Present or future incentives? On the optimality of fixed wages with moral hazard

This paper uses a laboratory experiment to show that principals can defer all incentives for present effort to future payments—and thus pay fixed wages—and still motivate workers at the least cost whenever outcomes are observable. This result contrasts with the prediction of the classical moral hazard model, according to which future and present payments must be made contingent on present outcomes to induce effort at the least cost. Even though risk aversion cannot explain this result, I estimate an expectation-based reference-dependent model to show that it is consistent with loss aversion.