Abstract:
This paper investigates the optimal design of incentives when agents distort probabilities. We show that the type of probability distortion displayed by the agent and its degree determine whether an incentivecompatible contract can be implemented, the strength of the incentives included in the optimal contract, and the location of incentives on the output space. Our framework demonstrates that incorporating descriptively-valid theories of risk in a principal-agent setting leads to incentive contracts that are typically observed in practice such as salaries, lump-sum bonuses, and high-performance commissions.
