Chicago Judge Says His Bail Reforms Were a Success. But Independent Reviews Show Flaws and More Crimes.

Did Cook County overdo it and let too many dangerous defendants free?


Chicago's bail reforms may not have had the rosy outcomes indicated by a top county judge's analysis, which independent researchers say is downplaying the new crimes that have resulted from allowing defendants to await trial outside of jail.

Those are the results of an analysis by a group of Chicago Tribune reporters in a new investigative piece as well as a just-published data analysis paper by University of Utah professors Paul Cassell and Richard Fowles.

In 2017, Cook County Chief Judge Timothy Evans implemented an order reforming how the Chicago area courts handled pretrial detention. The goal was to reduce the demands for cash bail, which tend to keep people trapped behind bars on the basis of poverty rather than risk. Cook County met its goal of detaining fewer defendants before their trials. The number of defendants who secured pretrial release between 2016 and 2018 jumped from 71.6 percent to 80.5 percent. When cash bail was ordered, the amount demanded was much lower than before. Cook County's jail population dropped from 7,443 to less than 6,000.

Last May, Evans released a report that showed releasing more defendants from jail did not put the community at greater risk of crime. A high proportion of defendants (83 percent) charged with felonies and released under the new system returned to court as ordered and did not commit new crimes while released. In all, Evans' report painted a positive picture that matched the narrative of those who support bail reform: That court systems in Cook County were accurately sorting defendants based on the risk they posed to public safety and their likelihood of showing up to trial, rather than simply leaving everyone in jail simply because they couldn't afford to pay what the courts ordered.

But further examination of Evans' data paints a less rosy picture. Last week, The Chicago Tribune reported that Evans' report left out hundreds of violent crime charges filed after the bail reforms were implemented. The reporters say he did this by including certain violent crimes (murder, attempted murder, non-negligent manslaughter, forcible rape, robbery, and aggravated battery) and excluding incidents like domestic violence, assault with a deadly weapon, battery, reckless homicide, and others. If Evans' report had included all these other crimes, the Chicago Tribune calculates the number of violent crimes allegedly committed by released defendants would jump from 147 to 578. The largest chunk of these charges—231 of them—were for domestic battery.

Furthermore, Evans' report stated that only three defendants who had been released under the new pretrial system had subsequently been charged with homicide. But the Tribune identified 21 defendants accused of murder who had been released during the 15 months of bail reform the report reviewed. Their exclusion from Evans' report is supposedly a result of incomplete records and some odd reporting decisions like only counting the first new charge a defendant received after being released (two of the defendants were arrested for another charge, then released, and then allegedly killed people); or not counting them because their initial charges weren't felonies (five murder defendants had been bonded out on misdemeanor charges).

In a separate review, Cassell and Fowles reanalyzed Evans' data and found other problems.

For one thing, there's a significant flaw in how Evans measured new crime charges prior to his bail reforms and afterward. When calculating the crime rate, Evans' report evaluated the "before" defendants for an average of 243 days and the "after" defendants for an average of just 154 days. This is a significant methodological problem because reducing the time frame in the post-reform evaluation gives these defendants less time to commit new crimes. Cassell and Fowles argue that this difference of nearly 100 days may well mean that, in actuality, the post-reform crime rate among those released might be even higher. Cassell and Fowles' report observes, "the second group will, other things being equal, undoubtedly commit fewer additional crimes simply because they have had less time to commit such crimes."

The two attempt to estimate what the crime rate might actually be if the report monitored the post-change pool for the same time frame. It's a challenging calculation, they note, because they couldn't find any studies showing month-to-month re-arrest rates among those released pretrial. So they used some modifications in stats from the Bureau of Justice Statistics for recidivism rates among those who have been released from prison, combined with some pretrial recidivism rates from Cook County's data. They conclude that in all likelihood, Cook County's report undercounted new crimes committed by released defendants by about 1,200. When they correct for the time frame, Cassell and Fowles estimate that there was actually a 45 percent increase in the number of new crimes caused by defendants who had been released.

It's important to make it clear that this is a mathematical model, and Cassell and Fowles aren't specifically detailing a bunch of concrete new crimes that have been committed by these defendants. But part of the problem here is that the court has been reluctant to share the data Evans used with Chicago Tribune reporters, which required the newspaper to file a petition with the Illinois Supreme Court. Evans has since agreed to share his data with the newspaper.

Cassell and Fowles write that their goal is not to kill off bail reforms or scare courts away from implementing them. Rather, they are concerned about biases in self-analysis that "always lurks when an entity implementing reform later studies whether that reform is successful. In this case, it appears that many dangers stemming from the court's expansion of pretrial release were not carefully assessed by the court's own subsequent study."

These kinds of independent assessments are extremely valuable in part because these reforms are still relatively new and they make a number of people very, very nervous. Part of that fear results from deliberate scaremongering by those who have a financial or political stake in protecting a harsh status quo, like bail bond companies and jail officials.

But as the Cassell and Fowles report notes, poorly managed pretrial reforms can backfire and cause additional harms. If you agree, for example, with the argument for reform—that it's a violation of a person's rights to keep them locked up before they're convicted only because they cannot pay bail—you must also consider the risk they pose to the rights of other people if they are released before trial. In a city like Chicago, it is mostly poor people whose rights are violated by the bail requirement and mostly poor people whose rights are violated by the defendants who commit additional crimes before their trials.

If, on the other hand, you make a utilitarian argument that keeping people locked up because they're too poor to pay bail but aren't dangerous is much more expensive than letting them return home, Cassell and Fowles note that the cost-benefit analysis changes if the person commits new crimes before trial. While there are established financial harms to pretrial detention (lost jobs and housing) and established benefits to letting them out to continue to work and care for families, the economic impact of a homicide wipes out the financial benefits of letting more people out of jail.

At some point, the cost of new crimes committed by a percentage of defendants free before trial financially outweighs the savings of freeing people who aren't dangerous. Data-driven bail reform is supposed to prevent a few bad defendants from spoiling it for everyone, yet Cassell and Fowles argue that it does not appear to be working as intended in Chicago: "Given equal weight to the benefits the pool of such defendants receive when compared to the costs inflicted on victims seems dubious."

The report ends not trying to bash reforms but warning that court systems need to really explore the impact of pretrial release data and make sure they're not perpetuating new harms: "To be sure, such pretrial release reforms can have significant benefits. But only if both benefits and costs are accurately measured can a sound decision be made about which way the scales tip and whether the 'reform' was truly an improvement."

Cassell wrote about the report's release over at The Volokh Conspiracy, hosted here at Reason.