Sabotage in dynamic tournaments

This paper studies sabotage in a dynamic tournament. Three players compete in two rounds. In the final round, a player who is leading in the race, but not yet beyond the reach of his competitors, gets sabotaged more heavily. As a consequence, if players are at the same position initially, they do not work productively or sabotage at all in the first round. Thus sabotage is not only directly destructive, but also depresses incentives to work productively. If players are heterogeneous ex ante, sabotage activities in the first round may be concentrated against an underdog, contrary to findings from static tournaments. We also discuss the robustness of our results in a less stylized environment.