The federal system: A patchwork of agencies rapidly expanding federal surveillance and punishment

Because it is so fragmented, the federal system can be particularly hard to understand, which only makes it harder to change effectively through policy. We explain the agencies involved, the loophole that undermines ‘sanctuary’ policies, and overlooked problems with federal pretrial detention.

by Jacob Kang-Brown, July 23, 2026

The federal system of incarceration and detention makes up a relatively small, but rapidly expanding, share of confinement in the United States. Between 1984 and 2026, the number of people confined by the federal government has exploded by 631%. And while that number is below its 2011 peak, federal incarceration and detention have grown by 12% in just the last two years.¹

The Trump administration’s recent, belligerent use of policing and prosecutions have already left their mark on the federal system, especially with regard to mass detention and deportation. But even so, it’s not easy to understand the scope of this corner of the system and how its practices and priorities have shifted over time. Opaque federal police and prison agencies make it hard to see these changes, or to know which agency is responsible. And it’s harder still to know what federal policy reforms or state and local resistance would be most effective to rein in runaway expansion plans.

In this briefing, we try to overcome these issues by piecing together data from various sources to provide a clearer picture of the patchwork federal system, from ICE detention and electronic monitoring to federal prisons and pretrial supervision. We show trends in detention and incarceration,² and examine the role that the criminalization of immigration (including the exploitation of a loophole in sanctuary policies) has played in shaping the system. Finally, we round out our analysis by spotlighting some of the glaring problems with federal pretrial detention and highlighting the fiscal costs of the system.

With a better grasp of the agencies involved and the mechanisms at play, people can work with their state and local governments to limit the reach of the federal system.

The size and scope of federal systems of confinement and supervision

Today, 268,000 people are detained or incarcerated on federal authority, and 321,000 people face the threat of being locked up while under federal community supervision. Taken together, more than half a million people are under some kind of federal correctional control.³ But they are spread across four major agencies or authorities, depending on the charges and the status of their cases. For example, after a federal arrest, people are locked up in one of three primary ways:

1. Pretrial detention while a criminal case proceeds in federal court. They may be held without bail or, more often, unable to afford release on bail.
2. A sentence to incarceration. When someone takes a guilty plea or is convicted at (an exceedingly rare) trial, a judge can sentence them to incarceration.
3. ICE detention: They are confined while a civil case proceeds in immigration court or while awaiting deportation.

Additionally, many who aren’t locked up are instead under community supervision by either the federal criminal court system or ICE. In this section, we briefly explain which agencies are involved at various stages of the legal process and how they share resources to detain, incarcerate, and monitor so many people.

Pretrial detention. While being prosecuted in a federal criminal court, people are booked into custody of the U.S. Marshals Service (USMS or “the Marshals”). The Marshals provide space for pretrial detention to the federal courts, primarily via contracts with a network of local jails.⁴ While the Marshals’ role in criminalizing immigration is less visible, their detention network presents a loophole that allows the use of local resources for immigration enforcement despite “sanctuary” policies. It’s important to understand that the Marshals often serve the same purposes as ICE and share the same resources. Almost half of all people who face federal criminal charges in the course of the year are there for some kind of “criminal” violation of immigration law, such as returning to the United States after an earlier deportation.⁵ Meanwhile, ICE relies heavily on the Marshals’ network of local jails to detain people for civil immigration matters; only a small share of these individuals have criminal convictions.

Incarceration. After sentencing in a federal criminal court, people with short terms of incarceration are sometimes left in U.S. Marshals custody instead of being transferred to a federal prison. Many people don’t know that 1 in 5 people serving a federal sentence of incarceration on any given day are in USMS custody. The Federal Bureau of Prisons handles everyone else with sentences to incarceration, using its detention centers, penitentiaries, correctional institutions, and prison camps.

Immigration detention. In federal civil immigration cases, people are detained in ICE’s detention network. This network consists of detention centers operated by private prison companies, but also include some federal detention centers, state prisons, Bureau of Prisons facilities, and ICE offices. Importantly, ICE often has access to the USMS detention network of local jails as well, as intergovernmental service agreements commonly include both agencies.⁶ ICE can also contract directly with local jails.

Community supervision: pretrial, probation, post-release, and for ICE. Beyond federal detention and incarceration, 321,000 people face community supervision and surveillance by the federal government, both of which come with the inherent threat of confinement for any alleged violation. Community supervision for people accused or sentenced for federal crimes is overseen by the federal courts’ Probation and Pretrial Services. ICE surveils hundreds of thousands of people on an “alternatives to detention” electronic monitoring program.⁷

This table breaks down how many people are involved in various parts of the system now:⁸

Immigrant criminalization and prosecution are driving federal confinement trends

The longer-run trends of federal detention and incarceration are similar to the overall expansion of mass incarceration through the 1980s and 1990s.¹⁰ Since then, the federal system has continued to add huge numbers of people in immigration detention, especially during the Trump administrations. That increase in immigration detention, combined with the failure of the federal government to keep pace with broader declines in criminal punishment at the state and local levels over the past six years, has led to an overall spike in federal confinement.

The number of people confined by federal agencies is more than six times greater in 2026 than it was in 1984, although it remains below its 2011 peak. Note that this chart does not include confinement by state or local criminal legal systems.

Moreover, Congress has already approved funding for ICE to expand detention capacity to 135,000, almost double the current level. If used, this would drive federal confinement to all-time highs. Extreme levels of ICE arrests — like the 10,000 people arrested in the last 5 days of June 2026, and 2,000 on the weekdays that followed — will assuredly lead to even more people held in ICE detention.

The role of immigration detention and criminalization

In addition to the well-documented expansion of ICE detention, federal prosecutors, the U.S. Marshals, and the Bureau of Prisons together have played an overlooked role in the expansion of confinement for immigration-related reasons, detaining and incarcerating people on criminal charges that had commonly been treated as civil matters in the past.¹¹

Federal criminal prosecution allows immigration arrests and detention even in places with sanctuary policies.¹² That’s because of a longstanding loophole: ICE and other federal agencies can refer people for federal prosecution on immigration-related “crimes” and thus hold them in contracted local jails in sanctuary cities, counties, and states. In doing so, the Trump administration is transforming what are normally civil immigration matters into more serious federal crimes. This expands the overall federal incarceration system. As we have explained, every county with a U.S. Marshals contract is effectively signed up for this part of the mass detention and deportation agenda, even if they have sanctuary policies on their books.

We estimate the federal Bureau of Prisons and U.S. Marshals detained or incarcerated roughly 25,000 people on an average day in fiscal year 2025 for immigration-related criminal charges, up 45% from around 17,300 people in 2024.¹³ Federal prosecutors have been filing more cases against people for immigration-related offenses, sending people into the federal criminal legal system. Almost 50,000 people were charged and booked into U.S. Marshals custody for an immigration-related offense in fiscal year 2025 alone. The Trump administration plans to take this even further.

Expanded prosecutions and more detention

The Trump Administration projects that it will book 250,000 people into jail in fiscal year 2027.¹⁴ There were fewer than 100,000 federal bookings in 2024.

The most dramatic part of this projection is the expected booking of almost 110,000 people on immigration-related criminal charges, followed by sharp increases in the catch-all “other charges” group. Historically speaking, bookings in the “other” category primarily have been for supervision violations, followed by property offenses.¹⁵

Notably, the Trump administration also plans a rapid increase in the number of people booked on federal drug charges, indicating a return to the failed policies of the “war on drugs.” This will only compound problems within the federal pretrial system: people facing federal drug charges tend to be detained pretrial for longer than others (7.5 months before sentencing), in part because they often face protracted plea negotiations to avoid harsh mandatory-minimum charges.¹⁶ Aggressively prosecuting drug-related crime will exacerbate the harmful impacts of pretrial detention in the federal system.

What federal, state, and local policymakers can do

The United States still incarcerates and detains its residents far more than peer nations on a per capita basis, and changes at the federal level remain absolutely necessary — though not sufficient on their own — to end mass incarceration. Federal policymakers have direct control over criminal and immigration law, agency budgets, and appointment and confirmation of agency directors, federal prosecutors, and judges. Their influence can be seen in the outsized growth of the federal criminal legal and immigration detention system, especially in recent years, when immigration detention has spiked.

At the top of any policy agenda must be addressing the injustice of ICE and civil immigration enforcement with their bloated budgets and lack of transparency and accountability. But the criminalization of immigration by the federal criminal courts, as well as the ongoing “war on drugs,” also demand action. Currently, federal judges and prosecutors use the vast detention network set up by the U.S. Marshals Service to lock people up in hundreds of different local jails and private prisons for the purposes of punishing what have not always been criminal offenses. Advocates and elected officials have fought back by ending law enforcement and jail collaborations with ICE. But policymakers need to go further and also block collaboration with other federal detention agencies like the U.S. Marshals. While federal policy action is essential to changing this system and reversing its buildup, states and local governments can take steps to limit it now.

How states use — or don’t use — medical parole to release sick and dying people from prison

An updated table of state-by-state data shows that medical parole (also known as compassionate release) is used so sparingly that very few people are actually released this way.

by Emily Widra, July 22, 2026

In almost every state, there are laws on the books offering a way out of prison for people who are seriously or terminally ill. Medical parole — also known in some states as compassionate release, medical release, or conditional release for medical reasons — is ostensibly an early release mechanism for extremely ill or incapacitated people who pose little or no threat to public safety. In reality, medical parole sets an extremely high bar for applicants, with a lengthy and confusing application and hearing process. As a result, most applicants remain locked up, with many people dying before their cases are resolved.

We compiled publicly available data from all 47 states with medical parole laws, the District of Columbia, and the federal prison system, in an effort to update data collected and published by FAMM which they updated in 2021 and 2025 where possible). Ultimately, these data show just how few people benefit from this type of early release, despite the level of need and public support for these laws. Because most states are not statutorily required to publish data regarding medical parole, in many cases we were not able to find data more recent than what FAMM collected from state agency reports and records requests. In most states where we were able to update the data, there has been no increase in the use of medical parole over the past few years, and the number of people granted medical parole each year remains extremely low. For example:

• Florida reported granting Conditional Medical Release to 45 people in 2019 and 35 people in 2020, but granted release to only 18 people in FY2025.
• Massachusetts reported granting Medical Parole to three people in 2019 and 31 people in 2020 (during COVID-19), but only 11 people in FY2024.
• New Jersey reported granting Medical Parole to ten people from 2011 to 2020, but only eight people from FY2021 to FY2025.

With a rapidly aging population, inadequate medical care inside, and too few care facilities that will accept people with felony convictions if they are approved for medical parole, states need to make better use of all early release mechanisms to keep prisons from becoming nursing homes.

Involuntary commitment: How forced hospitalization expands the criminal legal system

Civil and criminal courts both mandate the detention of people for extended — and often indefinite — periods of time in state hospitals, essentially extending the already-long arm of the criminal legal system.

by Emily Widra, July 21, 2026

Hundreds of thousands of people are forced into psychiatric and substance use treatment every year in the United States under the umbrella term “civil commitment.” Proponents try to justify these forced hospitalizations as necessary to “help” people who are unable to help themselves despite little evidence to back their medical efficacy.¹ In reality, this contact between the criminal legal and medical systems expands carceral control; as a legal tool for confinement on the basis of mental health conditions or substance use disorders, they effectively criminalize medical conditions. And, in keeping with both systems, forced treatment and hospitalizations are rife with racial disparities, and disproportionately impact people who are unhoused and living in poverty. ²

In this briefing, we catalogue the various forms of forced hospitalization, where they happen in the criminal legal system, and whom they impact. There are two broad types of forced hospitalizations, which we differentiate as non-forensic and forensic. The first type, non-forensic forced hospitalizations, are used for psychiatric and substance use disorders, and are the result of a civil court process unrelated to criminal charges. Even so, this process frequently involves the criminal legal system, as police often bring people to emergency rooms for psychiatric evaluations and can even petition the courts for forced hospitalizations. The second type, forensic forced hospitalizations, refer to the commitment of individuals to psychiatric hospitals because of their criminal legal system involvement. These hospitalizations are typically for (1) competency assessments and restoration services or treatment before trial, (2) the evaluation and treatment of people found “not guilty by reason of insanity” (NGRI) or “guilty but mentally ill” (GBMI), and (3) the commitment of people convicted of sex offenses after they have completed their prison sentences. Forensic forced hospitalizations can take place at any point throughout the criminal legal process. In essence, this fusion of the criminal legal and medical systems disguises the criminalization of mental illness and substance use as a benevolent act.

Civil courts forcibly hospitalize people for mental health and substance use disorders

In every state, civil courts can force individuals with symptoms of a serious mental condition into hospitals against their will if they are determined to be a danger to themselves or others, and/or are unable to provide for their “basic needs.”³ What exactly constitutes a person’s “basic needs” can vary by jurisdiction, but it’s typically understood to be food, clothing, and shelter. As of 2024, more than two-thirds of states have expanded these non-forensic hospitalization laws to target people with substance use disorders.

Psychiatric hospitalizations. The data on how many people are subject to forced psychiatric hospitalizations are sparse and suggest significant state variation.⁴ Still, even a rough, conservative estimate suggests hundreds of thousands of forced hospitalizations take place each year in the United States.⁵ People of color (particularly Black and Asian people) are disproportionately targeted and, across all categories of mental health conditions, people diagnosed with psychosis, schizophrenia, or bipolar disorder are among the most likely to be forced into mental health treatment.

Substance use disorder hospitalizations. Data about forced substance use treatment are even scarcer; there are no national estimates of how many people are court-ordered into treatment or how long they are typically confined.⁶ In Massachusetts — a frequent utilizer of laws to hospitalize people over substance use — more than 6,000 people are court-ordered into treatment each year. In 2018, there were more than 10,700 petitions and more than 7,000 evaluations for forced substance use treatment, and ultimately, nearly 6,000 treatment orders. A Massachusetts state commission found that people subjected to such treatment are frequently unhoused (24%) and unemployed (80%), and have no history of overdose (63%). While we cannot necessarily extrapolate estimates from Massachusetts to other states, we can see how involuntary treatment for substance use impacts unhoused people, people in poverty, and people of color disproportionately, mirroring disparities within the criminal legal system.

In general, non-forensic forced hospitalizations of both kinds follow a similar process: an individual is detained in a hospital or emergency room on an “emergency psychiatric hold” (typically a 72-hour detention) while a civil court determines whether they should be involuntarily hospitalized.⁷ In every jurisdiction, police can initiate an emergency hold, but beyond that, the rules vary about who can start the process.⁸ In some states, the court determination can proceed without an emergency hold while the individual remains in the community.⁹ In all cases, a judge ultimately determines whether an individual will be admitted to involuntary treatment in a state treatment facility.

In almost every state, law enforcement — not medical personnel — are responsible for transporting people to the facility once the court has committed them, even though they haven’t been criminally charged. Generally, people are hospitalized in state medical facilities, but it’s not uncommon for people to be held at other facilities (like private hospitals or emergency rooms) until a bed in a state facility becomes available.¹⁰ In Massachusetts, the court can confine people forced into substance use treatment within correctional facilities; in 2019, three in every four men forced into treatment in the state were in a jail or prison without any criminal charges or convictions. Scarce data exist regarding the actual length of time individuals spend in these settings, but state statutes cap the length of confinement (which varies by state from 14 days to one year) and civil courts can indefinitely renew the court orders.¹¹

In most jurisdictions, civil courts also determine when someone can be released from involuntary psychiatric hospitalization, presumably with input from the treatment team. When it comes to substance use hospitalizations, the state health authority can typically discharge people.¹² Individuals can also pursue their own discharge from either via civil court or “conditional release” programs that are contingent upon their participation in outpatient programs.¹³ Notably, “conditional release” from forced hospitalizations imitates probation and parole systems in the criminal legal context: people can be returned to confinement if they do not adequately adhere to a set of rules governing their behavior.

Forced hospitalizations are becoming more common in the criminal legal system

Forensic forced hospitalizations have been on the rise in recent years, and can occur at a number of points along a person’s route through the criminal legal system. Before or during a criminal trial, questions of defendants’ “competency” — that is, their ability to understand the charges and penalties they face, make decisions, and communicate with their legal representation during the trial — can result in forced hospitalizations. The verdict of the trial can also result in a forced hospitalization if the person is acquitted as “not guilty by reason of insanity” (NGRI) or, in some jurisdictions, convicted as “guilty but mentally ill” (GBMI). An acquittal as NGRI results in court-ordered psychiatric treatment, after which the individual is released and not subject to additional criminal penalties. In contrast, someone found GBMI and court-ordered to psychiatric treatment must serve their criminal sentence after completing treatment. It’s worth repeating that competency, NGRI, and GBMI are inherently legal concepts, not medical or clinical concepts or diagnoses, and are not based on best practices regarding mental illness and treatment.

Pretrial forced hospitalizations for competency evaluations and “restoration”

In criminal court, the defense, prosecution, or judge may question the defendant’s competency during pretrial hearings or at any point during the trial. In this scenario, “competency” is strictly limited to the defendant’s ability to participate in the trial and is unrelated to their mental state at the time of the alleged crime. The judge will order a competency assessment from the state mental health authority and if the individual is found “incompetent to stand trial,” the court will order competency “restoration services.”

An estimated 60,000 people are evaluated for competency each year in the United States, and about 20% of evaluations result in hospitalizations to “restore” competency. Based on data from the Treatment Advocacy Center (TAC) and NRI, we estimate that on any given day, nearly 3,000 people are in hospitals for competency evaluations, while over 8,000 people are in hospitals for competency restoration.¹⁵ The limited demographic research on competency finds that people of color are 1.4 times as likely to be found incompetent than white defendants,¹⁶ and suggests a statistically significant relationship between incompetency findings and defendants identified as female, minority race, or married.

The state mental health authority is responsible for the court-ordered assessment and subsequent competency restoration services or treatment. Such evaluations and restoration services are almost always provided in inpatient hospital settings, despite evidence that outpatient competency restoration programs in the community are effective. In a 2014 survey of officials in 43 states, respondents in nearly 80% of those states reported that courts can (and do) forcibly hospitalize people for competency assessments and, in most states, restoration services are exclusively provided in inpatient hospital settings.¹⁷

State statutes often limit how long a person can be hospitalized for an evaluation, typically ranging from 15 to 60 days.¹⁸ While most state mental health authorities agree that these evaluations can be completed in one or two interviews, people are often hospitalized for longer than is needed. The trial court then determines competency and can order forced hospitalization for “restoration services.” In every state, this can include medicating people against their will, even if they don’t meet the usual criteria (like danger to self or others due to mental disorders).¹⁹

Once the court determines that competency has been “restored,” the criminal trial proceeds.²⁰ However, it is worth noting that in many states, people will remain hospitalized even after competency is “restored” if the hospital staff believe they will “decompensate” upon return to jail. If the court determines that someone is “unrestorable,” the individual must be referred to a civil court for a civil commitment hearing (i.e., non-forensic, non-criminal) or released. Most states limit the amount of time someone in this scenario can be hospitalized²¹ but the National Association of Mental Health Program Directors states that “several studies” have found that these defendants often remain hospitalized long after they are determined to be unrestorable.

Competency evaluations and restorations generate tension between jails and state hospitals, as there are too few hospital beds and too many people in jail awaiting evaluation or restoration, often for low-level offenses. Reports of lengthy waitlists for evaluations are common, as well as long wait times for competency hearings in court and delays for criminal trials to start after being found competent. Once a person is found incompetent, there are more waitlists for state psychiatric hospital beds. In other words, people who may be in psychological distress or in need of psychiatric treatment are left to languish in jails until a hospital bed becomes available. Even a day or two in pretrial detention can destabilize a person’s life for years to come, jeopardizing employment, health, housing, and more — effects that may be more acute for people with mental health conditions. Despite the high rates of mental health and substance use disorders among incarcerated people, jails rarely have adequate resources to care for them, and are inappropriate settings to house or treat people with mental health disorders.

Forced hospitalizations as sentencing outcomes

“Not guilty by reason of insanity” (NGRI) and “guilty but mentally ill” (GBMI) verdicts are typically the result of legal defense efforts to contextualize a crime, mitigate punishment, and ensure treatment is included in the case outcome. Once charged, the defendant, the defense attorney, or the judge may request a psychiatric evaluation, which can be conducted on an outpatient basis (in the community or the jail) or in a state psychiatric hospital.²²In 2017, the New York Times estimated that nearly 90% of NGRI verdicts were the result of plea bargains based on the results of these evaluations. In other situations, the evaluation may be presented as part of the defense’s argument that the defendant is “not guilty by reason of insanity” or requires treatment.

“Not guilty by reason of insanity” (NGRI) is an acquittal based on the criminal court’s determination that the individual was unable to differentiate right from wrong due to their mental condition at the time of the offense. Usually, this is because they were experiencing severe symptoms of mental illness that gravely impaired their capacity to perceive reality or think clearly. The court typically determines a defendant is NGRI based on a psychological evaluation, a psychiatric diagnosis, and a “finding of dangerousness,” meaning the court believes them to be a danger to themselves or others. Unlike competency, a finding of NGRI reflects a decision about the person’s mental state at the time of the crime and is unrelated to their mental state at the time of the criminal trial: NGRI verdicts result in automatic forced hospitalizations, even if the person is no longer in the same mental state as they were at the time of crime.

Alternatively, at least 13 states offer a verdict of “guilty but mentally ill” (GBMI), which is a criminal conviction that can result in forced hospitalization, but still requires a person to additionally serve whatever the typical sentence would be for the offense.²³ The establishment of GBMI verdicts in a number of states was largely in reaction to public demands for fewer acquittals via NGRI.²⁴ GBMI verdicts do not result in reduced sentences; people must serve the same sentence as anyone convicted of the same crime, with the difference being that they receive psychiatric treatment “if available” or “if necessary.”²⁵ However, they must serve the full term of the sentence regardless of the outcome of the treatment. Under a GBMI conviction, a person can be incarcerated for the maximum prison term, even if they are no longer experiencing the symptoms of mental illness they had at the time of the offense.

An estimated 7,900 people are in hospitals following “not guilty by reason of insanity” or “guilty but mentally ill” verdicts on any given day.²⁶ While this estimate combines NGRI and GBMI hospitalized populations, older research suggests that people found GBMI are no more likely to receive treatment for mental health disorders than any other person in prison. And while no nationally-representative data exist to show the demographics of people hospitalized due to NGRI or GBMI verdicts, an expansive analysis of NGRI acquittals in Missouri from 1980-2009 found an overrepresentation of Black people, men,²⁷ and people with prior suicidal ideation or attempts.

Despite the reputation of the “insanity defense” as a “get out of jail free card,” NGRI and GBMI verdicts are actually pathways to a different — and often less accountable — system of punishment. An NGRI verdict most often results in automatic hospitalization, and a finding of GBMI may result in forced hospitalization and/or a sentence of incarceration. The average length of time someone is hospitalized after a NGRI verdict ranges from months to years, depending on the jurisdiction, and most often people are confined in the hospital until the court releases them. Some states offer “conditional release” provisions where people hospitalized as NGRI can be released into the community but remain under the jurisdiction of the court. Just like someone on probation or parole supervision in the criminal legal system, someone with this status can be re-hospitalized for a violation of any conditions of release. And most states do not have a specified limit for how long a person found NGRI may be subject to the terms of conditional release.

Little evidence supports the widespread use of forced hospitalizations in civil or criminal courts

Beyond their ethical implications — including the curtailment of civil liberties and patient rights — one of the most alarming aspects of all types of forced hospitalizations is how little information is available about who is confined in these settings, what kind of treatments they are subjected to, and the outcomes of forced treatment.³³ What does exist, however, is some troubling evidence that forced hospitalizations are ineffective and even harmful. Research suggests the practice can actually raise the risk of suicide or overdose while failing to prevent relapses or future arrests: for example, in a study focused on Allegheny County, Pennsylvania, researchers found that forced hospitalizations double the risk of being charged with a crime and the risk of death by suicide or overdose in the three months following discharge.³⁴ The implementation of laws allowing the forced hospitalization for substance use disorders has not helped with opioid overdose death rates: a 2024 analysis actually found a significant rise in opioid overdose death rates in states with these laws (compared to those without such laws) in 2020-2021 compared to 2010-2019.³⁵ And in Massachusetts, the risk of opioid overdose death was actually higher for people with a history of forced, court-ordered substance use treatment than those enrolled in other state-run substance use treatment services. Fewer than 1 in 10 people released from forced substance use treatment in Massachusetts actually attended their scheduled follow-up appointments, and more than one-third reported relapsing on the day of release.

Research also suggests that courts do not need to rely on forced hospitalizations. Many people subject to competency restoration or found to be “not guilty by reason of insanity” (NGRI) are able to remain in the community for treatment or safely return to their communities after a short period of treatment. And outpatient competency restoration, which is generally completed much faster than in a forced hospitalization setting, appears successful. In Washington, D.C., three-quarters of people referred to an outpatient program were “restored” in the initial 45 days, and in Virginia, over 80% of people found incompetent to stand trial had their competency “restored” via a jail-based program in an average of 77 days. People hospitalized for competency restoration typically take much longer: the median length of hospital stays for competency restoration was 147 days in a meta-analysis of more than 50 studies from 1975-2013.³⁶ Similarly, studies indicate that longer hospitalizations seem to have little impact on the risk of re-arrest, and that the success of NGRI treatment is primarily contingent on housing and psychosocial support at the time of discharge or conditional release. The majority of people on conditional release from NGRI hospitalizations (76% in Virginia and 66% in Oregon) were able to safely remain in the community and participate in outpatient court-ordered treatment. With little support for forced hospitalizations, the available data seem to indicate that states can safely reduce their reliance on forensic hospitalizations and, more broadly, reconsider their use of forced psychiatric and substance use hospitalizations.

Conclusion

Civil and criminal court-ordered hospitalizations confine hundreds of thousands of people into a murky space between the legal system and the healthcare system. The sparse data landscape is particularly troubling given the growing reliance on, and growing political support for, forced mental health and substance use treatment in the United States. The expansion of carceral control over people who are not charged with criminal offenses and people who have already served their criminal sentences reveals yet another way the legal system seeks to criminalize mental illness and substance use. In addition, the rampant use of forced hospitalizations as a “solution” to crime — one which simultaneously expands the definition of “criminal behavior” to include homelessness, mental illness, and substance use — turns systemic outcomes into individual failings, and disregards the real root of the problem: an unequal society that relies on the carceral system to (poorly) address societal failings.

The best way to address this misdiagnosis of the problem would be to expand access to community-based mental health and substance use treatment. For example, investing in mental health crisis care instead of arrests, and utilizing community-based diversion programs instead of pursuing prosecution for nonviolent misdemeanors, could reduce the number of people hospitalized against their will for competency evaluations and NGRI verdicts. Community-based competency evaluations and restoration programs also reduce the number of people confined in jails, while simultaneously relieving the burden on state hospitals by shrinking the waitlists for inpatient beds. As the Substance Abuse and Mental Health Services Administration explains, prosecutors and judges “must exercise discretion to determine if prosecution and continued involvement in the criminal legal system is appropriate for individuals with serious mental illness” throughout the process.

Prison Policy Initiative endorses the Extreme Temperatures Injustice in Prisons Act

Nearly 19,000 people in federal prisons experienced heat-related illnesses between 2022 and 2025. A new bill is looking to change that.

by Emmett Sanders, July 16, 2026

On Thursday, July 8, 2026, U.S. Congresswoman Alma S. Adams of North Carolina and Congresswoman Lucy McBath of Georgia introduced the Extreme Temperature Injustice in Prisons Act. This Act is a continuation of efforts by the Congresspeople to address the practice of subjecting people incarcerated in federal prisons to extreme, life-threatening temperatures. It stems from a 2025 letter of inquiry sent by the Congresspeople to the Federal Bureau of Prisons regarding conditions in their facilities. In their reply, the BOP revealed that nearly 19,000 people incarcerated in their facilities reported heat-related illnesses between January 2022 and September of 2025 alone. More than 40 were diagnosed with heat stroke.

The Prison Policy Initiative provided technical support to the drafters of the bill, and we are proud to endorse this legislation, which takes a much-needed step towards ensuring that people in federal prisons are afforded standardized safe living conditions.

The bill has several provisions designed to set standards, increase oversight, and mitigate harm. Among other things, this Act:

• Authorizes funding for climate control system installation and upgrades in all federal prisons within five years, with temperature monitors in every cell and workspace.
  Set clear health and safety standards for extreme heat and cold across work, recreational activities, and housing.
• Directs the Center for Disease Control & Prevention (CDC) to research how extreme temperatures harm incarcerated individuals and staff and develop extra protections for elderly, pregnant, and immunocompromised individuals.
• Provides free cooling and warming supplies — like cold water, fans, warm clothing, and blankets — when temperatures reach dangerous levels.
• Implements accountability measures regarding staff training for recognizing and responding to heat and cold emergencies
• Requires regular public reports on HVAC progress, temperature-related illness and deaths, and annual safety inspections.

As millions across the US face rising temperatures and heat advisories, people in prisons are routinely locked inside without access to air conditioning. Even those who are most vulnerable, like the many in prison with chronic medical issues, the elderly, and pregnant people, are left with few options to mitigate the deadly conditions. Advocates have long sounded the alarm about the dangers of extreme heat and cold temperatures in prisons, which put both individual lives and overall prison safety at extreme risk. No one should be forced to endure these conditions, particularly not those who safety is entrusted to government hands.

We at the Prison Policy Initiative strongly support the Extreme Temperature Injustice in Prisons Act, and applaud these efforts towards recognizing that people in prison are deserving of safe living conditions and basic human decency. A heatwave should not be a death sentence for people in prison.

Read the official press release for the Extreme Temperature Injustice in Prisons Act here.
For further information on the need for environmental justice in prisons, consider the following resources:

• Heat, floods, pests, disease, and death: What climate change means for people in prison
• Cruel and unusual punishment: When states don’t provide air conditioning in prison

Since you asked: What does early release look like in states that eliminated discretionary parole?

While 17 states and D.C. have taken discretionary parole off the table for most or all incarcerated people, they still have other forms of parole and conditional release that could safely release many more people from prison. Here, we examine these slimmed-down parole systems and other release mechanisms, and show they are not wildly different from states still using discretionary parole.

by Leah Wang, June 23, 2026

A common but misleading way to think about parole is that some states have it and some states do not. It’s true that many resources (including our own reports) will say that since 1976, more than a dozen states have done away with discretionary parole, requiring most or all criminally sentenced people to serve their full sentence instead of having an opportunity to demonstrate to a parole board their transformation and readiness for release. This approach taken by 17 states (and D.C.) to date is also known as part of a “truth-in-sentencing” structure for determining criminal sentences, and indeed took away a huge opportunity for early release for the vast majority of people entering prisons.

However, binary thinking about parole does not reflect the wide range of ways that states “do” parole and other forms of “early release”¹ — particularly those states that are thought to have no parole. Their parole systems are just as worthy of assessment and reform as those we included in our recent two-part report on discretionary parole. In this briefing, a companion to that report, we answer some common questions about states “without parole” and how they compare to states with active discretionary parole systems.²

Here, we also fill in data gaps, providing a comprehensive table describing these often overlooked “post-parole” states. Using state statutes, agency websites, and news media, we compiled information about how these hollowed-out parole systems work, including eligibility criteria, board composition, outcomes, and other ways (if any) incarcerated people can earn an early release. These details can be found in full in the appendix table accompanying this briefing.

In 17 states and D.C., people are eligible for discretionary parole under old laws or narrow criteria

Every state that abolished or curtailed discretionary parole — which can also be generally described as shifting to a determinate sentencing structure — still has a fully functioning paroling authority, holding hearings and making a variety of decisions related to release and supervision. Even where some states’ parole-eligible populations are dwindling, the boards are typically tasked with other work, such as processing clemency applications or setting conditions for other types of release.

Characteristics of parole release systems in states without discretionary parole, 2026

Our findings about parole in these states include:

• In all 17 states and D.C., people who committed crimes before parole was abolished are still eligible for parole. In states such as Maine and Indiana, this means that someone must still be incarcerated for a crime committed in the 1970s or earlier; Wisconsin and Virginia are among the states that passed their corresponding laws in the 1990s.
• In 10 states, there are other situations that make someone eligible for discretionary parole, such as having a life sentence (4 states), being a youth or young adult at the time of the crime (5 states), or being sentenced outside of the truth-in-sentencing guidelines (3 states).
• The number of parole-eligible people — in the few states where we could locate such data — ranges from zero (in Maine) to over 41,000 (in California). States where eligibility is based solely on the date of the crime generally have low and dwindling parole-eligible populations.
• Parole boards or their equivalent (such as a “review board”) in these states range in size from three to 21 members, but more than half (9 states) have five-member boards. In every state except Ohio, these board members are appointed by the governor.³
• Where data were available, we found parole grant rates ranging from 1% (in Virginia) to 23% (in Minnesota). By contrast, our analysis of states with discretionary parole found recent grant rates as high as 76%, with a dozen states reporting over 50% of hearings resulting in a grant. (Note that parole grant rates are difficult to compare to one another because parole applicants and eligibility criteria can vary so widely.)
• Most states (15) offer some type of medical parole or compassionate release, but in some cases this opportunity is framed as an “extraordinary” release; in Wisconsin, an approved medical situation can result in resentencing, rather than a direct release.

Looking at the structure and function of these parole systems, it’s evident that states without discretionary parole look a lot like states with it. Every state’s parole board, hearing format, and eligibility rules offer something for advocates to note and assess, for example against the core principles of parole.

Other opportunities for early release aren’t unique to states without discretionary parole, and aren’t actually letting many eligible people out

It is critical for states to have mechanisms for evaluating long sentences and releasing people whose incarceration is no longer justifiable. Other than discretionary parole, there are several statutory “release valves” that states could be using, such as medical parole, resentencing, and earned or good time systems; we introduce these and more in our 2018 report, Eight Keys to Mercy.

It would be one thing if the 17 states without discretionary parole (plus D.C.) had well-functioning release tools, with streamlined application processes and broad eligibility leading to thousands of releases annually, but this is objectively not the case. Plenty of states with discretionary parole have these release mechanisms, too, and they are universally underutilized. Below, we break down some of the most common early release opportunities found in prisons, showing that these 17 states and D.C. should (but do not) lean on them more heavily to relieve overcrowded prisons and safely return people to their communities.

Mandatory parole. Some states that ended discretionary parole replaced it with a system of mandatory parole, where at least some time on supervision is required at the end of someone’s sentence. For example, Illinois has had Mandatory Supervised Release since 1978, and New Mexico has post-release supervision, two programs for which the parole boards set conditions for people but do not have the final say in their readiness for release. But mandatory supervision is not an innovation from states that eliminated parole; it exists in many other states, like Alabama, Massachusetts, Texas, and West Virginia.⁴

Mandatory post-release supervision may sound like an improvement on discretionary parole because, for the most part, a parole board cannot stop someone’s “mandatory” release. But research suggests that this approach to release may not be effective, with higher re-arrest levels compared to discretionary parole release⁵ and lower levels of successful supervision completion.⁶ Even when mandatory supervision “works,” it is no replacement for discretionary parole.

Some states are revisiting their medical parole laws to expand eligibility or streamline the application process, but these efforts⁹ have not resulted in more releases. According to some prison systems’ own data, dozens of people are so sick that they require acute long-term care or 24-hour nursing care behind bars at any given time, vastly outnumbering those who are granted medical parole over an entire year.

Earned time and good time. These policies credit days off of a criminal sentence for completing programs or education behind bars, or maintaining a clean disciplinary record. Sentencing credit structures range from day-for-day (one day without disciplinary infractions leads to one day off of the sentence) to just a few days’ credit for a month of compliant behavior. States with strict truth-in-sentencing laws tend to have stingy earned and good time policies,¹⁰ so they should not be considered comparable to discretionary parole. Rather, sentencing credits should complement opportunities to discuss growth and transformation in front of a parole board.

Resentencing, sentence review, or “second look” sentencing. More and more, states are responding to the vast body of research showing that long prison sentences don’t work, passing legislation that authorizes judges to review a sentence after someone has served a specific amount of time. This mechanism, while fundamentally different from discretionary parole in its core question and structure, has great potential to redress excessive prison sentences rooted in outdated ideas about punishment. Disappointingly, only 15 states have second look laws on the books (beyond those that address juvenile or young adult life-without-parole sentences); seven of those states lack discretionary parole.

Finally, most governors have the authority to grant release or modify a sentence¹¹ as an act of clemency or under extraordinary circumstances. However, they often shy away from such merciful actions due to inevitable political blowback.

States that have done away with discretionary parole are still making sweeping changes to their parole systems

Advocates in states that don’t have discretionary parole should still pay attention to their parole or post-release supervision system because reforms or rollbacks can easily move the needle on decarceration in either direction. In a move that will place release further out of reach for incarcerated people, Kansas recently overhauled its parole board, increasing victim and law enforcement experiential requirements and transferring the power to appoint board members from the corrections director to the state attorney general and governor, further solidifying the board as an extension of political power rather than a professional agency.

Parole in Perspective See our deep-dive report on discretionary parole systems

New Mexico, on the other hand, passed legislation acknowledging that parole should be based on dynamic factors and demonstrated rehabilitation, rather than static factors such as the original crime or a conviction history. The static factors are still included in the law as part of the bevy of information the parole board will look at, but this statutory framing that prioritizes readiness for release is a huge step forward, grounding parole in forward-looking criteria.

The biggest recent change to parole in this cohort of states likely took place in Minnesota, where an entirely new Supervised Release Board (SRB) was established in 2024 to oversee post-release supervision for a growing number of incarcerated people. The SRB will now replace the commissioner of corrections as the decision-making authority for releasing eligible people (such as those with life sentences), increasing the capacity to hold hearings and post decisions.

Further, unlike many other states in our analysis, Minnesota’s SRB does not handle clemency or medical parole applications, potentially making its workload lighter than that of other boards.¹² Still, some incarcerated people in Minnesota have to serve 30 years behind bars before becoming eligible for supervised release, far longer than what many advocates suggest is reasonable.

Parole in every state is flawed, but discretionary parole should be an option for all incarcerated people

As our research over the years has shown, no state is “doing” parole well. But every incarcerated person should have access to discretionary parole as a bedrock opportunity for release after showing transformation and preparedness. Lawmakers in the 18 jurisdictions that have curtailed discretionary parole should consider that loved ones and taxpayers are footing the bill for such an ineffective approach.

Beyond some of the changes described above, some advocates in these states (namely Illinois and Maine) are going to the root of the problem and campaigning to reinstate discretionary parole. Still other states, such as Wisconsin and Washington, have advocates working on establishing or expanding compassionate release, resentencing, or other release mechanisms to combat excessively long sentences and the crisis of aging in prison. Ultimately, any reform related to parole, sentencing, or early release should be retroactive, set out broad criteria that do not carve out certain convictions, and offer a straightforward process for consideration.

Parole and other release mechanisms are found in every state but are vastly underused. When wielded more broadly, more urgently, and without politics, they can meaningfully reduce the nation’s overreliance on incarceration.

The Prison Policy Initiative submits testimony in opposition to HB 338, which would curtail higher education in prison in Ohio

HB 338 would end access to higher education for people in maximum security prisons in Ohio, making prisons less safe and worsening outcomes after release.

by Sarah Staudt, June 16, 2026

On June 10, 2026, the Prison Policy Initiative submitted written testimony in the Ohio Senate in opposition to HB 338, also known as “Andy’s Law”.

The bill claims to promote safety in prisons, but instead is counterproductive. It removes all higher education from maximum security prisons in Ohio, and bans access to tablets and contact visitation. As our testimony notes, higher education in prison is shown to reduce both recidivism and violent incidents in custody. In short, it makes prisons and communities safer.

If you are a criminal legal system reform advocate or legislator who would like to talk to our Advocacy department about providing testimony for or against a bill, please reach out to us using our contact form.

New research: How past incarceration affects people later in life

Two recent studies authored by Professor Carmen Gutierrez and her colleagues show that prior incarceration is associated with various geriatric health conditions and reduced life expectancy later in life.

by Emily Widra, June 9, 2026

As a consequence of the dramatic expansion of incarceration in the last 50 years, a significant and growing portion of older adults in the U.S. have experienced incarceration.¹ In fact, three in every four people released from state prisons between 1991 and 2021 — more than 12 million people — would be over the age of 50 by 2026.² Most existing research about the effects of incarceration on health and mortality have focused on the initial weeks, months, and years after people are released from prison,³ but two recent studies call attention to the health and mortality consequences of incarceration that can take place decades later, when people are 50 or older. The most recent of these studies finds that older adults who have experienced any incarceration have shorter life expectancies than those who haven’t. The other study, published in 2024, reveals troubling details about the health and wellbeing of older formerly incarcerated people compared to their peers who have never been locked up. Taken together, these studies highlight the serious health consequences in later life that stem from past incarceration.

Older adults who have ever been incarcerated face reduced life expectancy

The newest publication from Professor Carmen Gutierrez and her colleagues, published last month (May 2026) in the American Journal of Public Health, is one of the first studies to show that any history of incarceration increases mortality (death) risk for older adults. While prior research has shown that incarceration accelerates physiological aging and shortens life expectancy overall, there is little known about how incarceration impacts mortality risk in later adulthood specifically. The researchers find that older adults who have experienced any amount or form of incarceration in their lives have shorter life expectancies than their peers who have never been incarcerated.

Professor Gutierrez and her colleagues analyzed the results of the 2012 and 2014 iterations of the Health & Retirement Study, a longitudinal, nationally representative survey of nearly 14,000 U.S. adults over the age of 50. As a measure of incarceration history, the survey asks a simple yes-or-no question: “Have you ever been incarcerated, that is, spent time in a jail, prison, juvenile detention center or other correctional facility?” The study also links individual responses to participants’ administrative records, including their appearance in the National Death Index. The researchers measured participants’ rates of death from any cause (i.e., all-cause mortality rates) between the 2012 and 2014 surveys and the follow-up interviews conducted in 2022, when the respondents would be roughly 60 or older.

While controlling for other characteristics like age, sex, race and ethnicity, and childhood health and socioeconomic factors, the researchers found that formerly incarcerated older people have an 88% higher hazard of premature death compared to those with no incarceration history.⁴ More specifically, older people who have experienced incarceration are expected to live nearly six fewer years than similar peers who have never been incarcerated. Formerly incarcerated men and women died at significantly younger average ages (74.8 and 68 years, respectively) than men and women without incarceration histories (82.3 years and 82.9 years, respectively).⁵ For men between 50 and 75 years old, the difference was even greater at close to eight fewer years, but after age 75, incarceration history appeared to have less of an effect on life expectancy.

In addition to its findings about mortality, this study found significant demographic differences between older adults with histories of incarceration and those with none, consistent with what we know about the socioeconomic and racial disparities of incarceration. Compared to older adults with no prior incarceration, formerly incarcerated older adults were more likely to be Black and to report childhood poverty, childhood housing insecurity, lower educational attainment, and fair or poor self-rated health (as opposed to good, very good, or excellent health). In addition, formerly incarcerated older adults were more likely to be unmarried, unemployed, and have a disability than those never incarcerated.

Incarceration is tied to serious health problems in later adulthood

An earlier study authored by a team of academic researchers including Professor Gutierrez — published in the Journal of Gerontology in 2024 — sheds more light on the health of formerly incarcerated people in their older age, with particular focus on chronic conditions that undoubtedly contribute to their shortened life expectancy. Previous research has shown that incarceration takes a toll on both individual health and public health, but this study provides evidence that incarceration is associated with poor health outcomes specifically in later life. The researchers find that any past incarceration is associated with impairments of mobility, vision, and activities of daily living,⁶ as well as diagnoses of cancer, depression, and chronic lung disease among people 50 and older in the U.S.

This study is based on an analysis of data from the National Longitudinal Study of Adolescent to Adult Health (Add Health), which follows a nationally representative sample of over 20,000 people who were in grades 7-12 during the 1994-95 school year. The research team analyzed data collected from two specific groups of people⁷ in this massive, long-term study:

• Parents: A survey of parents — biological, adoptive, or stepparents — of children who participated in the Add Health study. Almost all (96%) of the surveyed “parents” were women (mothers).
• Partners of parents: A survey of the cohabitating partners and spouses of the adults in the “parent” sample, as of the 1994 survey. The vast majority (95%) of these respondents were men (fathers or other male partners). To be clear, “partners” could also be parents of Add Health study participants, but only one adult in each household was included in the “parent” sample.

At the time of the 2015, 2016, and 2017 rounds of the Add Health study, there were 2,007 parents and 976 partners of parents who were at least 50 years old. The researchers analyzed survey responses about age-related health concerns among these nearly 3,000 older adults to compare those who reported any history of incarceration and those with no previous incarceration.⁸

In terms of health outcomes, the researchers found that the prevalence rates of a number of age-related conditions were higher among formerly incarcerated older adults. In the sample of parents (again, primarily mothers), those who were formerly incarcerated faced significantly higher rates of cancer, heart disease, stroke, depression, and chronic lung disease. In addition, those who had experienced incarceration were more likely to rate their own health as poor or fair, as opposed to parents without incarceration histories. When controlling for age and sex differences, parents with histories of incarceration were more likely to report difficulty walking and difficulty with activities of daily living than those parents with no history of incarceration.

Among the partner sample, made up of mostly men, the researchers also found alarmingly elevated rates of age-related health concerns associated with previous incarceration. In particular, prevalence rates of cancer, depression, chronic lung disease, and “heavy” alcohol use⁹ were significantly higher among formerly incarcerated people in the partner sample compared to those with no prior incarceration. In addition, formerly incarcerated partners reported difficulty with walking, seeing, and with activities of daily living at higher rates than those who reported no time spent in jail or prison. While the researchers do not explain the reason for the different findings between the parent sample and the partner sample, sex is the most obvious difference.

Aside from the findings about health concerns among formerly incarcerated older adults, the analysis of these samples also helps paint a picture of formerly incarcerated parents. First, incarceration was more common in the sample of partners (mostly men) than among the sample of parents (mostly women): 15% of partners reported any previous incarceration compared to only 4.2% of parents. Compared to parents and partners with no incarceration history, a greater portion of parents and partners with incarceration histories were Black, reported parental childhood abuse, had lower educational attainment, and reported receiving public assistance in the past 12 months. In terms of their access to healthcare, parents and partners who had ever been incarcerated were more likely to identify the emergency department as their usual source of medical care, report at least one overnight hospital stay in the past 12 months, and/or to have Medicaid or no health insurance. These differences tell us that not only are formerly incarcerated older adults particularly vulnerable to a number of serious health issues, a number of socioeconomic factors also limit their access to the appropriate and necessary care.

Conclusion

Millions of formerly incarcerated people are aging into older adulthood: approximately 1 in 15 adults aged 50 or over has a history of incarceration and therefore faces both an elevated risk of death and a number of serious medical conditions. These two studies provide further evidence that incarceration is a social determinant of health and should be included in assessments of public health. In addition, the elevated mortality risk and prevalence of chronic age-related conditions among formerly incarcerated older adults in these studies supports the argument that healthcare providers need to consider individuals’ history of incarceration to better understand the health risks they may face.

WEBINAR: Fighting for families: How advocates are resisting attacks on visitation and physical mail

by Sarah Staudt, June 4, 2026

Protecting family connections and bonds with loved ones during incarceration is one of the most important and, increasingly, one of the most difficult things advocates can do. Contact visitation, handwritten cards and letters, and regular phone calls improve mental health outcomes for both people in prisons and the families who love and support them. They also strengthen parent-child relationships and familial bonds, improve in-prison conduct and make prisons safer, and can significantly increase a person’s chance of success upon release.

Nevertheless, jails and prisons across the country are implementing increasingly restrictive communication policies. Contact visitation is under attack, and more than 30 states no longer allow incarcerated people to receive physical mail from their loved ones. These moves, ostensibly in the name of security, not only devastate incarcerated people and their loved ones but make us all less safe in the long run. But families are fighting back, championing policies to protect these vital lifelines and the benefits that come with them.

On July 1, Prison Policy Initiative hosted a webinar Jodi Hocking, Executive Director of Return Strong Nevada, and Taj Mahon-Haft PhD, Director of The Humanization Project in Virginia. We took a closer look at the importance of family connections, learned more about the ways they are under attack, and shared strategies to help families fight back.

Spanish language interpretation is available by selecting the “Spanish” audio track in the “Settings” menu on the webinar’s YouTube video. The American Sign Language interpreter feed for this video is available here.

Webinar Resources

The slides for the webinar are available here.

Prison Policy Initiative Resources

• Mail Scanning Toolkit — This toolkit gives ideas for how best to fight against mail scanning in prisons.
• Research Roundup: The positive impacts of family contact for incarcerated people and their families — This briefing summarizes decades of research showing that visitation, phone calls, and mail are essential to the wellbeing of incarcerated people and their families, and shows how communication with loved ones reduces recidivism and misconduct.

Additional Resources

• The Humanization Project — Webinar guest Taj Mahon-Haft is the director and co-founder of this Virginia-based organization, which seeks to inspire more humane, compassionate treatment of all people behind bars by humanizing their lives via their voices, mutual education and research.
• Return Strong — Webinar guest Jodi Hocking is the executive director of this Nevada-based organization, which works to protect the humanity of those affected by incarceration right now while simultaneously setting the foundation for a future free from prisons.
• We can’t afford it: Mass incarceration and the family tax — This June 2025 report by fwd.us that estimates the costs that mass incarceration brings to families — including communication costs.
• The power of free communication in prisons and jails — This May 2026 report by Worth Rises details the immense impact of free jail and prison phone calls.

Probation and parole trends by state: A look back at the data on ‘alternatives’ to incarceration

We compiled, corrected, and analyzed 45 years of state-level probation and parole data, and found that while most (but not all) states have reduced the number of people they’ve placed under supervision, many states have had more people on probation than previously reported.

by Jacob Kang-Brown, June 2, 2026

Around two-thirds of all people under correctional control are on probation or parole. For data details, see the methodology.

Community supervision represents the largest segment of the criminal legal system and, much like incarceration, it is almost entirely a state- and locally-run enterprise. As we explain in Punishment Beyond Prisons, on an average day, around 3.5 million people are on probation or parole compared to 2 million people who are incarcerated. To put this into perspective, that means more people are on parole than are in local jails nationwide, and people on probation make up more than half of all people under correctional control of any kind.
But how has the number of people under supervision in each state changed over time? In this briefing, we clean up and piece together state-level community supervision data covering roughly the last 45 years to provide a clear view of supervision dynamics in each state, as well as how states compare to one another.

This briefing includes state-level annual data tables and detailed trendline graphics going back to the 1970s, at the start of a rapid expansion of incarceration and community supervision. Compiling these data required correcting data published by the Bureau of Justice Statistics and updating estimates that states had previously reported during the 1990s or early 2000s. ¹ Through this process, we discovered that many states have had more people on probation than previously reported. We also found that people under supervision are trending older, much like the rest of the criminal legal system.

To learn more about the data we used and the corrections we made, see the data discussion and the detailed methodology.

Community supervision has declined by nearly a third since its peak in 2009

After experiencing rapid growth alongside the jail and prison boom decades of the 1980s and 1990s, parole and probation populations have changed course. Between 2009 and 2024, the overall rate of people on community supervision across the United States fell by 31%. ² This decline has happened nearly across the board: 41 states and the District of Columbia have substantially reduced community supervision rates over the past 15 years.

Declining numbers of people on community supervision are primarily the result of courts placing fewer people on probation, specifically. Parole boards have also been stingier with granting release in recent decades, holding fewer hearings and handing down denials in greater numbers. In fact, nearly every state with discretionary parole is granting release to fewer people each year.

While the number of people on community supervision has been trending downward, supervision nonetheless remains at a high level compared to other forms of correctional control. Looking beneath the national-level data can help us understand what is driving these changes.

At the regional level, the South ³ had the highest rate of community supervision in both 2009 and 2024; the Northeast ⁴ had the lowest. While every region saw falling supervision rates, the drops were larger in the West ⁵ and Northeast than in the South and Midwest.⁶ Meanwhile, in both the West and Northeast, the total number of people on community supervision fell to less than 1% of the total state population (a rate of 1,000 per 100,000 residents). In 2009, only 12 states were below that 1% mark; by 2024, half of all states were.

Of the 10 jurisdictions that had, at minimum, a staggering 1 in 50 people on probation or parole in 2009, only two (Georgia and Idaho) continue to place people under supervision at such a massive scale today. ⁷ Even so, Georgia’s community supervision rate has fallen by almost 30% in that time.

Arkansas has earned the ignoble distinction of having the largest rise in the community supervision rate between 2009 and 2024. The state went from having 1,785 to 2,399 per 100,000 residents under supervision — an increase in the rate of 614 per 100,000 residents. Remarkably, Arkansas went from the state with the 14th highest rate of supervision in the country (below the overall rate in the South) to the second highest. It is also one of only six states that saw a growth in its supervision rate of 10% or more. ⁸ See our appendix tables for a detailed breakdown of supervision population counts and rates for each state since the 1970s.

The new, more accurate data we present offer new insights on historical trends in community supervision. For example, Illinois and Oklahoma both had more or less unchanged rates of community supervision in 2009 and 2024. ⁹ However, the uncorrected data published by the Bureau of Justice Statistics would give a different impression: that Illinois had reduced community supervision by 42% in that period.

Importantly, readers should note that, when interpreting trend lines in different states, changes in probation and parole populations over time are not inherently good or bad; they can result from a wide range of reforms across the criminal legal system. For example, more people on parole might mean that states are moving more people out of prisons in recognition of the fact that they don’t threaten public safety. Declining community supervision could reflect new policies lowering barriers to successfully completing a supervision term, like reducing the number of conditions or the use of carceral sanctions for violations. We argue that far more people should be granted parole, ¹⁰ which would reduce prison populations and return people to their communities, and that probation is overused for lower-level offenses. Determining the forces that have shaped probation and parole in each state is outside the scope of this briefing, but we hope the data series we provide can help others carry that work forward.

The age of people on community supervision has risen dramatically

As mass incarceration ramped up between the 1970s and 1990s, politicians stoked racism and criminalized youth as easy targets, leading to an estimated 2.8 million youth and young adults on community supervision in 2009. ¹¹ By 2024, however, their number had fallen by two-thirds to under 1 million. Youth and young adults went from representing close to half of all people who were under supervision in 2009 ¹² to only a quarter in 2024. ¹³

While primarily a sanction for youth and young adults in the early 2000s, community supervision now appears to be more concerned with older adults and those living in poverty. In stark contrast to the drop in the supervised youth and young adult population, the number of people aged 50 or older who were under supervision grew by 2% between 2009 and 2024. ¹⁴

It’s worth noting just how much the parole and probation age distribution has changed in this time span. In 2009, 1 in 17 young adults were on supervision; for youth under 18, the rate was 1 in 30. For young people today, community supervision is a far rarer circumstance, inflicted on 1 in 65 people under age 18, and 1 in 59 young adults aged 18 to 25. For others, there has been much less of a change: in 2009, 1 in 33 adults aged 26 to 49 were on supervision, a rate that grew to 1 in 45 in 2024. Meanwhile, the supervision rate for adults aged 50 and older went from 1 in 130 in 2004 to 1 in 164 in 2024.

What we’ve learned about supervision data quality (and ongoing issues in certain states)

The fractured nature of correctional control across local, state, and federal jurisdictions makes it hard to truly comprehend the scale of mass punishment. While we typically rely on and trust data published by government agencies to produce our analyses, this research provided an opportunity to compare state and federal sources and correct what were, in many cases, obviously distorted datapoints. In this section, we give a high-level overview of some of the most common problems we identified in the data; you can skip ahead to read the full methodology for more detailed information on how we compiled these data and arrived at our estimates.

There are three different ways we honed the data to clarify trends in each state (impacting probation figures in particular).

1. The Bureau of Justice Statistics’ (BJS) improved sample makes it looks like caseloads grew when they didn’t: Recent BJS methods are finally catching up to the extent of probation in states like Oklahoma, Tennessee, and Virginia by collecting data from local (in addition to state) probation agencies. However, this change in collection practices has appeared in the data as a rapid rise in probation in those states. To the contrary, by constructing an accurate and comparable data set, we find that the number of people on probation in those states in recent years has been generally trending downward.

2. Incomplete data reporting by probation agencies sometimes leads to over-estimates in BJS publications. The Bureau of Justice Statistics imputes (creates estimates for) missing or incomplete data, and sometimes does so in a way that leads to an otherwise unexplainable data spike in a single year. In Delaware, Kansas, Massachusetts, Minnesota, and Mississippi, BJS data appear to show sudden peaks in the number of people on probation, while alternative state-level data sources we found show those increases never actually took place.

3. States often correct their data when they find errors, but the updates don’t always make it to BJS: In states like California, Illinois, Maryland, New York, and Pennsylvania, state agencies collect and synthesize probation information in ways that lead to data improvements over time. Unfortunately, the Bureau of Justice Statistics does not have a mandate to go back and collect this updated data, which would otherwise correct double counting, typos, and other issues. Neglecting these updates can enshrine errors in the historical record, as is the case with California’s peak probation population, which is actually quite a bit lower than reported by BJS. Worse, ignoring updates can obscure real-world changes. Take, for example, the real and sudden drop in probation cases between 2014 and 2015 in California due to Prop 47 making many criminal cases eligible for misdemeanor sentences that would have otherwise been felonies. The BJS data show a smoother decline, but subsequent data improvements from the California Department of Justice show that, while relatively small, the change was nonetheless much more abrupt.

Conclusion

As a true “alternative” to prison or pretrial detention, parole or probation can be an improvement over the status quo. Unfortunately, community supervision tends to be poorly designed as a way to balance accountability with connecting people to the services and treatment they need to achieve stability. Both probation and parole impose an excessive number of conditions on people that can be nearly impossible to comply with all the time. When someone on supervision is accused of violating a condition — particularly through behaviors that are not crimes — they are often incarcerated as a response, which only further destabilizes their lives. And “failure” is indeed common: Only about 36% of people successfully completed a term of probation, and 62% completed parole, in 2024. ¹⁹

Correctional control is practiced at an enormous scale in the United States, too often shuffling people back and forth between detention, incarceration, and supervision to the detriment of community health and public safety. It’s important for all jurisdictions to implement reforms to reduce and avoid correctional control as safely, efficiently, and as quickly as possible. Since community supervision is, like most of the criminal legal system, almost entirely a state and local affair, more accurate state-specific data (like those we’ve provided in this briefing) can help us see how the myriad supervision systems have developed over time. For those states that have genuinely reduced the burden of community supervision, it is worth understanding whether they can share lessons with other states, especially if they have been able to reduce incarceration as well.

The California Supreme Court puts on the record the best arguments against money bail

A recent case, In re Kowalczyk, affirms that money bonds are unfair, harmful, and make communities less safe.

by Sarah Staudt, May 27, 2026

In January 2021, Gerald Kowalczyk, a disabled man experiencing homelessness, found a credit card on the ground in San Mateo, California. He attempted to use the credit card to buy a $7 hamburger, then had a change of heart, asked for a refund, and when the manager refused, left without the hamburger. He was arrested shortly afterwards and charged with identity theft and petty theft. California law would generally not have allowed Mr. Kowalczyk to be detained pretrial on such minor, non-violent charges. But the judge, like too many prosecutors and judges, set an astronomical $75,000 bond in the case, stranding Mr. Kowalczyk in jail for six months before he pled guilty.

On April 30, 2026, the California Supreme Court expanded on a previous case, In re Humphrey, and ruled that Mr. Kowalczyk’s detention was unconstitutional under the California State Constitution because detaining people using unaffordable cash bails is the exact same thing as detaining them without bail, something that can only be done in limited circumstances with rigorous due process protections. Because of this case, unaffordable bail is now never an option in the California courts — if they use monetary bail, courts must set bail at a level that an accused person can pay.

The opinion and concurrence state in no uncertain terms what pretrial reform advocates have been saying for years:

• Using money to decide who goes to jail and who goes free is a lawless system that puts the opinions of judges and prosecutors above the will of legislatures and the public.
• Money bail privileges the rich and leaves everyone else stranded in jail.
• The overuse of pretrial jailing does more harm than good and destabilizes the very communities prosecutors and judges claim to be protecting.

In re Kowalczyk could free tens of thousands of people every year in California, if appropriately implemented. Even though this case can’t be used as binding precedent in other states because its legal basis is specific to the California State Constitution, the court’s reasoning gives excellent examples of how to make these important arguments resonate with policymakers.

When prosecutors and judges use money bail to detain people, they are imposing their own judgements instead of following the law

“Courts cannot use artificially high or objectively unattainable bail as an end run to effectuate pretrial detention where such detention is not authorized under [the California Constitution].” — Majority opinion, p. 32

The debate around cash bail centers around a lesson from Civics 101: separation of powers. Legislatures decide what the law is. Judges then apply that law to individual cases. The Constitution separates these roles because legislatures are best suited to make policy decisions, like the decisions about what charges and situations should and should not make someone eligible for pretrial detention, and how those decisions should be made. Judges ensure that these legislative decisions are carried out in the cases before them.

California, like many states, has a list of offenses that are eligible for pretrial detention (a list that excludes misdemeanors and low-level felonies like the ones Mr. Kowalczyk was arrested for). The state constitution then imposes a series of due process protections to make sure that decisions are made fairly.

The opinion in Kowalczyk makes clear that when judges — often at the request of prosecutors — use unaffordable cash bail to detain people, they are ignoring these constitutional rules and imposing their own judgments instead. Before this decision, when a judge set an unaffordable bail to detain someone, they didn’t have to go through the due process steps set out by the state constitution, because technically, the person wasn’t being detained. In theory, they could pay money to get out. In practice, however, detaining an unhoused person on a $75,000 bail is the same as saying he can’t be released at all — and that’s exactly what the California Supreme Court recognized in this case.

The takeaway: Advocates throughout the country should look at what their laws and constitutions actually say about who can be detained pretrial, determine whether they think those laws are fair or need to be changed, and then close the loophole of allowing judges and prosecutors to use unaffordable cash bail to ignore the law.

Using money to determine who goes to jail is illogical and unsafe

“A system in which a person’s right to liberty turns on financial resources compromises public safety and raises equal protection and due process concerns.” — Groban concurrence, p. 8

Even if decisionmakers think that pretrial detention of some people is necessary to keep communities safe (more on that below), when judges set unaffordable money bails it produces a fundamentally unfair and illogical system.

First, money bail does nothing to guarantee that a rich person who poses a real threat to the public will actually stay in custody. Conversely, everyday people who do not pose a risk are stuck in jail simply because they can’t afford to pay.

Second, a money bail system provides no meaningful due process protections. There is no requirement that there be sufficient evidence that someone is guilty, no requirement that a prosecutor prove that the person poses a public safety risk, and no examination of whether alternatives other than jail would be appropriate. Instead, the judge simply sets a monetary amount and the person is detained if they can’t pay it. This heightens the risk that innocent or non-dangerous people spend time in jail and encourages hasty, unthinking judicial decision-making that can put communities at risk. Notably, when Illinois eliminated monetary bail in 2023, the length of detention hearings increased dramatically, and judges considered more factors when making their decisions.

The takeaway: Advocates throughout the country should remove money from playing any role in their pretrial release systems, since it doesn’t have any bearing on safety.

The overuse of pretrial detention is actively harmful and doesn’t keep communities safe

“Many people who are detained are never convicted of any crime; a substantial proportion of detainees is never even charged with one. Many others accept unfavorable plea offers, with potentially devastating repercussions for their employment, education, housing, access to public benefits, immigration status, and family stability. And some, forced to choose between remaining detained or pleading guilty to an offense they did not commit, choose to plead guilty to secure their release.” — Groban concurrence, p. 9

Pretrial release should always be the default in the justice system. When someone is first arrested, they have not been proven guilty of a crime, and many people later have their cases dismissed, plead guilty to less serious charges, or are found not guilty. Estimates in some jurisdictions suggest that more than 50% of arrests end without a conviction. Nonetheless, nationwide, over 75% of people in jail custody are pretrial. This huge number of unconvicted individuals in jails makes them less safe environments by contributing to overcrowding.

Pretrial jailing harms people in custody in a number of ways. People detained pretrial are more likely to:

• lose their jobs and experience housing instability;
• die by suicide or overdose after release;
• become ensnared in the criminal legal system again, as measured by long-term re-arrest rates;
• miss a scheduled court date;
• be convicted and be sentenced to longer jail and prison terms, worsening prison overcrowding and mass incarceration as a whole.

Jurisdictions that have reduced the use of pretrial jailing have not seen any rise in crime. There is simply no evidence-based reason to continue the extremely high levels of pretrial detention that currently exist in the United States.

The takeaway: Advocates around the country should pursue laws that minimize the use of pretrial detention, seek to shrink jail populations, and provide robust due process protections to make sure that people are only behind bars if they truly pose a danger to others or are likely to willfully flee prosecution.

In re Kowalczyk has the potential to free tens of thousands of Californians each year as they are released on their own recognizance — or on money bail levels they can actually afford to pay — instead of being held in custody because they’re poor. The next step for California is implementation. Unfortunately, without rigorous efforts to require courts to follow a ruling like this, it could have little impact, as happened after the In re Humphrey decision. But the reasoning in this case can provide a roadmap for people in other states interested in creating just, fair, and rational pretrial systems.