Three Essays on Unethical Behaviour and Deterrence Mechanisms in Contests

This thesis consists of three self-contained essays. We study undesirable behaviors such as cheating and self-sabotaging in contests in a laboratory environment. The rst paper proposes a new anti-doping policy. In a conditional superannuation scheme, athletes have to pay a certain fraction of their proceeds from sports into a fund from which they can draw only well after their careers and if they have never been caught doping. Theoretically, this fund has two important advantages over conventional anti-doping policies such as bans and nes. It does not lose its deterrence e ect when athletes approach the end of their careers (unlike bans), and it can deal with the widespread problem that drug cheats are often only found out much later when the detection technology has caught up with doping practices. We build a model of a dynamic sporting contest, implement it in the laboratory and compare the performance of our policy to that of traditional policies. Our policy compares favorably with respect to doping prevention and the quality of resulting sporting contests. In the second paper, we study a tournament that rewards not only winners but also losers with extremely bad performance, which creates an incentive for underdogs to under-perform deliberately. Such a contest scheme is often employed to improve the long term competitive balance. We design two treatments, with or without the leaderboard, to investigate whether social status can reduce this self-sabotaging behaviour. The leaderboard of all participants' ranked performance is used as a proxy of social status. Our results show that underdogs respond to the monetary self-sabotaging incentives in contests. In addition, individuals tend to self-sabotage just enough when the leaderboard is displayed to everybody. Without the leaderboard, players self-sabotage more excessively. We conjecture that by achieving exactly the level of performance that gives the consolation prize, tankers in the leaderboard treatment want to signal that they understand the game well and they tanked to receive the consolation prize. The third paper addresses an agency problem in a contest between two contestants, each with a manager. Individuals that are engaged in contests have strong incentives to cheat. Sanctions are designed to deter potential cheaters. Often other agents in the contestant's team (e.g., a coach of an athlete) or company (a manager of an R&D engineer) have a bene t from cheating and can in uence on the cheating decision. If only the contestant is punished for cheating, an agency problem arises. We show theoretically, that extending the liability from the contestant to the manager reduces cheating only if fines are suffciently high. Otherwise over-all cheating rate increases. Experimental tests confi rm that for high fines joint liability is effective in reducing cheating, while predicted detrimental effect of joint liability when fines are low does not materialise.Thesis (Ph.D.) -- University of Adelaide, School of Economics, 201