You don't always get what you pay for

Consider a principal-agent relationship in which more effort by the agent raises the likelihood of success. Does rewarding success, i.e., paying a bonus, increase effort in this case? I find that bonuses have not only an incentive but also an income effect. Overall, bonuses paid for success may well reduce effort and hence the probability of success. I also identify conditions under which the income effect dominates the incentive effect, and single out the hazard-rate of effort as a crucial determinant of this trade-off.