An experimental study on tournament design

Since recently, rank order tournaments have become quite popular for providing incentives in employment relationships. However, the consequences of different tournament designs are widely unexplored. This paper experimentally investigates different tournament design alternatives along two dimensions: tournament size and prize structure. We find that average effort tends to increase with a higher proportion of winner prizes. Additionally, variability of effort is lower if the number of winner prizes is high. Especially two-person tournaments are prone to collusion. Furthermore, we observe a restart effect. (C) 2003 Elsevier Science B.V All rights reserved.