We filed comments with the Judicial Council of California on two of its proposed new court rules on the use of pretrial risk assessment instruments. We argued that the proposed rules on how courts use pretrial risk assessment tools need significant modifications in order to be constitutionally defensible and to protect civil rights.
Re: Proposed California Rules of Court 4.10 and 4.40
On behalf of Upturn, The Leadership Conference on Civil and Human Rights, The Leadership Conference Education Fund, The NYU Center on Race, Inequality, and the Law, AI NOW, and Medial Mobilizing Project, we respectfully submit these comments regarding the Judicial Council of California’s proposed California Rules of Court 4.10 and 4.40. We appreciate the opportunity to engage with the Judicial Council on these critically important issues. We believe these rules must be substantially modified, as described further below, in order to maximize their chances of being constitutionally defensible and achieve the aims of bail reform.
This summer, a broad coalition of civil rights, digital justice, and community-based organizations jointly published “The Use of Pretrial ‘Risk Assessment’ Instruments: A Shared Statement of Civil Rights Concerns.” That statement argues that jurisdictions should not use risk assessment instruments in pretrial decision-making, and instead focus their limited resources on other strategies to strengthen the presumption of release and meaningful hearings for all defendants. At the same time, the statement recognizes that many jurisdictions are pursuing risk assessment instruments and offers principles and guidance on how these tools can be made least harmful to civil rights. Our comments on the proposed rules are guided by the principles set forth in that document.
The text of S.B. 10 is disheartening, particularly in light of the public discussion that preceded its passage. The now-enacted law lacks critical safeguards that were included in earlier versions of the bill, safeguards that community advocates impacted by mass incarceration rightfully saw as necessary. As the Committee develops these rules and future rules, it must revive and mandate these much-needed protections.
Comments on Rule 4.10
Proposed Rule 4.10 misstates the purpose of pretrial risk assessment information, inaccurately describes proper uses, and inadequately describes improper uses. It also authorizes statistically irresponsible and potentially unconstitutional uses of risk assessment scores. If pretrial risk assessment instruments are to be used at all, the only legitimate purpose they can meaningfully serve is to identify which people can be released immediately and which people are in need of non-punitive or restrictive services. Both subsections (b) and (c) of the proposed rule are missing critical protections.
Proposed Rule 4.10(a) misstates the purpose of pretrial risk assessment information.
Given the state of today’s instruments, pretrial risk assessment information cannot responsibly be used to justify detention decisions. These instruments provide fairly limited information: the likelihood of certain outcomes, such as rearrest for any reason or failure to appear, under the condition where no assistance is provided to the accused. At most, pretrial risk information may help Pretrial Assessment Services (PAS) or a court quickly determine whether an individual is such a low risk that they should be released immediately, and to identify the smaller set of people who may benefit from additional supervision.
Today’s risk assessment instruments make no attempt to measure the extent to which supportive services can reduce risk. Nor do they contemplate the ways that risk may be modified as a result of supportive, supervisory conditions that a judge, state-level policy, or court-rule may impose in a particular case. And yet, it is precisely this modified level of risk — the level of risk after supportive services or supervisory conditions are taken into account — that the judge must use as the basis for a detention decision.
Indeed, the relevant constitutional and legal question for detention decisions is whether there are no conditions that could reasonably assure an individual’s appearance or the safety of the community. Given that risk assessment tools do not consider the effect of such conditions, their purpose should not be to “assist Pretrial Assessment Services and the court to make release and detention decisions,” but rather to “assist Pretrial Assessment Services and the court to make release and supervision decisions.” Detention decisions must be rendered following an appropriate, individualized, adversarial hearing.
Finding that the defendant shares characteristics with a collectively higher risk group is the most specific observation that risk assessment instruments can make about any person. Such a finding does not answer, or even address, the question of whether detention is the only way to reasonably assure that person’s reappearance or the preservation of public safety. That question must be asked specifically about the individual whose liberty is at stake — and it must be answered in the affirmative in order for detention to be constitutionally justifiable.
This is not only the position of civil rights groups. The developer of one of the most popular pretrial risk assessment tools, the Laura and John Arnold Foundation, agrees: “The foundation “encourages [jurisdictions] to understand the [relevant state] law in order to use the [Public Safety Assessment] in the most legally sound manner; this necessarily includes holding a due process hearing prior to intentional pretrial detention.”
Proposed Rule 4.10(b)(1) insufficiently specifies proper use of pretrial risk assessment information and would not lead to uniform use.
Proposed Rule 4.10(b)(1) requires PAS and courts “give significant weight to the risk assessment score,” but does not detail what “significant weight” means. “Significant” can have a range of definitions: for example, it can mean “having meaning,” “having or likely to have influence or effect,” “large enough to be noticed or have an effect,” or “very important.” This vague phrase would allow courts or PAS to use pretrial risk assessment information as they wish, so long as the information is not “determinative,” or the “sole basis” for a decision regarding detention, release, or conditions of release. Moreover, given that proposed rule 4.10(b)(1) states that both courts and PAS “must give significant weight to the risk assessment score,” proposed rule 4.10(b)(3)’s admonition that the “risk score is not determinative but is a relevant factor” is confusing. Ultimately, the proposed rules provide inadequate guidance as to how the PAS and courts should understand and use pretrial risk assessment information. The shifting standards only compound the vagueness.
In the absence of clear, direct guidance, PAS and courts may use pretrial risk information in an inconsistent way, based on what best suits them in each individual case. This is a terrible result, and one that belies the “objectivity” and “evidence-based” processes promised by risk assessments in pretrial decision-making. In such a world, the Judicial Council may find that PAS and courts have low concurrence rates, diverging from the recommendations of pretrial risk assessment instruments and their associated decision-making frameworks.
Further, 4.10(b)(1) should be rewritten to reorder what the court and PAS must consider. Currently, as drafted, “the rights of the defendant” is the third item that either actor must consider, behind the “safety of the public” and the “safety and rights of the victim.” The rights of the defendant must be considered first, given the unique constitutional protections afforded to the accused.
Ultimately, pretrial risk assessment instruments must be implemented in ways that reduce and eliminate unwarranted racial disparities across the criminal justice system. Without clearer language on how PAS and courts may consider pretrial risk assessment information to further ensure that the vast majority of individuals are released, the proposed rule will not have an ameliorating effect and may actually lead to worse outcomes.
Proposed Rule 4.10(b)(3) authorizes statistically irresponsible and potentially unconstitutional uses of risk assessment scores.
Proposed Rule 4.10(b)(3)(A)
This proposed rule implies a 4. As a result, it is neither responsible nor constitutional, and must be struck from the list of permissible uses for risk scores.
Risk scores were not designed, and cannot responsibly be used, to inform detention decisions. As noted above in our comments on Proposed Rule 4.10(a), these scores make no attempt to measure the only risk that matters to a detention decision — namely, the modified risk that remains after supportive services are provided or supervisory conditions are imposed. The tools instead measure the ex ante risk before any risk management is undertaken, which is relevant information when setting conditions of release but does not speak to the basis for detention decisions.
Moreover, by S.B. 10’s own terms, the law “create[s] a presumption that the court will release the defendant on his or her own recognizance at arraignment with the least restrictive nonmonetary conditions that will reasonably assure public safety and the defendant’s return to court.” As such, the subsection should not contemplate using a risk score to inform the choice of “[w]hether” a defendant can be released. The presumption of release commands otherwise.
Instead, the rule should recognize that risk scores can only be a relevant consideration in determining “how” the court and PAS can ensure or promote release in the least restrictive ways possible. In order to ensure that pretrial detention remains the “carefully limited exception,” the Judicial Council should reference Principle 3 from the Shared Statement of Civil Rights Concerns regarding the minimally necessary procedures and safeguards to determine when detention must be used as a last resort.
The only risk information that is relevant in order for PAS or a court to respond to a violation of a condition of release is that individual’s risk of continued, repeated violation of that same condition. PAS or a court may able to determine that risk through other means, but risk scores are not applicable to this specific decision and should not be used.
Further, S.B. 10 commands that, “the results of a risk assessment using a validated risk assessment instrument, shall not be used for any purpose other than that provided for in this chapter.” Given that S.B. 10 does not provide for how PAS or the Court should respond to violations of conditions of release, neither PAS nor courts may permissibly use pretrial risk assessment information in response to violations.
Proposed Rule 4.10(b)(5) must more clearly communicate the limitations of risk assessment tools.
It is important that the Judicial Council will direct courts to consider the limitations of pretrial risk assessment tools. However this proposed rule, as written, does not do an effective job at communicating these limitations. Further, the draft rule does not describe other important limitations that courts should consider. Both deficiencies must be addressed in ways we describe below.
Proposed Rule 4.10(b)(5)(A)
Proposed Rule 4.10(b)(5)(A) rightly encourages courts to consider the limits of the data upon which pretrial risk assessment instruments were developed. But it stops short of describing two critical limitations: first, that the group data reflects decades of massive racial inequities within the criminal justice system and second, that the group data does not reflect the risk-mitigating effects of new bail reforms.
The Proposed Rule should be explicit about racial inequities reflected in the data upon which pretrial risk assessment instruments are based. Our system of justice is profoundly flawed: it is systematically biased against and disproportionately impacts communities of color. (Indeed, the need to address this reality was a motivating factor behind earlier versions of S.B. 10.) This fact is reflected in historical data: Critically, decades of research have shown that arrest data primarily documents the behavior and decisions of police officers and prosecutors, rather than the individuals or groups that the data are claiming to describe. The Committee must speak much more clearly about the limitations arising from this biased data.
Further, the Committee should make clear that the group data upon which pretrial risk assessment instruments are developed do not consider the risk-mitigating effects of other bail reform policies. For example, text message reminders and telephone check-ins have been shown to help reduce the rate of failure to appear. Indeed, there is an inherent, basic tension between statistical prediction and other bail reform efforts. On the one hand, to change a broken system, policymakers enact and implement new procedures that work to support individuals and reduce the risk of rearrest and failure to appear. On the other hand, policymakers ask statistical tools to forecast those very same risks based on data from the very system under reform. In other words: the development sample or training data for pretrial risk assessment tools often comes from times and places that are materially different from the ones where predictions are made. Without the right policies and procedures, pretrial risk assessment tools are blind to the helpful impact of new risk-mitigating policy changes and may consistently overestimate risk.
Given this, the committee should ensure that Rule 4.10(b)(5) instructs courts to consider the following limitation of pretrial risk assessment information: that pretrial risk assessment instruments do not consider the risk-mitigating impact of certain, least-restrictive conditions of release.
Proposed Rule 4.10(b)(5)(B)
Proposed Rule 4.10(b)(5)(B) should not exist because it should not be necessary. Though outside the scope of these rules and proceedings, no proprietary pretrial risk assessment tool should be used in the state of California. We strongly urge the Judicial Council to ensure that when the Council “[c]ompiles and maintain[s] a list of validate pretrial risk assessment tools,” it does not include a tool whose maker makes claims of trade secrecy or alleges that any relevant part of the tool is proprietary. At a minimum, the public, accused, and counsel must be given a meaningful opportunity to inspect how a pretrial instrument works. This includes access to the outcome data used to develop and validate a tool; a complete description of the design and testing process used to create the tool; and all input factors and weights of those factors.
There is simply no defensible reason to use tools that cannot be completely audited within a system that has such substantial impact on human lives.
Proposed Rule 4.10(b)(5)(C)
Proposed Rule 4.10(b)(5) indicates that a court must “consider any limitations of risk assessment tools in general,” and Proposed Rule 4.10(b)(5)(C) further specifies that a court must consider whether “any scientific research has raised questions that the particular instrument unfairly classifies offenders based on race, ethnicity, gender, or income level.” It is incredibly important for the Judicial Council to speak clearly and cautiously on this matter. As the Judicial Council wades into the waters of whether or not pretrial risk assessment instruments “unfairly classif[y]” offenders, it must tread carefully.
Research has shown that if the base rate of a predicted outcome (here, rearrest) differs across racial groups, it is impossible to achieve predictive parity, parity in false-positive rates, and parity in false-negative rates. In other words, so long as the base rate of the predicted outcomes is racially imbalanced, it is impossible to both ensure that risk scores mean essentially the same thing regardless of a person’s race, and to ensure equal false positive and false negative rates across racial groups. Computer scientists have detailed mathematical proofs of this fact. When base rates of rearrest differ, “race-neutrality” is not attainable.
Given this research, the correct consideration for a court is not whether “any scientific research has raised questions” about whether a particular tool “unfairly classifies offenders.” Instead, this observation applies to all pretrial risk assessment instruments. Predictive parity and error rate balance — both of which are important measures of fairness in their own respects — are mutually exclusive so long as the base rate of rearrest is unequal.
Proposed Rule 4.10(b)(5)(D)
It is difficult to comment on Proposed Rule 4.10(b)(5)(D) given that the Judicial Council will “[d]escribe the elements of ‘validation’” in another rulemaking process, which will, in all likelihood, define what a “relevant population” means. We look forward to seeing and offering comments on the Judicial Council’s forthcoming rules regarding the elements of validation. Nevertheless, similar to Proposed Rule 4.10(b)(5)(B), Proposed Rule 4.10(b)(5)(D) should not exist because the Judicial Council should not allow a pretrial risk assessment instrument to be used if it is not validated on a relevant population. The validation of a pretrial risk assessment instrument on a relevant population is a necessary, not a sufficient condition of use. Using pretrial risk assessment tools that are not validated on relevant populations would contradict principles 4 and 6 from the Shared Statement of Civil Rights Concern.
Several Necessary Protections are Missing from Proposed Rule 4.10(b)
Add a new provision regarding the gap between predicted outcomes and legally significant outcomes.
The Judicial Council should add a provision to Rule 4.10(b)(5) describing one of the central limitations of pretrial risk assessment instruments: the gap between the outcomes that pretrial risk assessment tools predict (i.e., rearrest or failure to appear), and those that courts must assess (i.e., public safety and flight risk). In both instances, pretrial risk assessment instruments use what can be measured to speak to what risks courts may care about. Critically, the two concepts are not the same.
Pretrial risk assessment tools do not speak to “public safety.” They in fact predict something much more limited: rearrest. Though the predicted outcome of rearrest is sometimes treated as a proxy for a person’s public safety threat, its use as a measure is imperfect at best and dangerous at worst. To the extent that the relevant concern is preventing violent crime, risk scores that reflect the general probability of rearrest are counterproductive: individuals that are at the highest risk of any rearrest are not at the highest risk of arrest for violent crime. Worse, a focus on generalized rearrest as a proxy for public safety will unnecessarily compound racial disparities in prediction, given that arrests rates for low-level offenses and drug offenses are racially biased.
Similarly, while pretrial risk assessment tools predict future nonappearance in court, generalized nonappearance risk and flight risk are distinct. In most cases, nonappearance risk does not involve people who have the means or the desire to flee a jurisdiction, and is best addressed by other means than detention. As one scholar has argued, there are three “subcategories” of “nonappearing defendants”: true flight, local absconders (those who remain in the jurisdiction but actively and persistently avoid court), and low-cost appearances (defendants who remain in the jurisdiction but whose failures to appear are easily preventable and/or non-willful). Risk scores do not capture these nuances, and courts should be warned of this fact.
To the extent that Rule 4.10(b)(5) directs courts to consider the limitations of pretrial risk assessment instruments, the Committee should draft language that clearly distinguishes between the predicted outcomes that pretrial risk assessment tools can measure on the one hand (rearrest and failure to appear), and the only risks courts may consider to order detention on the other (prevention of violent crime and flight from the jurisdiction).
Add a provision directing PAS or the courts to consider the costs that supervision or jailing impose on defendants and communities.
The Committee should add a provision to 4.10(b)(1), which describes what information PAS or a court must consider, Instructing the PAS or court to consider the negative effect that detention can have on a person’s life.
A large body of scholarship shows that being sent to jail increases the likelihood the accused will plead guilty, even after controlling for the seriousness of the case and other factors. While studies recognize that short-term detention does incapacitate individuals from re-offending pending adjudication of their case, the literature also finds that “pretrial detention increases new crime after case disposition through a medium-run criminogenic effect.” On net, a careful econometric study and a separate thorough literature review have found, the public safety gains from pretrial jailing are nullified by the added risk of crime that is created by sending someone to jail.
As one study in New York observed, most of the reduction in pretrial rearrests were “reversed as individuals who were detained pretrial are more likely to be rearrested after their cases are resolved.” Relatedly, another study not only finds that pretrial detention reduces employment, especially for individuals who had the strongest ties to the labor market, but also reduces the likelihood of receiving unemployment insurance and credits under the Earned Income Tax Credit for wages earned while incarcerated.
Given that proposed Rule 4.40(b)(3)(D) requires, via local rule, the PAS to consider the impact of detention on an individual assessed as medium risk pending arraignment, the Committee should ensure all individuals potentially subject to detention enjoy this crucial consideration by amending 4.10(b)(1).
Add a provision on how pretrial risk assessment information is communicated.
In accordance with basic concepts of fairness, the presumption of innocence, and due process, pretrial risk assessment instruments must frame their predictions in terms of success upon release, not failure.
Proposed Rule 4.10(c) does not adequately describe improper uses of pretrial risk assessment information.
Proposed Rule 4.10(c)(1)
We are heartened to see that the draft rule says that PAS and the court “must not use the risk score as the sole basis to detain … a person,” other than as required by other troubling provisions of S.B. 10.
Nevertheless, the Committee must be more assertive in protecting against improper uses. On this point, we would point the Committee to a recent decision by the New Jersey Supreme Court. There, a previous court rule directed that:
“The Court may consider as prima facie evidence sufficient to overcome the presumption of release a recommendation by the Pretrial Services Program … that the defendant’s release is not recommended (i.e. a determination that ‘release not recommended or if released, maximum conditions.)” (emphasis added).
But the New Jersey Supreme Court recently “conclude[d] that parts of the Rule should be revised,” because part of the original rule “suggests a court can order detention based solely on a recommendation against release.” As Justice Albin noted in his concurrence, the recommendation, based on pretrial risk assessment information, “could have operated to undermine the rebuttable presumption favoring release.”
The Committee should update this rule to reflect the New Jersey Supreme Court’s wisdom: Pretrial risk assessment information should never on its own constitute prima facie evidence sufficient to overcome the presumption of release.
Proposed Rule 4.10(c)(2)
Where PAS or the court seek to manage the risk of reoffense during the pretrial stage, they can only do so “during the pretrial stage of the case,” and cannot consider “the long-term risk of reoffense.” This rule correctly prohibits courts and PAS from considering “long-term risk of reoffense” when evaluating public safety.
But ensuring that PAS and courts cannot consider the “long-term risk of reoffense” should not bar PAS or courts from considering the longer-term effect of detention on the individual. That’s because the effect of interest would be directly caused by state action against an individual’s liberty. As we described in our general comments on Proposed Rule 4.10(b), the “available empirical evidence suggests that pre-trial detention is indeed criminogenic, imposing long-term costs on society.” It is well within the proper boundary of inquiry to understand the consequences of a state-imposed sanction on a presumptively innocent individual. The Committee should not blind PAS or courts to these very real consequences.
Separately, we note that different pretrial risk assessment tools work on different time horizons, which could create confusion about what counts as “long-term” for purposes of this rule. For example, the Ohio Risk Assessment System defines recidivism as any “arrest for a new crime” within one year from the date of assessment, whereas COMPAS defines recidivism as any arrest within two years of the assessment, whereas the Los Angeles Countywide Criminal Justice Coordination Committee defines recidivism within a three-year period following the individual’s release. Given variation between how tool developers define “the risk of reoffense during the pretrial stage of the case,” as well as variations amongst local jurisdictions, the Judicial Council should direct that judges be explicitly notified of the precise time horizon over which predictions are being made.
Proposed Rule 4.10(c)(3)
We note that this proposed rule stands in tension with the Committee’s proposed Rule 4.10(b)(5)(B). Were the Judicial Council to allow the use of a proprietary pretrial risk assessment instrument — which, to reiterate, we strongly oppose — then, it would likely follow that a judge would not be able to “be familiar with the factors included in the particular risk assessment tool,” or the weights of those factors.
Further, as Proposed Rule 4.10(b)(5)(B) describes, if a proprietary tool would “prevent the disclosure of information relating to how [the tool] weighs risk factors,” a judge would have a Herculean task of understanding how to not “give additional undue weight to [those] factors,” given that a proprietary tool would likely not allow them to understand the tool’s weighting of those factors. The judge would not have a baseline to work from in order to ensure that they did not place “undue weight” on the factors or combination of factors present in the tool.
To resolve this conflict, we believe the Judicial Council must ensure that no proprietary pretrial risk assessment instrument be used in California.
Several necessary protections are missing from 4.10(c)
Add a rule regarding subsequent use of pretrial risk assessment information.
The Criminal Law Advisory Committee should add a provision in Rule 4.10 prohibiting subsequent, downstream use of pretrial risk assessment information, bringing it in line with the clear requirements of S.B. 10’s text. It would make most sense for the Committee to do so in 4.10(c).
The plain language of § 1320.9(d) states that “The report described in subdivision (c), including the results of a risk assessment using a validated risk assessment instrument, shall not be used for any purpose other than that provided for in this chapter.” (emphasis added). Given that Rule 4.10 is about prescribing the “proper use of pretrial risk assessment information,” the Committee should consider adding the following language: “Pretrial risk assessment information shall not be used in any court proceeding or criminal justice decision point, other than pretrial.” This would prevent pretrial risk assessment information from improperly being used at sentencing — where, for example, the time-horizon of a forecasted prediction may no longer be applicable.
Add a rule clearly allowing for independent study of pretrial risk assessment information.
The Committee should be sure to add a provision which allows for independent third-party testing for predictive bias, and for disparate outcomes in implementation.
Comments on Proposed Rule 4.40
As an initial matter, it is difficult to provide detailed comments to the Committee without knowledge of how many people will be assessed as medium risk. If a significant number of people are likely to be assessed as medium risk, then the Committee must take special effort to ensure there are heightened requirements for local rules to add exceptions. In any case, it is improper to ask local courts to develop local rules regarding review and release standards for individuals assessed as medium risk without knowing what “medium risk” means. Therefore, the Council should consider delaying this rule until the risk level definitions are established.
Further, Proposed Rule 4.40 is missing critical protections for the setting of release conditions, the considerations for expanding the list of exclusions, and the requirements of courts in the development and review of local rules. Individuals assessed as medium risk should not, because of their pretrial risk assessment score, lose their presumption of release on recognizance.
Proposed Rule 4.40(a) must further emphasize the presumption of release.
Proposed Rule 4.40(a)(2)
The Committee should revise this rule in two ways.
First, the opening clause of the rule should reads as follows: “Each local rule must clearly authorize release for as many arrested persons as possible.” Second, the Committee should add a sentence to this proposed rule that clarifies that each local rule must clearly state that individuals assessed as medium risk enjoy the presumption of release on their own recognizance with the least restrictive nonmonetary conditions that will reasonably assure public safety and the defendant’s return to court.
Proposed Rule 4.40(b) must clarify the criteria and methods allowed to influence decisions about pretrial release.
Proposed Rule 4.40(b)(3)
Subsection (b)(3) must clarify and provide definitions for the list of factors to be considered by PAS. This should include clear definitions as well as explanations of the sources of information that will be accepted for consideration, including the types of pre-arraignment interviews. No one should be required to provide information to an employee of a law enforcement agency in order to be eligible for pretrial release.
Based on the text of S.B. 10, the Judicial Council’s own materials, and the Committee’s draft rules, it is unclear how PAS will accomplish Proposed Rule 4.40(b)(3)(B), which calls for PAS to consider an arrested person’s “family and community ties.” Though § 1320.9(a)(3) directs PAS to obtain “[a]ny supplemental information reasonably available that directly addresses the arrested person’s risk to public safety or risk of failure to appear in court as required,” it remains unclear how PAS will accomplish this.
Here, it is difficult to comment because we do not know what procedure or set of procedures will be used to gather information regarding an individual’s family and community ties. Specifically, it is important for the Committee and the Judicial Council to provide clearer guidance here. Are PAS expected to interview an individual about their family and community ties? It is important for this information to be available and considered, but PAS should not interview clients who have not otherwise been appointed counsel.
Further, this information should only be used in a way that ensures release. Individuals should not be penalized if, for example, they have lost touch with their family or they just moved to California.
As to the consideration of “criminal history” in Rule 4.40(b)(3)(B), which the Committee requested specific comment on: yes, the history considered should exclude arrests that did not result in the filing of charges. Moreover, in specifying that PAS consider additional factors including “criminal history, and record concerning appearance in court proceedings”, appears to be contradicting the intent of Rule 4.10 (c)(3) which specifies that additional undue weight should not be given to factors already considered by the risk assessment tool.
Finally, a new subcategory should be added to Subsection (b)(3) that prioritizes consideration of the impact on dependent minor children of detainee parents.
Proposed Rule 4.40(b)(4)
Subsection (b)(4) is too vague with many undefined terms. The phrase “substantial likelihood” is vague, potentially investing too much discretionary power in PAS to retain an arrested person in custody. The proposed rules must define how PAS or a judge must make the determination of “substantial likelihood.” Does “substantial” mean a 5% likelihood? 20%? 50%? Without clearer language, PAS may interpret this standard in radically different ways. In order for these local rules to “authorize release for as many arrested persons as possible,” the Committee must establish a higher, clearer bar for local rules. Without guidance or limitation, it remains a vague and subjective standard.
Proposed Rule 4.40(c) misses an important opportunity to limit the proliferation of intrusive conditions of release.
This subsection must better define rules for setting conditions of release. Although there is a presumption that a person will be released under the least restrictive nonmonetary conditions, and individuals cannot be required to pay for any supervision conditions that are imposed, the “least restrictive” standard needs to be clearly defined. The rules should impose some defined limits on when the most severe conditions are allowed and establish a process for modifying conditions of release based on progress and updated circumstances.
Proposed Rules 4.40(c)(1)-(4)
We support the Committee’s draft rules 4.40(c)(2) and 4.40(c)(4). The Committee should consider slightly revising 4.40(c)(4) from “an undue burden” to “any undue burden” to strengthen its protections.
In regards to 4.40(c)(1), PAS is empowered to “exercise independent judgment and to tailor release conditions to the individual person,” essentially playing the role of judge or magistrate without the training or democratically bestowed authority.
We believe Proposed Rule 4.40(c)(3) misses the mark and potentially creates confusion regarding the standard to apply in determining potential conditions of release. PAS must only impose the least restrictive conditions that are reasonably related to assuring the individual’s return to court and public safety, a more specific requirement than simply conditions that are reasonably related to those aims. As noted in The Use of Pretrial ‘Risk Assessment’ Instruments: A Shared Statement of Civil Rights Concerns, a hearing must be held promptly to determine whether there is a credible danger to specifically identified individuals in the community.
Given that proposed rule 4.40(c)(6) already has this stronger language, we believe rule 4.40(c)(3) is unnecessary, and current proposed rule 4.40(c)(6) could take its place.
Proposed Rule 4.40(c)(5)
The Committee should require local rules to set a significantly higher bar for a finding that GPS or electronic monitoring is a necessary condition of release. Specifically, the Committee should require that local rules set a higher bar for making transdermal monitoring (4.40(c)(5)(J)) and passive or active GPS monitoring (4.40(c)(5)(K)) a condition of release. For both, there should be a presumption against their use. Both should only be allowed to be used after a specific, written finding has been issued as to why these conditions satisfy the “least restrictive condition” test articulated in proposed rule 4.40(c)(6).
As the Committee considers a higher standard for active or passive GPS monitoring, it should be borrow heavily from the Guidelines for Respecting the Rights of Individuals on Electronic Monitoring, a set of guidelines which more than 50 organizations signed onto. The Committee should also take note of New Jersey’s experience with electronic monitoring before trial. The New Jersey Judiciary told the state that, “[e]lectronic monitoring has . . . been a significant challenge.” Given electronic monitoring’s especially pernicious restraints and effects on the individual, electronic monitoring must only be issued in rare, specific circumstances.
Proposed Rule 4.40(c)(6)
This subsection should be clarified so that it is explicit that PAS must justify any conditions imposed by their necessity to facilitate return to court and guard against a credible danger to specifically identified individuals in the community, and that punitive/rehabilitative conditions are inappropriate.
Proposed Rule 4.40(d) misses an opportunity to give clear guidance to ensure local rules authorize release for as many people as possible.
Proposed Rule 4.40(d)(1)
This proposed rule’s “[a]ll or nearly all” language is not only imprecise, but also aims for the wrong target. The relevant policy goal for the Committee with respect to pretrial release decisions is to ensure that release is authorized for as many people as possible. As a result, the Committee must set a brighter line than “all or nearly all.” There are many ways in which the Committee could do this. For example, the Committee could consider setting a defined threshold, which could read as follows: “If a court chooses to add to the list of exclusionary offenses or factors, the court must not adopt a rule that includes exclusions which effectively exclude more than 25 percent of persons assessed as medium risk from prearraignment release.”
The Committee should also consider adding robust procedural requirements for local courts to add exclusionary offense or factors to local rules. For example, one such requirement could be that for every exclusion or factor a local court wishes to add, that decision must be unanimous among the relevant judges of the court. Further, another potential requirement would force local courts to get permission from the Judicial Council and issue a public statement of reasons for their decision.
Proposed Rule 4.40(d)(6)
We are heartened to see that the Committee will require courts to consider the potential impact of new exclusions on marginalized groups. However, the Committee should go further than just requiring a court to consider this impact. The Committee should require courts to publicly state why the addition of an exclusion would not increase disparity in detention rates among marginalized groups in the local population.
Proposed Rule 4.40(e) is a step in the right direction, but the Judicial Council should require more.
Proposed Rule 4.40(e)(1)
We are encouraged to see that in developing the local rule, the court “must” meet with justice system partners, like the county behavioral health agency and community-based organizations. However, the “as appropriate” modifier, though well-intentioned, may lead some courts to believe this consultation is pro forma. The Committee should consider language to avoid this result. For example, the Committee could consider language like: the court “must consult with other justice system resources, on an ongoing basis,” or “must regularly consult.” Such modification could help ensure that community-based organizations and directly-impacted communities have an empowered voice.
Further, the court should seek guidance on “evidence-based practices in pretrial release and detention” not only from state and local organizations with relevant expertise. National organizations can also be of significant help. The draft rule should be modified so that courts may avail themselves of the best possible resources.
Proposed Rule 4.40(e)(2)
This subsection creates an annual review of racial and ethnic disparities in the criminal justice system to be conducted by the local courts. That review must show consultation with community stakeholders and detailed analysis of race data. However, we don’t know what metrics the courts will use to decide whether a tool or PAS department is “racist or not” and there is no condition of release of that report contained in rule 4.40. Such reports and findings risk to be ineffective, inaccurate and unpublicized.
It is important that this proposed rule directs local courts to study the impact of their rules on communities that are overrepresented in the criminal justice system. But local courts should not just “examine whether the rule has had a disproportionate impact based on race or ethnicity, gender or other demographics.” It is certainly a necessary step. But to the extent that a court’s examination finds a disproportionate impact, the Judicial Council should require that court to modify its rule or take some steps to remedy and reduce that burden. This subsection needs to have a requirement, standard, and process for changing any part of the local rule that is found to have a discriminatory impact.
For the reasons outlined above, we oppose the proposed rules and offered recommendations for modification. If you have any additional questions, please feel free to contact Logan Koepke, Senior Policy Analyst, Upturn, at email@example.com or Sakira Cook, Director, Justice Program, The Leadership Conference on Civil and Human Rights at firstname.lastname@example.org.
The Leadership Conference on Civil and Human Rights
Media Mobilizing Project
NYU Center on Race, Inequality, and the Law