Kentaro Fukumoto and Mikitaka Masuyama

Prepared for delivery at the 2006 Annual Meeting of the American Political Science Association, Philadelphia, August 31-September 3, 2006.

Abstract

Ramseyer and Rasmusen (2003) argue that a judge who once belonged to a leftist group takes longer to reach a moderately prestigious administrative status, and that Japanese judges are conservatively biased. Their findings are, however, dependent on the assumption that they observe or completely estimate the timing of promotion for all judges, which is not correct. In an attempt to deal with the issues of censoring and time dependence, we apply survival analysis to the data set of judicial careers, which provides us with a methodological challenge to solve simultaneously the problems of left truncation and split population. Moreover, we propose a way to estimate an average treatment effect on duration time and to perform matching to reduce model dependence of our analysis. A most important finding is that leftist judges are less likely to be on the track to a prestigious post, but once they are on the track, they are not discriminated in terms of the timing of promotion.