Racial Sentencing Disparities and Differential Progression Through the Criminal Justice System: Evidence From Linked Federal and State Court Data

TL;DR

This paper investigates racial disparities in sentencing within the criminal justice system, highlighting the significant role of federal judges and the impact of sample selection bias in state courts on disparity estimates.

Contribution

It provides new evidence on the influence of different actors and sample selection bias in estimating racial sentencing disparities using linked federal and state court data.

Findings

Federal judges largely responsible for disparities

Sample selection bias underestimates true disparities in state courts

Federal data shows no sample selection bias

Abstract

Several key actors -- police, prosecutors, judges -- can alter the course of individuals passing through the multi-staged criminal justice system. I use linked arrest-sentencing data for federal courts from 1994-2010 to examine the role that earlier stages play when estimating Black-white sentencing gaps. I find no evidence of sample selection at play in the federal setting, suggesting federal judges are largely responsible for racial sentencing disparities. In contrast, I document substantial sample selection bias in two different state courts systems. Estimates of racial and ethnic sentencing gaps that ignore selection underestimate the true disparities by 15% and 13% respectively.