Race, Prediction, and Pretrial Detention

Professor Shima Baradaran has recently posted her newest article, Race, Prediction, and Pretrial Detention, to SSRN.   The article is co-authored with Professor Frank McIntyre.  Here’s the abstract:

“We use a large, nationally representative and detailed database on criminal defendants to predict the likelihood that the defendant will commit a crime while on bail or other forms of pretrial release. We use this likelihood to estimate an empirical model of judicial behavior. We find that large and persistent gaps in pretrial detention across race is completely explainable as a form of statistical discrimination due to the higher probability a black defendant will commit a pretrial violent crime. The results are robust to a variety of possible statistical biases, including selection bias and misspecification of functional form. We also consider more nuanced models of discrimination and find evidence that judges may be more sensitive to violent crimes committed by white defendants than black defendants.”

See more of Professor Baradaran’s articles here.