Judicial Council group: Review of pre-trial tools is separate from bail debate

Recommendations to fix pretrial risk assessment tools presented at a Judicial Council meeting Friday included greater training for judges and further research on how to remove racial bias.

Appellate Justice Marsha G. Slough, who chairs the Pretrial Reform and Operations Workgroup set up by the council, told colleagues their work was completely separate from a larger debate about the statewide referendum that showed voters favored keeping cash bail.

"The pretrial pilot program is often conflated with SB 10," said the 4th District Court of Appeal justice. "Regardless of SB 10, the Judicial Council continues to have legislative responsibility to oversee this pretrial pilot program, which will continue to operate through June of 2021."

SB 10 is the 2018 law that would have phased out cash bail in the state. The bail industry, and many defense attorneys, challenged it through the referendum process and ultimately won, defeating Proposition 25 by a 12 point margin.

Slough said had voters approved the proposition, it would have been "a huge sea change." But it would have had little effect on the project that began when Gov. Gavin Newsom signed AB 74 last year. This vast omnibus budget bill included $75 million to study risk assessment tools.

"It is a stand-alone, limited term project with 17 pilot courts from throughout the state," Slough said. "The pretrial program is and has been required to operate under existing law from its inception."

Risk assessment tools ended up playing an outsized role in the Proposition 25 campaign. The bail industry, law enforcement groups and conservative organizations generally opposed SB 10. But these foes could also point to many civil rights groups and public defenders who were leery of the changes under SB 10 or actively worked against the measure because of the fear that they could further entrench racial bias in pretrial release decisions.

Another 2019 law, SB 36, requires the council to come with standards by the end of this year that can evaluate pretrial risk assessment tools for racial bias. But as the report issued Friday noted, real world factors make eliminating bias difficult.

One method courts have used in the past is to look at risk assessment scores and compare them to later arrests. But the report cited research showing people of color were rearrested more often even when the assessments conclude they are not necessarily more likely to engage in further criminal activity.

The report recommends instead comparing these scores to later conviction rates and possibly discarding data from some rearrests for minor charges. It also cites the need to continue updating risk assessment tools for bias.

Meanwhile, the working group behind the report stressed the need for better training to help judges "identify appropriate circumstances in which to follow the recommendation of the tool versus depart from it and to avoid departures which may inadvertently result in bias."

This training could also help the state become more successful in using pretrial services and developing more accurate risk assessment scores. The report recommended against using tools that give a suspect a single numerical score. Instead, tools should provide a more nuanced view of a particular person to help a judge match them up with the right type and amount of supervision that could allow them to stay out of custody, the group suggested.

The report also explicitly recommended omitting mental illness from risk assessment. The tools currently under use in counties around California do not use mental illness to test risk. But many outside the state do.

"Many studies have found that the presence of a mental illness does not necessarily increase a person's likelihood of pretrial failure," the report found.