Revisiting the Tradeoff between Risk and Incentives: The Shocking Effect of Random Shocks

Despite its central role in the theory of incentives, empirical evidence of a tradeoff between risk and incentives remains scarce. We reexamine this empirical puzzle in a controlled laboratory environment so as to isolate possible confounding factors encountered in the field. In line with the principal-agent model, we find that principals increase fixed pay while lowering performance pay when the relationship between effort and output is noisier. Unexpectedly, agents produce substantially more in the noisy environment than in the baseline despite lesser pay for performance. We show that this result can be accounted for by introducing agents’ loss aversion in the principal-agent model. Our findings call for an extension of standard agency models and for a reassessment of apparently inefficient management practices.