The 'Science' of Bail Reform

By Chuck Dinerstein, MD, MBA — May 17, 2023
The Sixth Amendment states, "In all criminal prosecutions, the accused shall enjoy the right to a speedy and public trial….” In 2019, felony charges required a median of 424 days “to reach some final resolution.” Fourteen months in pretrial custody is not speedy. Bail “reform” seeks to end cash bail for certain crimes in the hopes of reducing pretrial jail populations. A new study tries to get past the sound bites of the media.
Image by Mohamed Hassan from Pixabay

In New York State, roughly two-thirds of individuals incarcerated in local jails are unsentenced. In 2019 that represented 12,933 individuals; bail reform reduced that number to 10,669 by 2021. In New York City, the incarcerated population dropped from 7,100 to 5,500. There can be little doubt that bail reform reduces pretrial jail populations.

Of course, that social and economic good must be weighed against the harm of increased criminal activity. Those increases may come from individuals released without bail reoffending – lack of incarceration allows them to commit additional crimes. Or increases may be due to bail reform reducing the impact of arrest as a “general deterrent” – incarceration while awaiting judgment increases the “cost” of crime to the individual – while bail reform may create a “sense of lawlessness.”

New York State’s bail reform initially took effect in January 2020, and the list of “bail eligible” offenses has been modified twice. In general, “the court must release defendants on their own recognizance unless there is a demonstrated risk of flight to avoid prosecution. If there is a risk of flight, the court must select the “least restrictive” condition(s) permissible to reasonably assure a defendant’s court appearance and compliance with court conditions.” You can find a more detailed breakdown here.  

The study, published in Justice Quarterly, looks at crime in New York before and after the institution of bail reform and the pandemic. The data consist of New York State’s monthly crime statistics for “seven index crimes (murder, rape, robbery, aggravated assault, burglary, larceny, and motor vehicle theft)” and similar FBI Uniform Crime Reports for the same crimes in the fifty states and District of Columbia during the same period. Consider a few of their findings:

  • A Siena College survey of 806 New York State voters “showed most respondents (58%) believed the bail law had increased crime. This belief was the majority or plurality response across all party, gender, race/ethnicity, and age groups.
  • A study comparing reoffending pre and post-expansion of Cook County’s pretrial release program found 16.7% of individuals reoffended before the reform and 17.8% after. Drilling down to violent offenses, 2.9% committed new violent crimes before and 3.2% after reform. The authors found this not to be statistically significant.
  • Bail reform increased the rate of murders by 0.1, larcenies [1] by 15.23, and auto thefts by 1.86 cases per 100,000 people per month. All were statistically significant. Bail reform did not impact the rates of rapes, robbery, burglary, or aggravated assaults.

The researchers created a “synthetic” control group. That is, they used FBI crime statistics from other states without bail reform for the same period to create a “weighted” control that considered the pandemic's impact on crime – lockdowns reduced some crimes and increased others. When comparing the New York crime statistics with their synthetic control, they concluded

  • “…that the rates of murder, larceny, and motor vehicle theft increased after the bail reform, but the contribution of bail reform to the crime rate increases was not [statistically] significant.”

Two quick points. Construction of a “synthetic” control group comes with many decisions, specifically which states you choose to use and how they are weighted. There is an opportunity for chicanery here, whether it is acted upon or not.

More importantly, as with many issues, statistical significance is less important than our perceptions and the salience of those perceptions in our life – increasing numbers of crimes make us feel less safe. There is a difference in the rate and number of crimes. This study reports on rates; the media reports numbers.

During the study period, rearrest rates in New York City were similar, 5% in 2019 and 4% in 2020. Because fewer individuals were released in 2020 than in 2019 (45,000 vs. 57,000), the “raw counts” of crimes did not increase; they decreased. Considering a subset of those “releasees” committing violent felonies, the reoffending rate diminished between 2019 and 2020, going from 4.5% to 4.2%. But in 2020, there were roughly 14,000 additional releasees meaning 345 additional violent felonies. As the authors write, whether these new violent offenses

“…can be considered as a negligible cost or the failure of incapacitation is still a question.”

Perhaps it does remain a scientific question, but the daily reporting of crimes provides a salience to individuals that all of the studies one can find may not alter. Optics are important when it comes to our feeling of personal safety. Stories outperform statistics. It is hard to follow the science when your media diet tells a different story.


[1] Wrongfully taking, obtaining, or withholding someone else's property with the intent to deprive them of its possession and benefit, either temporarily or permanently.


Source: Does Bail Reform Increase Crime in New York State: Evidence from Interrupted Time Series Analyses and Synthetic Control Methods Justice Quarterly DOI: 10.1080/07418825.2023.2209145


Chuck Dinerstein, MD, MBA

Director of Medicine

Dr. Charles Dinerstein, M.D., MBA, FACS is Director of Medicine at the American Council on Science and Health. He has over 25 years of experience as a vascular surgeon.

Recent articles by this author:
ACSH relies on donors like you. If you enjoy our work, please contribute.

Make your tax-deductible gift today!



Popular articles