Does Transparency Really Increase Corruption? Evidence from the 'reform' of Figure Skating Judging

Transparency is usually thought to reduce corruption by facilitating monitoring by outsiders, but there is concern it can have the perverse effect of facilitating collusion by insiders. In response to vote trading scandals in the 1998 and 2002 Olympics, the International Skating Union (ISU) introduced a number of changes to its judging system, including obscuring which judge issued which mark. The stated intent was to disrupt collusion by groups of judges, but this change also frustrates most attempts by outsiders to monitor judge behavior. In Zitzewitz (2006), I found that prior to the changes, skaters with a compatriot on the judging panel benefited both from a nationalistic bias from their same‐country judge and from vote trading. The ISU's anonymization of judges prevents me from separately estimating the nationalism and vote‐trading components of the "compatriot‐judge effect", but I can still estimate its aggregate. I find that substantial judging biases remain. In my pre‐scandal sample, a compatriot judge increased a skater's score by 0.043 standard deviations (SDs). Immediately after the 2002 scandal, judges were anonymized, and the compatriot judge effect increased to 0.052 SDs. The scoring system was subsequently changed to one that put less weight on judges' scores, but the compatriot‐judge effect on those scores again increased slightly (to about 0.06 SDs). Further evidence suggests that the compatriot‐judge effect continues to reflect both nationalistic bias and vote trading. Taken together, the results suggest that reducing transparency did not have the promised effect of reducing corruption.