ABSTRACT
Pretrial risk assessment instruments (PRAIs) have captured national attention in recent years. These instruments utilize computer algorithms to aid judges in making two predictions: whether a person will return to court while on pretrial release and whether a person will pose a danger to the public while on pretrial release.
This Article contends that PRAI advocates have been inattentive to how criminal law and policy sustains the subordination of nonwhite communities. This dominant framework is concerned with the disproportionate incarceration of Black and Latinx people but concludes that racial disparities emerge from an overreliance on monetary bail and subjective risk assessment. Within the dominant framework, PRAIs will: evaluate risk according to objective criteria, identify exceptionally dangerous people for pretrial detention, and ensure monetary bail is set in an amount that facilitates the release of pretrial defendants.
The Article develops an anti-subordination framework to assess contemporary bail practices in relation to PRAIs. This framework emphasizes the unreliability of dangerousness predictions and how these predictions contributed to the dramatic growth in detention populations following the Civil Rights Movement. PRAIs do not resolve the inaccuracies that attend these predictions, but instead derive their risk scores from criminal data that reflects the biased and subjective predictions of the past. Rather than facilitate racial equality, the anti-subordination framework concludes that PRAIs legitimize dangerousness predictions and usher in bail regimes that continue to disproportionately harm Black and Latinx people.
The Article directs attention to two particular states—California and New York—in which grassroots coalitions launched prominent bail campaigns, seeking legislation that would dramatically reduce pretrial detention populations and rectify racial disparities within those
[pdf-embedder url="https://www.uclalawreview.org/wp-content/uploads/securepdfs/2022/03/Hill-68-4.pdf"]