Match-Rigging in Professional Sumo: Elucidation of Incentive Structures and Empirical Analysis

This paper first develops a simple model to show that incentives for match-rigging in professional sumo differ depending on the rank to which a wrestler belongs. While incentives for match-rigging can arise for wrestlers in the top divisions (sekitori), few incentives arise for wrestlers-in-training (makushita and below). We then report the results of our empirical analysis, which show that match-rigging in the top divisions has declined in the post-Duggan and Levitt (2002) period, and that at the wrestler-in-training level there was no statistically detectable match-rigging during any period. We find further that match-rigging in professional sumo has stabilized at statistically undetectable levels following the mobile phone text message match-rigging scandal, and that on average the wrestlers selected for sanctions following the scandal were in fact those most involved in match-rigging.