Working Paper: NBER ID: w26999
Authors: David Arnold; Will S. Dobbie; Peter Hull
Abstract: We develop new quasi-experimental tools to measure disparate impact, regardless of its source, in the context of bail decisions. We show that omitted variables bias in pretrial release rate comparisons can be purged by using the quasi-random assignment of judges to estimate average pretrial misconduct risk by race. We find that two-thirds of the release rate disparity between white and Black defendants in New York City is due to the disparate impact of release decisions. We then develop a hierarchical marginal treatment effect model to study the drivers of disparate impact, finding evidence of both racial bias and statistical discrimination.
Keywords: Racial Discrimination; Bail Decisions; Criminal Justice
JEL Codes: C26; J15; K42
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| judge assignment (K40) | release decisions (Y40) |
| racial discrimination (J71) | release decisions (Y40) |
| misconduct risk (K42) | release decisions (Y40) |
| racial discrimination (J71) | disparity in release rates (D39) |
| average misconduct risk among black defendants (K42) | average misconduct risk among white defendants (K49) |
| disparity in release rates (D39) | racial bias (J15) |
| disparity in release rates (D39) | statistical discrimination (J71) |