New York Releasing Incarcerated Individuals Early, Cites Staffing Shortage

New York’s prison system began releasing people early in April 2025 because it could not staff the facilities it built to hold them. Michigan is managing a version of the same crisis. Both states are treating the symptom. Neither is examining the structure that produced it.

The New York Crisis: What the Memo Actually Said

DOCCS Commissioner Daniel Martuscello issued his memo on March 31, 2025 — three weeks after the end of a 22-day wildcat strike by correction officers who had been protesting mandatory overtime and the state’s HALT Act limiting solitary confinement.^[1] More than 11,000 officers participated in the strike. Governor Hochul’s response was to deploy approximately 6,500 National Guard members to maintain operations and to fire approximately 2,000 officers who did not return to work when ordered. The financial cost of that response exceeded $100 million per month.^[2]

That context matters for reading the memo accurately. The staffing crisis that prompted early releases was not created by the strike. A February 2025 memo from Martuscello directed prisons to operate as though 70 percent staffing levels were equivalent to full staffing — an acknowledgment that DOCCS was already short approximately 2,200 positions before the first officer walked off the job.^[3] The strike and the subsequent terminations deepened a hole that already existed. By the time the early release memo was issued, DOCCS had 10,000 officers and sergeants employed against a workforce that was being sustained at scale only through National Guard deployment.

The memo directed prison superintendents to compile spreadsheets of eligible candidates. The eligibility criteria were specific: within 15 to 110 days of an approved release date from the Parole Board, Time Allowance Committee, or maximum sentence expiration; no conviction for violent felonies, sex offenses, murder, terrorism, or arson; an approved residence that was not a shelter or a Department of Social Services placement.^[4] Martuscello described the approach as invoking Correction Law Section 73, the same provision used during the COVID-19 pandemic to manage population amid health concerns. The commissioner said the released individuals were already approaching release and had already completed parole board approval processes. He was not technically wrong. He was also not acknowledging what the housing requirement would do to the actual release numbers.

The Housing Requirement That Swallowed the Policy

The early release policy did not produce the numbers the announcement suggested. By late April 2025, DOCCS reported that 766 incarcerated people had been initially identified as eligible candidates. Of those, 151 had been released. The gap — more than 600 people who qualified on paper but were not released — was explained primarily by the housing requirement.^[5]

The requirement that eligible individuals have an approved non-shelter residence is a condition that sounds reasonable until you examine what it excludes. Approximately 42 percent of people released from New York state prisons transferred directly to New York City shelters in 2024, according to a report by the Coalition for the Homeless citing DOCCS records.^[6] Under DOCCS’s interpretation of the early release law, nonprofit transitional housing organizations — including the Osborne Association, which provides reentry services — also fell under the shelter prohibition because that housing is technically labeled as shelter. People who had been approved for release by the Parole Board and met every other criterion were nonetheless held because the housing they had available did not satisfy the residence requirement.

The Center for Community Alternatives, a New York advocacy organization, named this for what it was: the policy acknowledged over-incarceration as a contributing factor to the staffing crisis while simultaneously maintaining the housing barriers that keep people incarcerated after their legal basis for detention has ended. Prison reform advocates called for the governor to use clemency power to reach people who had demonstrated rehabilitation but remained incarcerated because the early release criteria were too narrow.

The Political Frame Versus the Structural Problem

Republican criticism of the early release directive focused on public safety: that releasing people early compromised community safety, that the right solution was to rehire the fired officers. That argument has political traction regardless of its evidentiary basis. It also does not engage with the structural reality Martuscello’s own memo described.

DOCCS had documented a 24.4 percent decline in corrections officer positions between 2000 and 2023, a trend predating the strike by more than two decades.^[7] The staffing problem is not a management failure of any particular administration. It is the predictable result of building a large incarceration system during an era of mass incarceration and then failing to sustain the workforce required to operate it as conditions — pay, working environment, social attitudes toward corrections work, and competition from other labor markets — shifted against recruitment and retention.

The argument that rehiring fired officers would solve the problem is also complicated by the numbers. Even before the strike, DOCCS had 2,200 unfilled positions. Rehiring the 2,000 terminated officers would substantially address the immediate crisis. It would not address the structural trend line. It would not change the working conditions that prompted officers to strike in the first place. And it would not change the fact that New York was operating 70 percent staffed and calling it full staffing before the strike began.

Michigan: A Slower Version of the Same Story

Michigan has not had a corrections officer strike. It also has not had the acute crisis that produces dramatic policy responses. What it has had is a persistent, documented staffing shortage that has operated below the threshold for emergency action while accumulating the same structural pressures.

As of early 2025, the Michigan Department of Corrections operated 26 facilities with 5,530 total corrections officer positions and 956 vacancies — an overall statewide vacancy rate of approximately 17 percent.^[8] Eleven of those 26 facilities had vacancy rates of 20 percent or higher. Three had rates above 30 percent. In July 2024, the Michigan Corrections Organization’s president described conditions at some facilities as operating at crisis levels, with corrections officers being forced to work mandatory 16-hour doubles and in some cases sleeping in their vehicles between shifts.^[9]

The MCO formally requested National Guard assistance from Governor Whitmer in July 2024. The state declined, with MDOC characterizing the National Guard as a temporary measure rather than a solution and noting that Guard members are not trained for corrections environments.^[10] New York deployed the Guard anyway and spent more than $100 million a month doing so. Michigan has so far avoided that outcome by managing the shortage through mandatory overtime and a sustained recruitment push — but the 2025 per-prisoner cost reaching $52,050, up from roughly $48,000 the prior year, reflects a system that is becoming more expensive to operate even as its workforce struggles to keep pace with attrition.^[11]

The state has graduated two large academy classes in late 2024 and early 2025 — 281 new recruits in the fall 2024 class, the largest since 2015, and 224 in the spring 2025 class — bringing over 500 new officers into service across two classes.^[12] The MCO’s president acknowledged the numbers but said new graduates were barely keeping pace with the rate of retirements and resignations. Michigan is running to stay in place on staffing, at increasing cost per prisoner, in a system that is approaching 95 percent capacity as available beds decline following facility closures.

What Early Release as a Staffing Tool Reveals

When a state uses early release as a mechanism to manage corrections staffing, it is making a practical admission that it built more incarceration capacity than it can operate. The admission is not made in those terms. It is made through memo language about “appropriate balance” between staff safety and public safety, through 70 percent staffing directives called full staffing, through National Guard deployments to fill the gaps.

But the underlying logic is the same: the system cannot sustain the population it is holding at the staffing level it can maintain. The response — releasing people who were going to be released soon anyway — is operational triage, not justice policy. The people released are not chosen based on any principled evaluation of who should have been incarcerated in the first place. They are chosen based on eligibility criteria designed to minimize political risk: non-violent, non-sexual, near-release, with approved housing. The people who have served long sentences and demonstrated genuine rehabilitation remain inside, because their release requires a decision, and triage does not produce decisions.

Prison reform advocates in New York named the contradiction directly: the policy acknowledged over-incarceration as a systemic problem while releasing only the people whose release required the least political justification. The Center for Community Alternatives called it an overdue acknowledgment that New York’s prison crisis was one of over-incarceration, not just staffing shortages — and then noted the directive was far too narrow to address that.

The Diversion Argument

The corrections staffing crisis in both New York and Michigan is a back-end expression of a front-end policy failure. Diversion programs, mental health courts, and community-based treatment address the question before it becomes a corrections management problem: which people need to be incarcerated, and which people have underlying conditions that incarceration cannot address and that community treatment can?

The Vera Institute documents that diversion programs reduce rearrest rates and cost substantially less than incarceration per participant. Michigan’s annual cost of $52,050 per prisoner means that the state is spending that amount per person to operate a system that, at its current vacancy rates, is simultaneously unsafe for corrections officers and unable to deliver the rehabilitative programming that would make incarceration meaningful rather than purely punitive.

Diversion does not reduce the corrections staffing crisis today. It addresses the structural condition that produces that crisis over time. A system with fewer people in it requires fewer staff to operate safely. It generates lower overtime costs. It does not require National Guard deployment. It does not require governors to issue emergency early release memos and then discover that housing requirements prevent most eligible people from actually leaving.

The counterargument — that diversion compromises public safety — does not engage with the evidence on recidivism outcomes, and it does not address what the current system is actually producing. A corrections system operating at 70 to 83 percent staffing, with officers sleeping in their vehicles between mandatory overtime shifts, deploying the National Guard to fill gaps at $100 million per month, and releasing people through triage rather than justice decisions, is not a public safety system functioning at design capacity. It is a system in managed decline, making decisions by default that it has avoided making by choice.