent apps on average. Furthermore, the elimination improves consumer surplus by $2.2 million and profits of independent developers by $1.6 million per month.