Judging Judge Fixed Effects

We propose a nonparametric test for the exclusion and monotonicity assumptions invoked in instrumental variable (IV) designs based on random assignment of cases to judges. show its asymptotic validity demonstrate finite-sample performance simulations. apply our an empirical setting from literature examining effects pretrial detention defendant outcomes New York. When are not satisfied, we weaker versions usual restrictions under which IV estimator still converges proper weighted average treatment effects. (JEL H76, K41)