Evidence, Beliefs, and the Design of Judicial Mechanisms

Motivated by a growing body of empirical evidence documenting overconfidence among litigants, this paper studies the design of judicial mechanisms when defendants hold biased beliefs about their likelihood of conviction. Subjective biases have heterogeneous effects across defendant types and may fundamentally alter the structure of optimal judicial procedures. Relative to the benchmark with unbiased defendants, biased beliefs reduce the benefits of plea offers and, in some cases, the value of offering a menu of trial procedures to screen defendants. Overconfidence does not facilitate surplus extraction; instead, it reduces court welfare while increasing defendants’ expected utility. A calibration of the model using empirical evidence on judicial error rates and litigant overconfidence suggests that these effects are of significant magnitude.