Release Mechanism and Pretrial Failure in Large Urban Counties

A central question in contemporary bail reform is whether the different forms of monetary release used in large U.S. jurisdictions, commercial surety bonds, deposit bonds, full cash bonds, and property bonds, produce systematically different pretrial outcomes. The commercial bail industry has long defended its role on the grounds that bondsman-supervised release produces superior pretrial outcomes through a private enforcement function not available under alternative mechanisms. The present study tests this claim using data from the 2009 State Court Processing Statistics program on 5271 felony defendants released on financial conditions in 35 large urban counties. Logistic regression models with county fixed effects and cluster-robust standard errors estimate the association between release mechanism and two outcomes, pretrial rearrest and failure to appear (FTA), net of bail amount, prior criminal record, seriousness of offense, criminal justice status at arrest, time from arrest to release, type of legal representation, and demographic characteristics. Three findings emerge. First, defendants released on deposit bonds exhibit substantially lower odds of pretrial rearrest than otherwise comparable defendants released on commercial surety bonds, a finding that is robust across a battery of sensitivity analyses. Second, defendants released on full cash bonds exhibit substantially lower odds of FTA than otherwise comparable defendants released on commercial surety bonds, although this finding is somewhat sensitive to specification choice and is partly mediated by bail amount. Third, no specification supports the public-safety claim made on behalf of commercial bail because surety bonds do not outperform the alternatives for either outcome. These findings indicate that the principal empirical justification for the commercial bail industry is not supported by nationally representative data, and that a shift away from commercial bail toward court-administered alternatives is unlikely to impose behavioral costs and may produce modest public-safety gains.