Optimal information censorship

Abstract

This paper analyses Bayesian persuasion of a privately informed receiver in a linear framework. The sender is restricted to censorship, that is, to strategies in which each state is either perfectly revealed or hidden. I develop a new approach to finding optimal censorship strategies based on direct optimisation. I also show how this approach can be used to restrict the set of optimal censorship schemes, and to analyse optimal censorship under certain classes of distributions of the receiver’s type.