New York City Mayor Bill De Blasio unveiled a $17.8 million “bail alternative” plan to release more people under supervision while they await trial.

Expanding supervised release is important not only due to the city’s troubling history with pretrial detention, but because incarceration stacks the deck against defendants, who are left to fight their cases from within violent and overcrowded jails, away from the support of family and friends. Behind bars, their access to important services—such law libraries and resources, or mental health and substance abuse treatment—is nearly impossible to gain. Moreover, even if the person has not been fully tried, convicted and sentenced, they may still be subject to administrative punishments, like pretrial solitary confinement, which can have adverse effects on their defense.

Supervised release pilot programs have been in place in Queens, Manhattan and Brooklyn since 2009, and boast an 85% completion rate. The expanded program can potentially supervise around 3,000 people and will be funded by $4 million from taxpayers and $13.8 million in asset forfeiture funds from the Manhattan District Attorney, Cyrus Vance Jr.

According to the mayor’s office, the program’s expansion was developed “in close partnership with the Office of Court Administration, the five District Attorneys’ offices, the defense bar, and national experts to ensure that it is both evidence-driven and tailored to New York City’s unique criminal justice system.”

Yung-Mi Lee, a supervising attorney at Brooklyn Defender Services who was on the planning committee for the supervised release program, told Shadowproof that “any program that gets more clients out of Rikers Island or keeps them out of Rikers Island is a good program.”

“These are people who would have otherwise maybe gone to jail with bail set,” Lee said. “I think any effort to try to keep people out of jail is a good thing, so I can’t really overly criticize it.”

The program focuses on misdemeanor and non-violent felony defendants, using an updated risk assessment tool that purports to calculate an individual’s likelihood of re-arrest, as opposed to their risk of failing to appear in court. Under current state law, risk of flight is the sole consideration available for judges in the state of New York when determining whether or not to release a defendant.

Some advocates have said letting judges consider the risk an individual poses to public safety as another factor in bail determinations could reduce the number of people incarcerated before trial. Basing the risk assessment on public safety is one way to inject this consideration into the process.

The De Blasio administration acknowledged the limits imposed on judges and said it would “pursue a change to state law so that New York is no longer one of only two states that prohibit judges from considering public safety risk when making bail determinations.”

Risk assessment works by calculating a score from an array of “risk prediction factors,” including a person’s age, whether they have any open cases or prior warrants and convictions, whether it’s their first arrest, and if they report full-time to an activity, such as work or school. Risk points were calculated for each factor according to their influence on a person’s likelihood of being re-arrested.

Once the individual is assigned to a risk category, if they are eligible, a judge can decide to release them under the supervision program. Those in the program will receive phone calls and text messages, and are to participate in regular face-to-face sessions with caseworkers to make sure they have access to the services they need and show up to court when required.

But the complex issue of risk assessment and eligibility places some people at an undeniable disadvantage, despite other mitigating factors that could ensure they appear in court and not be re-arrested.

The new supervised release risk assessment tool gives youth between the ages of 16 and 19 the highest number of risk points (six) for any single factor, due to their age group’s comparatively high likelihood of re-arrest. But this may be a particularly counterproductive assessment to make.

One study by the Justice Policy Institute found around 70% of the nation’s juvenile defendants were detained before trial on non-violent offenses in 2014, in facilities where they were “physically and emotionally separated from the families and communities who are the most invested in their recovery and success.” Juveniles in pretrial detention were housed in “overcrowded, understaffed facilities—an environment that conspires to breed neglect and violence.” Rikers Island has a reputation for fostering no different an environment.

The study cited a psychologist, who found that one third of youth in detention suffered from depression, which they developed after incarceration. Another psychologist “suggests that poor mental health, and the conditions of confinement together conspire to make it more likely that incarcerated teens will engage in suicide and self-harm.”

But most importantly for the purpose of basing supervised release on risk to public safety, the study noted that “there is credible and significant research that suggests that the experience of detention may make it more likely that youth will continue to engage in delinquent behavior, and that the detention experience may increase the odds that youth will recidivate, further compromising public safety.”

When asked about the risk points assigned to the 16-19 year old age group, Yung Mi Lee said, “The problem, I think, any time you use statistics and probabilities, is when you’re looking at this one person in the courtroom, and the judge says ‘He’s not eligible for supervised release, he’s high risk,’ then there might be that mindset—that the judge will say, ‘Well I can’t release this person,’ even though that child might have a mother in the courtroom, a teacher in the courtroom.”

“Even though he lives right in the borough and has been living at the same address and has nowhere else to go. Even though that child might have, say, a cognitive issue, where the mother is going to guarantee that this child will return to court.”

“These are the kinds of individual factors we want judges to take into consideration, and not just looking at the risk assessment tool,” Lee said.

Thanks to the New York Police Department’s war on the city’s youth, particularly young black males, the NYC Criminal Justice Agency reported over 70% of juvenile arrests in 2014 were “non-docketed,” or not prosecuted at all–but many of those kids spent a night in jail nonetheless.

For youth detained until their arraignment, between 22-55% were then held on bail or remand with no bail set. Roughly 38% of juvenile defenders were unable to meet bail, and were left behind bars.

The irony is that New York City youth who are released, either on bail or on recognizance, fail to appear in court at very low rates: of those who were released, only 2% failed to appear in court. 0% of those who met bail failed to appear.

“Yes it’s problematic, because we still want judges to be able to look at an individual case,” Yung-Mi Lee said of the data-driven approach to risk assessment for youth. “Because if you look at the re-arrest rate for 16 to 19 year olds, it’s not even close to 50%. It’s just they have a higher arrest rate than, say, 30-34 year olds. That’s all it is.”

“So if you’re looking at someone and saying, ‘He’s 17,’ you can’t even say he has a higher chance of him getting re-arrested, because that’s not the case,” Lee continued. “Because this particular individual also may not get re-arrested, because he has a lot of different family members surrounding him, and other factors such as the fact that he may not even be guilty if you look at the facts of the case. So of course you’re going to want to come back and fight the case.”

“But I think even the people who did this study for the mayor’s office for the supervised release program clearly acknowledge that even brief periods of incarceration have bad impacts on everyone, not just the young clients.”

Individuals charged with violent offenses and felonies of a higher level are ineligible for supervised release, based more on perception than whether or not there is data correlating their charge with their likelihood to return to court or re-offend.

The New York City Criminal Justice Agency found that, across all five boroughs, the second most-common top charge (or most severe charge) facing individuals in 2014 were “physically injurious” charges, such as those involving homicide, arson, assault, violent sex offenses, kidnapping, robbery, and “other crimes of physical harm,” representing nearly 20% of all cases.

The same report found those facing felony charges have lower failure-to-appear rates than those facing non-felony charges. But the reality is those facing more serious charges simultaneously make prosecutors look like they’re bringing serious criminals to account, while making it seem unacceptable to release them during trial.

“There’s always room for improvement,” Yung-Mi Lee said when asked how she’d like to see the program develop in the future. “I would like to see obviously more cases included, but I do think it’s a step by step process.”

Citing her belief that those facing domestic violence charges should be considered for the program, Lee explained how blanket prohibitions on release for more serious charges can be problematic, and why one shouldn’t judge a charge by its name. “The definition of a domestic violence case is very broad,” Lee said, “so it’s not your usual romantic partner versus romantic partner domestic violence.”

“If you live with someone, even if it’s a family member by blood relation, that can be considered a domestic violence case. So you can have two sisters who fight with each other and it’s not your typical batterer situation,” Lee explained. “But if that’s defined as domestic violence, which it is, then that person is not going to be considered for supervised release, if they are in that risk category where they may get sent to Rikers Island.”

“I think any program that’s very data driven, where you’re just basing risk factors on statistics and probabilities–which is, I guess, the only way you can determine someone’s risk factor—you’re excluding the ability for judges to make individual, independent, discretionary judgements,” Lee said.

In New York City and across the country, prosecutors regularly over-charge defendants to pressure them to accept a plea deal on lesser charges, which may still carry serious penalties. The decision to give prosecutors such a significant seat at the table, instead of making a unilateral decision in the public’s best interest, has the potential to dull reform by pitting vested and disparately powerful interests against each other.

The reason for the program’s limited scope, and the perceptions they enforce, may have something to do with the roles played by city prosecutors in shaping the program, most notably Manhattan District Attorney Cyrus Vance, Jr.—the program’s main funder.

District Attorney Vance has has been dubbed a “kind of Santa Claus for law enforcement” by the New York Times, for using most of the $808 million in criminal penalties from cases against banks HBSC, Standard Chartered, and BNP Paribas on justice programs and equipment.

In November, the Times reported Vance had disbursed $69.5 million of that $808 million, with another $459 million earmarked for research, projects, organizations, and more. His office’s donations rival those from major philanthropic organizations, like the Open Society Foundation and the Ford Foundation.

Vance’s office sunk $90 million into NYPD smartphones and tablets; $101 million to improve locks and lighting in housing projects; funds for cybercrime labs, training and equipment for medical examiners, license plate readers and a photographic database of convicted felons.

Vance’s apparent charity has proven useful in other ways as well. The Times noted the district attorney “skirted battles with Mr. de Blasio, partly by contributing about $40 million to the mayor’s efforts to help the mentally ill and reduce the population at Rikers Island.” To put it another way, he avoided reproach while also affording praise by spending large sums of money on the mayor’s reform plan.

Allowing these programs to be funded through the prosecutors office means there is no public debate or other budgetary process for appropriating those funds. Once the initial funding dries out, most of these kinds of programs are automatically folded into their institutions. And some of the major institutional organizations tapped to provide caseworker services for supervised release, who have contracts and agreements with with other city agencies, are on the receiving end of this money and can expand their operations. While everyone gets some of what they want through compromise, those facing more serious charges are left behind.

Those closest to people languishing in the system are rightfully desperate to see change and some relief for the populations they serve, who have had to face down significant hardships brought by the city’s justice system for far, far too long. The expanded supervise release program is a significant victory, and the city should be commended for taking this step.

But it’s hard to deny meek approaches allow abuse to continue against people trapped in the criminal justice system, as elites ensconce themselves in the rhetoric of justice and change—while avoiding both for the foreseeable future. Real lives are impacted, and some are lost, along this path of incremental change.

“I’m hoping, because we’re going to continue to have meetings with the mayor’s office and the supervised release program, and everyone seems open to taking suggestions and moving forward, if the program is successful, they might open it up to domestic violence cases and maybe eventually, somewhere down the line, violent felonies. I’m not sure,” Yung-Mi Lee said.

“I have a client right now who’s doing supervised release. He was charged with a nonviolent felony, and when I told him he was eligible for the program and this was a way for him to get out of jail, he was so happy.”

“He’s back with his wife, he’s going to have to come back and fight the case,” Lee said. “But the fact that, even for one client, that he’s not sitting in jail two months after the case started, and he’s still able to fight the case on the outside—it’s a big deal to him, so it’s a big deal to me.”