Incentives or Persuasion? An Experimental Investigation

There are two theoretically parallel ways in which principals can manipulate agents’ choices: with monetary incentives (mechanism design) or Bayesian persuasion (information design). We are interested in whether incentives or persuasion is a better strategy for principals. We conduct an experiment that investigates the behavioral side of the theoretical parallelism between these approaches. We find that principals are more successful when persuading than when incentivizing. Agents appear to be more demanding in mechanism design than in information design. Our analysis also identifies many features that make mechanism and information design behaviorally distinct in practice