Website CV Paper
Fields
Industrial Organization, Applied Micro
Job market paper
Algorithmic Ranking and Delivery Foreclosure in Retail Platforms: Evidence from Amazon and Mercado Libre
Retail platforms can tie search visibility to sellers' purchase of the platform's integrated warehousing and delivery services. Competition authorities are investigating this practice, but its welfare effects remain unquantified. Furthermore, when platforms compete for consumers, a rival can limit how much any one platform gains from favoring its own services, and a remedy imposed on one platform changes the incentives of the other. Using a panel of more than 415,000 search-result observations from Amazon and Mercado Libre, Mexico's two largest retail platforms, I estimate a structural model of platform competition in which consumers choose between platforms and search within them, sellers choose prices and delivery methods, and platforms set fees and ranking algorithms. Integrated delivery is higher quality than third-party alternatives but up to 96% more expensive, and both platforms rank it systematically higher. Separating Mercado Libre's delivery service from its platform lowers total surplus by 4.5% and consumer surplus on Amazon by 8.4%, because Amazon no longer faces a rival with a comparable integrated offering. Barring both platforms from favoring integrated delivery reduces the adoption of the integrated delivery by up to 9.1%, lowers prices by up to 8.0%, and raises total surplus by 3.7%, with consumer surplus raising 10.8% on Amazon and 7.2% on Mercado Libre.