Incentive schemes, moral hazard, and deterrence: evidence from British horseracing

A fundamental challenge in principal–agent relationships is ensuring agents exert effort when their actions are not perfectly observable. We examine a natural experiment in effort enforcement in British horseracing, where jockeys are required not only to try to win but to be visibly seen to be trying. Jockeys who appear not to be trying face suspensions and significant penalties in terms of income forgone. Using a dataset of some 60,000 races involving more than 1,500 jockeys and 17,000 horses, we test whether these apparent non-trying events represent attempts to fix a race or influence future races. We find no evidence to support either hypothesis. Performance neither improves in the races before, nor deteriorates in the races after, a ban. Punishments for non-trying may nonetheless be rational as a deterrence mechanism–making an example of a small number of jockeys serves as a motivation for the rest: pour encourager les autres.