WorksEconomics, Econometrics and Finance
Platform Recommendation and Competition in Search Markets
Abstract
Recommendation rules affect both sellers' incentives and the alternatives consumers expect to find. We develop a framework for evaluating these rules with heterogeneous sellers, endogenous pricing and advertising, and costly consumer search. For a broad class of separable recommendation scores, we establish conditions under which a unique market equilibrium exists. Three applications illustrate the use of the framework. A stronger reward for low prices does not necessarily lower equilibrium prices and benefit consumers when products are sufficiently differentiated in qualities. Quality guidance can lower the price of each product and benefit consumers while raising the average transaction price. With heterogeneous match distributions, a reward for the best initial inspection can help consumers with advertising fixed and harm them when advertising adjusts. A two-type design extension shows that no finite optimum over implementable outcomes uses positive weights on both quality and low prices, even when the platform values consumer surplus.