85 Pages Posted: 17 Jul 2016 Last revised: 29 Mar 2017
Date Written: 2017
In misdemeanor cases, pretrial detention poses a particular problem because it may induce otherwise innocent defendants to plead guilty in order to exit jail, potentially creating widespread error in case adjudication. While practitioners have long recognized this possibility, empirical evidence on the downstream impacts of pretrial detention on misdemeanor defendants and their cases remains limited. This Article uses detailed data on hundreds of thousands of misdemeanor cases resolved in Harris County, Texas — the third largest county in the U.S. — to measure the effects of pretrial detention on case outcomes and future crime. We find that detained defendants are 25% more likely than similarly situated releases to plead guilty, 43% more likely to be sentenced to jail, and receive jail sentences that are more than twice as long on average. Furthermore, those detained pretrial are more likely to commit future crime, suggesting that detention may have a criminogenic effect. These differences persist even after fully controlling for the initial bail amount as well as detailed offense, demographic, and criminal history characteristics. Use of more limited sets of controls, as in prior research, overstates the adverse impacts of detention. A quasi-experimental analysis based upon case timing confirms that these differences likely reflect the causal effect of detention. These results raise important constitutional questions, and suggest that Harris County could save millions of dollars a year, increase public safety, and reduce wrongful convictions with better pretrial release policy.
Keywords: Pretrial Detention, Bail, Recidivism, Guilty Pleas
Suggested Citation: Suggested Citation
Heaton, Paul S. and Mayson, Sandra G. and Stevenson, Megan, The Downstream Consequences of Misdemeanor Pretrial Detention (2017). Stanford Law Review, Vol. 69, Pg. 711, 2017; U of Penn, Inst for Law & Econ Research Paper No. 16-18. Available at SSRN: https://ssrn.com/abstract=2809840 or http://dx.doi.org/10.2139/ssrn.2809840