2026/02/19 by Hector Chade, Héctor Chade, Jan Eeckhout
Mathematics · Computer Science · #Markov Chains and Monte Carlo Methods #Distributed systems and fault tolerance #Algorithms and Data Compression
paper · doi:10.1016/j.jet.2026.106146
In many economic settings, matching takes place under imperfect information about agents’ characteristics that are payoff-relevant to the match. This paper develops a matching model in which agents’ payoff-relevant characteristics are realized after matching takes place, and matching is based on the agents’ observable ex-ante characteristics, which stochastically determine the payoff-relevant ones. We derive conditions for positive and negative sorting. The conditions differ depending on whether agents’ characteristics are stochastically larger or more variable as a function of the agents’ ex-ante attributes. We apply our framework to risk sharing and household income inequality.