Beyond Cash Bail: Unintended Consequences of Maryland's Pretrial Reform

• In Event: Method Meets Mission: Empirical Applications Across Crime, Justice, and Community Response

Abstract

Bail reform efforts across the United States promise to reduce pretrial detention inequities, yet fundamental questions remain about their true impacts. When Maryland implemented new pretrial release rules in 2016/2017, decision-makers faced a critical challenge: how to reduce monetary bail without compromising public safety. This paper examines this policy shift through a methodological lens that combines traditional evaluation with machine learning. Using a regression-discontinuity in time design with comprehensive statewide court data, we analyze judicial detention decisions before and after reform implementation. We find that while monetary bail decreased as intended, detention without bail simultaneously increased—creating an unexpected tradeoff between different forms of pretrial constraint. We develop a machine learning approach to assess whether these shifting detention patterns aligned with empirical risk profiles. Our analysis reveals large heterogeneity in the relationship between release decisions and a priori risk across jurisdictions, suggesting inconsistent interpretation of the same statewide mandate. This research demonstrates how integrating machine learning with policy analysis can illuminate implementation nuances invisible to traditional methods and offer deeper insights into how reform impacts judicial discretion.

Authors

• Lydia Becker, University of Maryland, College Park

• Zubin Jelveh, University of Maryland, College Park