Learning or Networking? The Causal Effect of Judges Sitting by Designation
利用台湾近百万案件数据，研究发现上诉法院法官若与初审法官曾是同事，改判率显著降低（人脉效应），而临时晋升制度后整体改判率下降（学习效应）。
When lower court judges sit on panels with higher court judges, the former may learn from the latter, enter into personal relationships with the latter, or both. Lemley and Miller’s empirical study on US federal courts found that judges sitting by designation see a subsequent reduction in their reversal rates. Leveraging a unique dataset encompassing close to 1 million cases from Taiwan, in particular 2,591 appeals, our study examines whether Taiwan’s transitory promotion system has led to lower reversal rates and, if so, why. Using a difference-in-differences framework, we find that when cases appealed to appellate courts were assigned to judges who were formerly colleagues of the judges who penned the district court opinions, reversal rates drop statistically significantly (demonstrating the “networking effect”). Without a causal identification design, we show that the overall reversal rates are lower in the period after transitory promotions (consistent with the “learning effect”).