| Abstract: | Contrary to conventional wisdom, a right to silence can reduce wrongful convictions of innocent suspects who tell police the truth, and may reduce the overall conviction rate without affecting the confession rate. We establish these conclusions by modelling interrogation as a game in which suspects are privately informed of their type, but do not know the content of witness reports. Our model has testable implications which distinguish it from competing models. The best data rejects competing models, but does not reject our model. |
