The Effects of Pre-Trial Detention on Conviction, Future Crime, and Employment: Evidence from Randomly Assigned Judges

Key Takeaways

Pre-trial detention materially increases convictions and guilty pleas, showing that the bail decision can effectively predetermine case outcomes; For marginal defendants, especially those without a recent prior record, pre-trial release produces large long-run gains in formal employment and earnings; Detention’s short-run crime reduction via incapacitation is offset by higher post-disposition offending, yielding no net decline in overall crime; The main system cost of greater pre-trial release is higher failure-to-appear rates, not more total crime, suggesting court-support tools may be more efficient than detention; Judge-specific variation in bail leniency produces large, measurable differences in case and life outcomes, making judicial assignment and tendencies central to pre-trial strategy and policy design.

Dataset Description

The study links court records for over 420,000 criminal cases involving more than 238,000 unique defendants in Philadelphia County (2007–2014) and Miami-Dade County (2006–2014) to U.S. Internal Revenue Service (IRS) administrative tax data. Court data provide detailed information on bail hearings, judge identity, bail type and amount, pre-trial release status, subsequent case processing, convictions, sentences, failures to appear, and rearrests. IRS data supply pre- and post-arrest formal earnings, employment, tax filing, Unemployment Insurance (UI) receipt, and Earned Income Tax Credit (EITC) claims. Defendants are quasi-randomly assigned to multiple bail judges with differing propensities to grant release, enabling an instrumental variables design that identifies local average treatment effects of pre-trial release for marginal defendants in large urban U.S. state trial courts.

Key Findings

Using bail-judge leniency as an instrument, the authors show that pre-trial detention substantially worsens case and life outcomes for marginal defendants. Being detained rather than released increases conviction rates by roughly 15 percentage points (about 26 percent), largely by inducing guilty pleas and reducing defendants’ bargaining power. Detention significantly raises the likelihood of pleading guilty to a charge that results in time served, but has relatively small effects on the length of post-sentencing incarceration, conditional on conviction. On public safety, pre-trial detention reduces crime mechanically before case disposition but increases crime after disposition through criminogenic effects of jail time and the collateral consequences of a conviction; over a two-year horizon, these opposing effects roughly net out to no statistically significant reduction in total crime. In contrast, pre-trial release significantly improves long-run formal employment, earnings, tax filing, and access to UI and EITC, with especially large effects for defendants with no recent prior record. A back-of-the-envelope cost–benefit analysis suggests that, at the margin, releasing these defendants yields a large positive net social benefit, on the order of tens of thousands of dollars per marginal defendant, driven mainly by fewer convictions and stronger labor-market attachment, while the primary measurable cost is a higher failure-to-appear rate rather than an increase in net crime.

Summary

This article exploits quasi-random assignment of criminal defendants to bail judges in Philadelphia and Miami-Dade to isolate the causal effects of pre-trial detention versus release. Because judges vary systematically in their propensity to grant release, the authors use judge leniency as an instrument for pre-trial detention in a two-stage least squares framework. Linking rich court records to federal IRS administrative data allows them to trace not only case outcomes and subsequent criminal justice contact, but also longer-run labor-market consequences several years after the bail hearing.

For criminal case outcomes, the results show that pre-trial detention sharply shifts the playing field. Detained defendants are significantly more likely to be convicted and to plead guilty, with the conviction rate rising by about 15 percentage points compared with marginally similar defendants who are released. Much of this effect reflects pressure to accept plea deals, including pleas that convert pre-trial jail time into a sentence of time served. Detention thus operates less by increasing formal post-sentence incarceration and more by changing bargaining leverage, making the bail hearing a pivotal substantive decision point in the case.

Turning to public safety, the study evaluates a core justification for detention: crime prevention. Detention does reduce rearrests prior to case disposition through incapacitation and logistical barriers. However, this short-run benefit is offset by increased crime in the months and years after case resolution, consistent with the criminogenic impact of jail and the stigma and legal disabilities associated with a conviction. Over a two-year window, the authors find no statistically significant net reduction in total offending attributable to pre-trial detention for marginal defendants, undermining the notion that routine detention is an effective crime-control tool in this context.

The most enduring impacts emerge in the labor market. By following defendants’ IRS records for several years, the authors show that pre-trial release meaningfully boosts formal employment, wage income, and participation in programs like UI and the EITC. These gains are particularly large for individuals with no recent prior offenses, for whom avoiding a new conviction appears to preserve access to stable, legal work. The evidence suggests that pre-trial detention can entrench economic marginalization, whereas release can help maintain or restore attachment to the formal labor market.

Finally, the authors synthesize their empirical estimates in a partial cost–benefit analysis. Accounting for jail costs, changes in future earnings and tax payments, social-transfer receipt, and crime, they conclude that marginally expanding pre-trial release for defendants at the cusp of detention yields large net social benefits, on the order of $37,000–$40,000 per marginal defendant. The primary measurable downside is an increase in failures to appear, not an increase in total crime. These results highlight how judge-level discretion in bail practices can generate substantial and uneven social costs and benefits, and they provide a powerful empirical foundation for bail reform proposals that favor greater use of noncustodial pre-trial supervision over detention.

How the Study Advances Empirical Understanding of Legal Outcomes

The study finds that criminal case outcomes are not the product of randomness but instead reflect structured, repeatable patterns arising from institutional decision points, particularly the pretrial detention process and its downstream procedural effects. By exploiting quasi-random assignment within existing court systems, the analysis shows how formally similar cases can diverge systematically due to embedded procedural mechanisms that shape bargaining dynamics, case resolution, and longer-term outcomes. This methodologically rigorous approach aligns with Pre/Dicta’s emphasis on empirical, case-based analysis of legal outcomes and decision contexts as a necessary foundation for high-level strategic litigation practice grounded in observable institutional regularities.

PreDicta graphic about predictive judicial intelligence with four feature icons

Pre/Dicta Debuts Suite of New Capabilities, Further Expanding its Predictive Insights into Judicial Decision-Making