Regulating Speech with Bayesian Audiences

Abstract

We analyze the regulation of false statements in the presence of Bayesian audiences. We find that: (a) Often, moderate sanctions are optimal even though strict sanctions can fully deter all false statements; (b) the existence of separating equilibria—where only truthful statements are made—critically depends on judicial accuracy; (c) the magnitude of sanctions trades-off false information, chilling of truthful statements, and litigation costs; and (d) private enforcement often dominates public enforcement despite the lack of commitment. We emphasize the case of defamation law and discuss other contexts including securities regulation, whistle-blower incentives, jury trials, and reports of criminal activity.