Local Randomization Regression Discontinuity for Event Studies: Application to a Penal Procedure Reform in Uruguay

Title: “Local Randomization Regression Discontinuity for Event Studies: Application to a Penal Procedure Reform in Uruguay” (with M. Cattaneo and R. Titiunik). Speaker: Carlos Díaz (UAH) Abstract: We explore the potential unintended consequences of a new code of penal procedure on criminal behavior. In November 2017, Uruguay switched from an inquisitorial system (where a single judge leads the investigation and decides the appropriate punishment for a particular crime) to an adversarial system (where the judge serves as a referee to ensure correct procedures but the investigation is led by prosecutors). Using a regression-discontinuity design for the most common offenses reported in Montevideo, we find a local and significant increase in the total number of reports filed when the new adversarial code became effective. Results show an increase of 21-24 police reports per day, accounting for 23%-26% of the average annual increase. We also discuss the conditions that are required to use the regression-discontinuity framework in applications where time periods both index the units and function as the score variable.