Policing Practices That Curb Domestic Violence

Featured in print Digest

The lifetime incidence of intimate partner violence is 24 percent in the United Kingdom and slightly higher, 26 percent, in the United States. Two studies of policing in England conclude that taking legal action against the perpetrator reduces the likelihood of repeated abuse, refuting concerns that arrests and prosecutions boomerang by triggering retaliation and discouraging future calls for help.

In Deterrence or Backlash? Arrests and the Dynamics of Domestic Violence (NBER Working Paper 30855) Sofia AmaralGordon B. DahlVictoria Endl-GeyerTimo Hener, and Helmut Rainer report that an arrest halves domestic violence calls about the arrested person in the ensuing year. Most domestic violence calls do not, however, lead to an arrest or prosecution.

In Criminal Charges, Risk Assessment, and Violent Recidivism in Cases of Domestic Abuse (NBER Working Paper 30884), Dan A. BlackJeffrey GroggerTom Kirchmaier, and Koen Sanders find that pressing charges substantially reduces violent recidivism while providing risk-based protective services to victims does little to break the cycle of abuse.

Arresting or pressing charges against those who abuse their intimate partners reduces the likelihood of repeat behavior.


The first study uses arrest data as the key variable and also uses the fact that police officers are quasi-randomly assigned to emergency calls, with these officers differing in their propensities to make arrests. The second study uses data on whether charges were pressed and controls for other possible explanatory factors such as the severity of an incident and the responding officers’ propensity to make arrests or press for charges.

Deterrence or Backlash? focuses on the West Midlands, the second most populous county in England, and examines data from more than 124,000 domestic violence emergency calls made over the last decade. These calls include reports of threatening behavior, violence, or abuse.

In the absence of an arrest, abuse reoccurs in an estimated 23 percent of cases within 48 hours of the emergency call. Short-term revictimization is unlikely when there is an arrest, which allows for a cooling-off period during which the offender is no longer interacting with the victim. Additional reductions in revictimization occur over the following year, consistent with a longer-term deterrence effect. The initial arrest reduces the probability of another domestic violence call over the subsequent year by 51 percent. The deterrence effect is reinforced by the increased probability of prosecution: offenders who are arrested face a 12 percent chance of formal charges, compared with just 2 percent for those who are not arrested.

The researchers cite two indicators that suggest that the reduction in emergency calls reflects an actual decline in violence as opposed to a change in reporting behavior induced by fear of retaliation. First, post-arrest calls on average were for less-violent behavior, suggesting victims had a lower tolerance for abuse. Second, the composition of callers changed: repeat calls by third parties, such as neighbors, dropped by 57 percent, while those by victims fell 36 percent, the opposite of what one would expect if victims felt intimidated.

These findings are reinforced by those of Criminal Charges, Risk Assessment. In this study, the researchers assess two approaches used by the Greater Manchester Police to reduce serious domestic violence recidivism. One entails pressing criminal charges, the other sparing the offender and instead providing protective services to the victim based on a risk assessment made at the scene of the incident. The sample consists of more than 154,000 incidents between 2013 and 2019.

The study finds that pressing charges reduces violent recidivism over the following year by about 5 percentage points, a nearly 40 percent reduction relative to the average recidivism rate. Providing victims with protective services does not reduce violent recidivism on average. As in the previous study, pressing charges resulted in a greater drop in repeat calls from third parties than from victims.


—Steve Maas