All Deceptions Are Not Alike: Bayesian Mechanism Design with a Social Norm against Lying

We say that a society has a weak norm against lying if, all other things being equal, agents try to avoid getting caught lying. We show that if this is the case, and it usually is, then Bayesian monotonicity is no longer a constraint in implementation and all incentive-compatible social-choice functions are Bayesian implementable. In contrast to the previous literature, our result does not rely on any kind of intrinsic lying aversion, on which the experimental evidence is mixed.