We study the large-scale implementation of a system of specialized domestic violence courts (SDVCs), an innovation in access to justice programs for potential victims of intimate partner violence (IPV) and offenders. Using individual-level administrative data from the universe of civil domestic violence cases in Puerto Rico during the period 2014-2020, we leverage the staggered opening of SDVCs across judicial regions to examine the consequences for victims’ judicial protection and offender recidivism. Access to SDVCs leads to a considerable 8 percentage points increase in the probability that judges issue a protection order and a 1.7 percentage point (15 percent) decrease in victim and offender reappearance rates within one year of the start of the case. Effects are more pronounced for cases in which parties have children in common and in which access to SDVCs is more limited. Linking the case data to administrative and survey data on judges, we show that the priorities of judges assigned to SDVCs play a prominent role in explaining these outcomes.

Cambridge, MA: National Bureau of Economic Research, 2025. 

By Min-Taec Kim

The aim of this study was to assess the impact of participating in a program through a High Intensity Program Unit (HIPU) on reoffending, violent reoffending and reimprisonment within 3, 6 and 12 months of leaving custody. METHOD We estimate the impact of the HIPUs on reoffending, violent reoffending and reimprisonment within 3, 6 and 12 months of leaving custody by conducting three different comparisons between people who were identified as eligible for the program: 1. Comparing individuals by program status (Did not start, Started but did not complete, Finished); 2. Comparing individuals by the number of therapeutic hours received at a HIPU, regardless of program status; 3. Comparing individuals using variation in therapeutic hours recieved, restricted to just those who exited due to “insufficient time”. We use regression models across all analyses to adjust for observable differences between individuals. The key threat to our estimates is selection bias, which could be caused by unobserved differences between those starting and/or completing a program through the HIPUs and those who do not. The first two comparisons are (possibly) biased towards a higher estimate, while the third comparison attempts to avoid this bias by leveraging variation in time left to serve.

(Crime and Justice bulletin No. 250), Sydney: NSW Bureau of Crime Statistics and Research.  2022. 22p.