Bayesian privacy

Modern information technologies make it possible to store, analyze, and trade unprecedented amounts of detailed about individuals. This has led public discussions on whether individuals' privacy should be better protected by restricting the amount or precision that is collected commercial institutions their participants. We contribute this discussion proposing a Bayesian approach measure loss in mechanism. Specifically, we define associated with mechanism as difference between designer's prior posterior beliefs an agent's type, where calculated using Kullback–Leibler divergence, change triggered actions taken agent consider both ex post (for every realized maximal cannot exceed some threshold κ ) ante (the expected over all type realizations measures loss. Applying these notions monopolistic screening environment Mussa Rosen (1978), study properties optimal privacy‐constrained mechanisms relation welfare/profits levels.