Persuasion as Matching;Alexander Wolitzky (Massachusetts Institute of Technology)

Abstract

In a standard persuasion problem where the receiver's action is one-dimensional and her utility is single-peaked in the action, optimal signals pool at most two states in each realization. This result reduces persuasion to the problem of which states to match, and with what weights. We derive simple conditions--driven by economic properties such as super modularity and submodularity of preferences--for the optimality of positive and negative assortative matching/disclosure. Unlike in classical matching, there are many negative assortative matches, which we characterize further. Our approach unifies a wide range of previous findings and generates new applications.

Click here to view the paper (earlier version).