Citation

Sloan FA, Platt AC, Chepke LM, Blevins CE. J. Risk Uncertain. 2013; 46(1): 51-80.

Copyright

(Copyright © 2013, Holtzbrinck Springer Nature Publishing Group)

DOI

10.1007/s11166-012-9159-z

PMID

unavailable

Abstract

This study presents an empirical analysis of domestic violence case resolution in North Carolina for the years 2004 to 2010. The key hypothesis is that penalties at the level set for domestic violence crimes reduce recidivism (re-arrest on domestic violence charges or conviction in 2 years following an index arrest). We use state court data for all domestic violence-related arrests. Decisions to commit an act of domestic violence are based on a Bayesian process of updating subjective beliefs. Individuals have prior beliefs about penalties for domestic violence based on actual practice in their areas. An individual's experience with an index arrest leads to belief updating. To address endogeneity of case outcomes, we use an instrumental variables strategy based on decisions of prosecutors and judges assigned to each index arrest in our sample. Contrary to our hypothesis, we find that penalities, at least as set at the current levels, do not deter future arrests and convictions.

Language: en