A California bill replaces cash bail with risk-assessment algorithms, which critics argue will perpetuate the pre-trial detention of minority and low-income defendants.

Last August, California's former governor, Jerry Brown, signed Senate Bill No. 10, better known as the California Money Bail Reform Act. The legislation made the Golden State the first in the nation to end cash bail, or the practice of detaining defendants until their trials unless they are able to pay a bond ensuring their return. After signing the bill into law, the then-governor stated, "Today, California reforms its bail system so that rich and poor alike are treated fairly."

In place of bail, the act relies on risk assessments to determine which defendants are released pre-trial. Rather than universally releasing defendants on their own recognizance, the legislation mandates algorithm-based judgments, which incorporate both the defendant's previous criminal history and general criminal justice data to evaluate the supposed likelihood of their court attendance and their pre-trial criminal behavior. Accordingly, people deemed high-risk would be detained until their court dates—whether or not they can make bail.

That same day the bill was signed, the Bail Project, a national non-profit organization dedicated to freeing defendants from pre-trial detention, released a statement of its own on the passage of the California Money Bail Reform Act. "This is not bail reform," it read. "This is not a model to be followed by other states and it is no way to end the injustice of cash bail."

Brown's celebration and the Bail Project's condemnation of the California Money Bail Reform Act reveal a schism in the movement to end cash bail. Forces as disparate as politicians, non-profits, and think tanks all agree that bail must end. What should replace it, though, is anything but settled.

"It's a cautionary tale for why ending cash bail is not enough," says the Bail Project Chief Executive Officer Robin Steinberg of California's legislation. "True pretrial reform must be grounded in a vision and a plan for decarceration and racial and economic equity."

Steinberg in particular believes that algorithmic assessments to determine release are not a solution for reducing incarceration. "These algorithmic assessments are only as good as the data they use," Steinberg explains, "and criminal justice data will inevitably reflect the deeply imbedded racism and economic inequity that has driven mass incarceration in the first place."

In other words: Whereas defendants were condemned to pre-trial detention prior to the bill's passage due to their poverty and inability to make bail, they will now be in pre-trial detention due to their risk assessment—which will inevitably be based on their poverty, plus their race.

The Hamilton Project, an economic policy initiative from the centrist Brookings Institution, has similarly criticized the California Money Bail Reform Act. In "Proposals for Improving the US Pretrial System," a policy paper issued by the Hamilton Project last month, economic researchers Will Dobbie and Crystal S. Yang write:

The system is designed both to release low- and medium-risk defendants on their own recognizance or with the least-restrictive nonmonetary conditions, and to detain high-risk defendants without bail. However, what constitutes a high-risk defendant is largely left to the discretion of courts. ... In an extreme case, Senate Bill 10 may simply replace the money bail system with a system based on preventive detention.

The Hamilton Project's director, Jay Shambaugh, also criticizes risk-assessment tools along the same lines as the Bail Project.

"A danger of pre-trial risk-assessment tools is that, if the data fed into a statistical model is based on a system that is discriminatory, then the outputs will be discriminatory as well," Shambaugh says. "For example: If non-white individuals are stopped by police more often due to discrimination, they will have higher arrest rates even if they commit crimes no more frequently than white individuals. If prior arrests are included as part of a risk assessment, this might simply reinforce that discrimination."

But while agreeing that the California act is flawed, the Hamilton Project and the Bail Project don't see eye to eye on what should actually replace cash bail. In "Proposals for Improving the US Pretrial System," Dobbie and Yang suggest that it be replaced by a combination of behavioral nudges, such as text message reminders about upcoming court dates for low-risk defendants, and surveillance schemes, such as electronic monitoring, for high-risk ones. Risk assessment, in their estimation, must be based on transparent algorithms that encourage auditing and accountability. Even on the question of whether cash bail should be eliminated entirely, the authors hedge their bets, writing that "existing evidence does not yet support the wholesale elimination of the cash bail system." (For his part, Shambaugh has a charitable reading of Dobbie and Yang, saying that their paper "in no way advocates a continuation of the current system.")

The Bail Project, on the other hand, is less wavering in its "bail abolitionism"— a perspective informed deeply by its own experience buying people's freedom through paying their bail. Although the Bail Project was only founded in 2017, it has already freed over 4,000 people. (The organization also grew out of The Bronx Freedom Fund, which has been freeing hundreds of defendants from pre-trial detention each year since 2007.)

The Bail Project's process is simple: Using donations, it pays a defendant's bail; when that defendant returns to court, that bail money is returned to the organization—which it then uses to free more people. Because it recycles its funds, the Bail Project has a vested economic interest in ensuring that its defendants appear for their trials. Unsurprisingly, then, it also has a phenomenal track-record of getting people to court, with 96 percent returning for their trials.

How do they accomplish such feats without electronic monitoring or risk-assessment algorithms?

"We call our model 'community release with support,'" Steinberg explains, "a simple version of release on recognizance supplemented by effective court reminders, referrals to voluntary social services, and client advocacy."

Ultimately, the Bail Project advocates that cash bail should not be replaced by other systems that coerce defendants into showing up for court, but by methods that actually help them show up for court. The organization's vision for whom these methods should extend to—whether high risk or low risk—is straightforward as well. As Steinberg puts it: "We believe everyone—regardless of wealth, race, and, yes, accusation—should be entitled to the same presumption of innocence and treated with respect and dignity."

Related