Eight Keys to Mercy:

How to shorten excessive prison sentences

By Jorge Renaud   Tweet this
November 2018
Press release
Printable version


After decades of explosive growth, prison populations have mostly flattened. Much of that is due to lawmakers lessening penalties for drug possession or low-level property offenses. While a welcome start, a bolder approach is necessary to truly begin to make a dent in the numbers of individuals who have served and will serve decades behind bars. This approach will take political courage from legislators, judges, and the executive branch of state governments.

Approximately 200,000 individuals are in state prisons serving natural life or “virtual” life sentences.1 And as of year’s end 2015, one in every six individuals in a state prison had been there at least for 10 years.2

Graph showing the growth in the number of people in state prisons who have been incarcerated for over 10 years. Figure 1. The number of people spending 10 years or more in state prisons nearly tripled between 1999 and 2015, according to the Bureau of Justice Statistics’ National Corrections Reporting Program, 1991-2015: Selected Variables, Year-End Population. (1999 is the earliest year for which these data are available from this dataset.)

These are not merely statistics. These are people, sentenced to unimaginably long sentences in ways that do little to advance justice, provide deterrence, or offer solace to survivors of violence. The damage done to these individuals because of the time they must do in prison cells - as well as to their families and their communities - is incalculable.

People should not spend decades in prison without a meaningful chance of release. There exist vastly underused strategies that policy makers can employ to halt, and meaningfully reverse, our overreliance on incarceration. We present eight of those strategies below.

Understanding long prison terms and mechanisms for release

Too many state prisons hold too many individuals doing too much time. The goal of our eight strategies is to bring immediate relief to these individuals, by creating or expanding opportunities for their release. However, to discuss such reforms, we first need to understand the basic mechanisms by which someone is released from prison. In particular, it’s important to have a general idea of how parole works.

In general, when someone is convicted of a felony and sentenced, that person loses their liberty for a period of time. A portion of this period is typically served in a prison, and often a portion is served in the community under supervision, also known as parole.3 When parole boards have discretionary power, they periodically review someone’s case to determine if they should be released, beginning on their earliest release date. (One’s earliest release date may be well before the end of their punishment, or close to the end of their punishment, depending on state- specific statutes and requirements set by the judge.4)

Conceptual overview of time served. Figure 2. This schematic diagram illustrates how a sentence is served, from beginning to earliest possible release to the sentence’s eventual expiration (or maximum term). Where the release decision is made through a discretionary parole process (see sidebar), regular parole hearings present further opportunities for release. This graphic is not to scale and the ratios between the parts will vary by state and sentence type. Parole is very often available much later than this graphic implies or not at all.

Discretionary parole, mandatory release, and determinate and indeterminate sentencing explained

There are two basic types of release systems. Most states say they have one or the other, but it’s more helpful to think of your state’s system as incorporating elements of both. Here we explain the two types.

First, a note about vocabulary. This report does not focus on sentencing, which of course largely determines when and under what conditions individuals are eligible for or are released from prison. Instead of focusing on this “front-end” process, we are offering back-end reforms: how individuals can be returned to their communities.

States have different systems for deciding when to release people from prison. Most notably, some states primarily use discretionary parole, others primarily use mandatory release, and most states use a combination of the two.

  • With discretionary parole, an individual is granted release from prison by a panel of officials belonging to a “parole board.” This is the mode of release with indeterminate sentencing, wherein individuals become eligible for release to community supervision after a specified minimum amount of time in prison (for example, a quarter or half of a given sentence). An individual’s earliest release date varies depending on their conviction, their accrual of “good time” credits, and any statutory restrictions defining how much time they must spend before being eligible for parole. When deciding whether to release incarcerated people eligible for parole, the parole board reviews each individual’s criminal history, program participation, in-prison behavior, and other factors.
  • With mandatory release (sometimes called non-discretionary parole), an individual is released at a predetermined point and supervised in the community for the remainder of the sentence. This type of release is typical of determinate sentencing, which restricts the power of the parole board to make discretionary release decisions. An individual’s release is thus decided not by the parole board, but at sentencing or by statute. In some states, this “mandatory release” includes credit for in-prison behavior and program participation.

The main difference between the two systems is who decides when someone can be released (the legislature, via statute; or the parole board, through vote of the members). People released under either system can often “earn” an earlier release through participation in rehabilitative programming or good time, as we’ll discuss. But critically, people released under either type of supervision (or even probation) are equally subject to conditions of supervision, the violation of which can lead to re-incarceration until their sentence expires.

Most national efforts to discuss state parole try to classify each state as either using discretionary parole or mandatory release (or as having determinate or indeterminate sentencing).5 Were this so, either every incarcerated person in a given state would be eligible for parole, or none would ever be.

The reality is more complicated. Most state parole systems consist of a patchwork of discretionary parole and mandatory release.6 The vagaries of sentencing rules are to blame for this complexity. Every state has at various points changed its sentencing laws with respect to certain crimes (including rules about parole eligibility) in response to a change in the political mood, or a recent and highly-publicized violent crime. As a result, someone who commits a robbery or murder in one year may be eligible for parole after serving a fraction of a relatively short sentence. A person committing the exact same crime the next year may be denied parole eligibility and have to serve a quarter century or the rest of his or her life in prison. Every state has gone through these spasms, and they contribute hugely to the lack of equal justice in sentencing and parole.

This complexity also explains why some people say that states like Illinois or Florida “do not have parole,” even though those states do have people on parole, parole boards and parole processes (which in 2016 approved 0 and 2% of those eligible for parole in Illinois and Florida, respectively).7 Both of these states abolished discretionary parole for offenses committed after a certain year, so there are only a small number of people left in the system who are eligible for discretionary parole. Moreover, this population tends to be precisely the groups that the politically risk-averse parole boards are reluctant to release (i.e. those convicted of violent and sex-related offenses).

Because the discretionary-mandatory distinction is not a clean one, it is more helpful to evaluate how well a state parole system works by looking at the big picture, including:

  • The state’s sentencing structure (e.g. who is eligible for parole and parole eligibility criteria).
  • The prison system’s effectiveness at preparing people for parole (i.e. available programming and job training).
  • The parole board’s willingness to parole people who are eligible (i.e. grant rates, composition and philosophy).
  • The frequency with which the parole board sends people back to prison for technical violations (i.e. revocations for minor transgressions as opposed to new offenses).

For instance, someone convicted of aggravated robbery might be sentenced to a maximum of 30 years in prison, and in most states would be eligible for release after a certain period of time, let’s say 10 years.8 At that 10-year mark, this individual reaches their earliest release date, and the parole board considers their release on parole for the first time. If not released on parole, the parole board continues to consider release at regular intervals until that person is granted parole or maxes out their sentence.

Our 8 strategies

The eight suggested reforms in this report can shorten time served in different ways:

  • Several ways to make people eligible for release on parole sooner.
  • One way to make it more likely that the parole board will approve conditional release on parole.
  • Several ways to shorten the time that must be served, regardless of sentencing and parole decisions.
  • One simple way to ensure that people are not returned to prison.

Of course, states vary in many ways, most critically in how they structure parole eligibility (see sidebar above), and policymakers reading this report should anticipate tailoring our suggested reforms to their state systems. Each of the reforms laid out in this report could be effective independent of the others. However, we encourage states to use as many of the following tools as possible to shorten excessive sentences:

  1. Presumptive parole
  2. Second-look sentencing
  3. Granting of good time
  4. Universal parole eligibility after 15 years
  5. Retroactive application of sentence reduction reforms
  6. Elimination of parole revocations for technical violations
  7. Compassionate release
  8. Commutation

Presumptive parole

Presumptive parole is a system in which incarcerated individuals are released upon first becoming eligible for parole unless the parole board finds explicit reasons to not release them. This approach flips the current parole approach on its head, so that release on parole is the expected outcome, rather than one that must be argued for. Under this framework, an incarcerated person who meets certain preset conditions will automatically be released at a predetermined date.

How parole boards make decisions

Parole decisions are often tricky to predict or interpret. We explain the (often faulty) logic that informs parole decisions.

In most states, incarcerated people (and by extension, their families) are unsure what they must do in order to be granted parole. Parole boards generally rely on a mix of objective data9 and subjective judgment, but their decisions often lack predictability and transparency.10

One subjective factor that almost all parole boards weigh heavily - and which virtually all parole reformers say should not factor into release decisions - is the nature or seriousness of the crime of conviction.11 That factor will never change - it cannot become less or more serious - and in any event, the nature and seriousness of the crime was considered at sentencing.

The apparent randomness of parole decisions is not surprising: Rather than simply look for evidence that an incarcerated person will be a threat to public safety should they be released, parole board members often base their decisions on criteria so subjective it is unlikely any two people would agree on whether that criteria have been met. For example:

  • South Dakota asks parole board members to determine if an incarcerated people has “been confined for a sufficient length of time to accomplish the inmate’s rehabilitation.”12
  • In New Hampshire, a prospective parolee has to worry if the conviction carried with it over-the-top media coverage, because the board there is instructed to deny parole if there is “the existence of adverse public concern or notoriety [that] would hinder the inmate’s transition to the community.”13
  • In Utah, exercising your Constitutional right to challenge your conditions of confinement may count against you, because you can be denied parole for bringing a “claim that [any state or federal] court finds to be without merit.”14
  • New Mexico’s parole statutes include a set of conditions that may enable racial discrimination, as the Board is ordered to consider the incarcerated individual’s “culture, language, values, mores, judgments, communicative ability and other unique abilities.”15
  • Inevitably, the parole board will also consider the recommendations of prosecutors and crime survivors, who, by statute in most states, are invited to provide testimony. Both can be highly prejudicial, as well as uninformed about any programming or transformative experience the prospective parolee has undergone since being convicted and sentenced.

How should parole boards decide whether to release someone? Three of the most preeminent writers and thinkers on parole say it best: “The only ground for denial of release [on parole] should be the board’s finding, based on credible evidence, that the prisoner presents an unacceptable risk of reoffending if released.”16 No other criteria should matter.

Currently, parole boards treat continued confinement as the default and must justify why someone should be released. Logically, parole should only be denied if the board can prove that the individual has exhibited specific behaviors that indicate a public safety risk (repeated violent episodes in prison, refusal to participate in programming, aggressive correspondence with the victim, etc). But parole board members - who are almost exclusively gubernatorial appointees - may lose their jobs for merely considering to release someone sentenced to life,17 or for releasing someone who unexpectedly goes on to commit another crime.18 As a result, many parole boards and their controlling statutes routinely stray from evidence-based questions about safety (see sidebar, right).

The subjectivity of the current process is powerfully illustrated by the tremendous variations in the rate at which states grant parole at parole hearings, which vary from a high of 87% in Nebraska to a low of 7% in Ohio, with many states granting parole to just 20% to 30% of the individuals who are eligible.

Figure 3. Not all states make their parole grant rates available, but the existing data - collected by the Robina Institute - shows a tremendous variation between states (often resulting from the differences in state parole systems). In a few states, the data is more complicated for other reasons. Washington State, for example, eliminated discretionary parole in July 1984, leaving that option only to individuals who committed crimes before that date. However, Washington’s Indeterminate Sentence Review Board says it also reviews for release on parole anyone convicted “of certain sex offenses on or after Sept. 1, 2001” along with “persons who committed crimes prior to their 18th birthday and were sentenced as adults.” Therefore, it’s unclear how to report the parole grant rate for Washington.

An effective parole system that wants people to succeed will start with the assumption that success is possible. Instead of asking “why” the parole board should believe in the person coming before them, it should ask “why not” let that person go, then outline a plan that includes in-prison program participation and post-release community-based programming to help the potential parolee overcome barriers to release.

Changing this presumption would also create powerful new incentives for the entire system. The Department of Corrections would have an incentive to create meaningful programs, and incarcerated people would have an incentive to enroll and successfully complete them.

An effective presumptive parole system would have elements like those often found in Mississippi, New Jersey, Michigan, and Hawai’i:

  • Give clear instructions to incarcerated people on what they need to do in order to be released on a specific date.
  • Give clear instructions to incarcerated people, if they are denied release, on what they need to do to be released at the next hearing.
  • Require re-hearings in no more than 1 or 2 years.19
  • Provide case managers to help incarcerated people develop a plan to be successful at parole decision time.
  • Provide transparency to incarcerated people by sharing as much information as possible about how the parole board reached its decision.20
  • Provide transparency and accountability to the legislative branch by requiring annual reports on the numbers of, and reasons for, denials of parole, especially denials of individuals whose release has been recommended by guidelines supported by validated risk assessments.

Of course, those four state models have limitations that other states should be cautious about repeating:

  • Limiting presumptive parole to only certain offenses or for certain sentences. 21
  • Allowing parole boards to set aside official guidelines and deny release for subjective reasons.22

Second look sentencing

Second-look sentencing provides a legal mechanism for judges to review and modify individual sentences. The most effective way to do this is described in the newly revised Model Penal Code, published by the American Law Institute.23

The Model Penal Code recommends a process by which long sentences are automatically reviewed by a panel of retired judges after 15 years, with an eye toward possible sentence modification or release, and for subsequent review within 10 years, regardless of the sentence’s minimum parole eligibility date.24 This proposal also requires that state Departments of Corrections inform incarcerated people of this review, and provide staff resources to help them prepare for it.

Conceptual graphic showing how second-look sentencing impacts time served. Figure 4. “Second look” sentencing gives judges the power to reconsider long sentences, recognizing that our understanding of just punishment evolves over time, and allowing judges to adjust overly punitive sentences as appropriate.

To be sure, many states may have statutes that allow sentencing judges to reconsider an original sentence, although except for in Maryland,25 this doesn’t happen very often.

The reality is that people and societies change, as do views about punishment. Second-look provides the opportunity for judges to weigh the transformation of an incarcerated individual against the perceived retributive benefit to society of 15 years of incarceration.

Second-look is the only proposal in this report in which the judiciary would play a leading role, and that makes it particularly powerful tool in a reformist toolkit because polls show that people trust the judiciary much more than they trust the legislative or executive branches of government.26

Granting of good time

States can award credit to incarcerated individuals for obeying prison rules or for participating in programs during their incarceration. Commonly called things like “good time,” “meritorious credit” or something similar, these systems shorten the time incarcerated people must serve before becoming parole eligible or completing their sentences.

Conceptual graphic showing how good time impacts time served. Figure 5. This schematic of time served shows how a good time system can make people eligible for release on parole sooner and how it can make the ultimate completion of their sentence shorter. As this section describes, states vary greatly in how much good time they award and in what circumstances they award it.

States are unnecessarily frugal in granting good time and irrationally quick to revoke it. Good time should be granted to all incarcerated individuals, regardless of conviction and independent of program participation. Prisons should refrain from revoking accrued good time except for the most serious of offenses, and after five years, any good time earned should be vested and immune from forfeiture.

As the name implies, good time is doled out in units of time. Good time systems vary between states, as the National Conference of State Legislatures has previously discussed.27 In some states, the average amount of good time granted is negligible (North Dakota) or non-existent (Montana and South Dakota.) But in others, administrators are empowered by statute to award far more. For example:

  • Alabama can award up to 75 days for every 30 days served;
  • Nebraska can award six months per year of sentence, and can grant an additional three days per month for clean disciplinary records;
  • Oklahoma can award up to 60 days a month, plus additional credits for various kinds of positive disciplinary records, and a number of one-time grants for various educational or vocational accomplishments.

Procedures will vary by state and incarcerated people may not automatically be awarded the statutorily authorized maximum. In Texas, for example, the statute authorizes up to 45 days per 30 served, but the more typical amount awarded is 30, with the full amount reserved for people with non-violent sentences assigned to work outside the fence or in close proximity to correctional officers.

The most robust good time systems will:

  • Make good time eligible to every incarcerated person regardless of conviction, and ensure that every incarcerated person can apply good time toward initial parole or discharge. (For instance, Rhode Island prohibits individuals convicted of murder, sexual assault, child molestation, or kidnapping a minor from earning good time. And while Texas allows all individuals to earn good time, people with certain convictions are not allowed to apply it in the only two ways allowed - to lessen the time they must serve before initial parole eligibility or to shorten their actual time served.)
  • Fully fund any programs in which participation can result in receiving good time. For example, if drug treatment or educational classes make someone eligible for additional good time credits, there should not be a significant waiting list.28
  • Avoid the common pitfall of restricting valuable rehabilitative programs to only those close to release and low-risk and justifying those restrictions by pointing to lean budgets. This runs contrary to best practices, which say that “targeting high-risk offenders for intensive levels of treatment and services has the greatest effect on recidivism, and low-risk inmates should receive minimal or even no intervention.”29
  • Grant additional good time to people who are physically or mentally unable to take advantage of a program that gives good time. Many incarcerated people are mentally or physically incapable of engaging in programs, and anyone in that category should be awarded the maximum offered to those who can engage in programs.
  • Allow good time to be forfeited only for serious rule and law violations and allow forfeited good time to be restored. Texas, for example, prohibits the restoration of forfeited good time,30 while Alabama allows restoration by the Commissioner of the state Department of Corrections upon the warden’s recommendation.31 Finally, states should not allow one incident to result in a loss of good-time accrued over years, by vesting earned good-time after a certain period. We again rely on the Model Penal Code, which suggests good-time credits earned over five years be vested and untouchable.

Universal parole eligibility after 15 years

While many states will retain the option of imposing long sentences, their sentencing structures should presume that both individuals and society transform over time. This proposal uses the same 15-year timeline as proposed by the Model Penal Code for Second Look Sentencing discussed above.32

States will vary in how they structure sentences and how parole eligibility is calculated, but states should ensure that people are not serving more than 15 years without being considered for parole.

Conceptual graphic showing how broad changes to parole eligibility impact time served. Figure 6. Schematic graphic showing how states can shorten excessive prison sentences by making all incarcerated people eligible for parole at the 15-year mark.

Retroactive application of sentence reduction reforms

Sentences are determined based on the laws in place at the time the crime was committed. Unfortunately, when sentencing reform is achieved, it almost always applies only to future convictions. This means people currently incarcerated experience unequal justice and fail to benefit from progressive reform. Our statutes should be kept current with our most evolved understanding of justice, and our ongoing punishments like incarceration should always be consistent with that progress, regardless of when the sentence was originally imposed.33

For example, one significant sentencing reform that was not made retroactive was Congress’ modifications to the Anti-Drug Abuse Act of 1986, which created the infamous crack cocaine/powder cocaine disparity that treated possession of small amounts of crack cocaine as equivalent to possession of 100 times as much powder cocaine. Congress recognized that this law was based on irrational science and resulted in disproportionate arrests for people of color and changed it in 2010, but the reform was for new drug crimes only. People sentenced under the old law were forced to continue to serve sentences that were now considered unjust.34

Delaware passed a justice reform package in 2016 that not only reformed three- strikes laws but allowed those convicted on three-strikes statutes to apply for a modification of their sentences. Delaware took the common-sense step of making its reforms retroactive, but far too few legislatures do.

Historically, when sentencing reforms do grant relief to individuals already serving lengthy sentences, it is more often the result of a judicial order. (Courts make their decisions retroactive either by requiring states to change their laws, or by having the states erect frameworks for incarcerated people to apply for resentencing.)

For instance:

  • When the U.S. Supreme Court reversed an earlier decision and declared in 1963 that it was unconstitutional to put poor people on trial without first appointing them a lawyer, the Supreme Court ignored the State of Florida’s plea to not make the ruling retroactive.35 The Supreme Court did so knowing that it would apply to many thousands of people serving prison sentences in five southern states, including a substantial portion of Florida’s prison population.36
  • In 2002, the U.S. Supreme Court reversed its earlier decision and, in Atkins v. Virginia barred the execution of the intellectually disabled — the Court used the term “mentally retarded” — instructing that the Eighth Amendment’s prohibition against cruel and unusual punishment should be interpreted in light of the “evolving standards of decency that mark the progress of a mature society.”37 The Court did not define “mentally retarded,” leaving each state to devise its own standards. Over the next 11 years, at least 83 individuals condemned to die instead had their sentences reduced because of a finding of “mental retardation” stemming from Atkins.38
  • The Supreme Court has made other improvements in sentencing retroactive as well, including barring execution for offenses committed before age 1839 and barring mandatory life without parole sentences for offenses committed before age 18.40
  • State courts have also made changes retroactive. For example, in 2012 the Maryland Court of Appeals ruled in Unger v. Maryland that that jury instructions in capital murder convictions prior to 1981 were flawed and ordered new trials for the approximately 130 individuals still incarcerated with life sentences. (Most of those people were released by the state and placed on probation to great success.)

Elimination of parole revocations for technical violations

Parole supervision should focus on strengthening ties between individuals on parole and their communities. Unfortunately, the emphasis is more often on pulling parolees out of the community and returning them to incarceration at the first sign that they are struggling,41 with parole officers intent on “catching mistakes through surveillance and monitoring, rather than on promoting success via rehabilitation and support.”42 Parole officers have the power to return people to prison for “technical violations” that represent no threat to public safety and may simply indicate that a person on parole needs more assistance, or less stringent rules, not more incarceration.

Approximately 60,000 parolees were returned to state prisons in 2016 not because they were convicted of a new offense, but because of a “technical violation” such as missing a meeting with a parole officer or traveling to another state to visit a relative without permission. (Parole officers in Massachusetts can even re-incarcerate a parolee if they believe the person “is about to” engage in criminal behavior.43) For people who have already served years in prison and worked hard to earn their release, states should make sure that parole officers are supporting their reentry, rather than sending them back.

Parole revocations for technical violations are a problem in most states, but 10 states in particular were responsible for a majority of such revocations in 2016:

Graph showing that in 2016, 10 states were responsible for 67% of all parolees sent back to state prisons for technical violations of parole. Figure 7. These ten states accounted for two-thirds of all returns to incarceration for technical violations of parole conditions in 2016, according to the Bureau of Justice Statistics (BJS). See an alternate view of how much parole revocations for technical violations cost these states annually. Note that Wisconsin is not included in these graphs, because the state did not report any revocations to BJS; however, it reports 2,923 “revocation only” admissions to prison in 2016 (and even more in 2017) on its Department of Corrections website.

States should stop putting parolees behind bars for behaviors that, were the individual not on parole, would not warrant prison time. If a parole condition is itself a law violation, it can be dealt with by the criminal justice system. For example, a parole condition common to all states prohibits parolees from possessing firearms. Since states make it a criminal offense to be a felon in possession of a firearm, traditional criminal justice procedures can be brought to bear when a parolee is found with a firearm. All other, non-criminal violations should be addressed through community intervention and should never subject someone on parole to re-incarceration.

How often does your state re-incarcerate people on parole for technical violations?

See state-by-state data from 2016.

Source: Bureau of Justice Statistics, Probation and Parole in the United States, 2016, Appendix Table 7. For further state-level data, see the Appendix.
State Total number on parole, 2016 Total returns to incar­ceration, 2016 Total returns to incar­ceration that were for technical violations without a new offense, 2016
Alabama 8,562 473 58
Alaska 1,812 559 369
Arizona 7,500 2,488 2,472
Arkansas 23,792 6,003 5,741
California 93,598 n/a n/a
Colorado 10,186 3,226 2,397
Connecticut 3,379 878 0
Delaware 387 8 6
Florida 6,155 1,101 764
Georgia 22,836 2,561 22
Hawaii 1,367 336 336
Idaho 5,054 379 n/a
Illinois 29,248 8,340 6,570
Indiana 8,385 2,374 1,985
Iowa 6,051 1,502 789
Kansas 4,830 162 0
Kentucky 15,383 6,210 2,093
Louisiana 30,907 4,067 1,232
Maine 21 1 0
Maryland 10,305 1,009 516
Massachusetts 1,851 490 408
Michigan n/a n/a n/a
Minnesota 7,075 3,182 2,883
Mississippi 8,645 2,013 0
Missouri 17,792 6,613 3,439
Montana 1,074 228 206
Nebraska 1,088 429 416
Nevada 5,261 1,219 525
New Hampshire 2,436 797 797
New Jersey 15,128 1,485 1,387
New Mexico 2,780 1,644 1,172
New York 44,426 9,736 6,362
North Carolina 12,726 1,328 397
North Dakota 804 311 239
Ohio 19,634 1,722 127
Oklahoma 1,895 24 10
Oregon 24,711 2,906 2,041
Pennsylvania 111,087 11,595 5,302
Rhode Island 460 49 41
South Carolina 4,347 244 212
South Dakota 2,687 834 583
Tennessee 12,092 1,591 704
Texas 111,287 7,142 1,273
Utah 3,707 1,821 1,547
Vermont 935 n/a n/a
Virginia 1,650 140 35
Washington 11,322 1,728 760
West Virginia 3,550 330 272
Wisconsin 20,401 6,134 3,280
Wyoming 842 160 133
Total 741,451 107,572 56,621

Some states take great care to avoid sending people to prison on technical violations, but other states allow high rates of re-incarceration. In order to increase the likelihood that individuals on parole succeed, and to lighten the load on overwhelmed parole officers, states should adopt suggestions advanced by the Robina Institute44 and Columbia University Justice Lab:45

  • “Front-load” supervision resources immediately after release, when individuals released from prison are most likely to need support;
  • Tailor conditions to individual parolees instead of using boiler-plate language intended to cover every possible situation;
  • Limit the length of time an individual can be on parole regardless of sentence, and shorten parole terms by granting good-time for compliance with conditions.

Compassionate release

Compassionate release is the release of incarcerated individuals, usually but not exclusively aged, who are typically facing imminent death, and who pose no threat to the public. This process is often lengthy and cumbersome, which is unfortunate given that people recommended for compassionate release are almost always terminally ill or profoundly incapacitated and the complicated nature of this process means many die before their cases are resolved.46

All states but Iowa have a framework for compassionate release, but currently few states use compassionate release to a meaningful degree.47 The processes vary tremendously, but the basic framework is the same: An incarcerated person is recommended for release48 on compassionate grounds to prison administrators, who then solicit a medical recommendation, and then administrators or members of the parole board approve or deny a conditional release. These programs are plagued by many shortcomings, including:

  • Requirements that a person be extremely close to death, or so incapacitated that they do not understand why they are being punished.49
  • Requiring medical professionals to attest that someone is within six months, or nine months, of death. Health professionals are reluctant to give such exact prognoses, which means prison officials will default to the “it’s safer just to not let this person go.”50
  • Allowing prison personnel to overrule medical prognoses.51

To be sure, some states do certain facets of compassionate release better than others, but states would be wise to implement the recommendations52 of the Model Penal Code on compassionate release, along with FAMM’s excellent suggestions.53 Particularly robust compassionate release systems will:

  • Be available to all incarcerated people regardless of the underlying offense.
  • Streamline all processes and set reachable deadlines so that petitioners don’t die due to bureaucratic bottlenecks before they are released.
  • Limit the ability of prison officials to overrule, on medical grounds, a recommendation of release by medical professionals.


Commutations are modifications of a sentence by the executive branch to either make someone eligible for release before they otherwise would be, or to release them outright. These decisions are usually made by the governor, or some combination of the governor and a board, whose members are themselves often appointed by the governor. (For a detailed description of the process and structure in each state see The Criminal Justice Policy Foundation’s helpful summary.)

Definitions: “Pardons” vs. “clemency”

“Commutation,” “pardon” and “clemency” all mean slightly different things. We explain pardons and clemency and how they relate to commutation.

Pardons involve an official forgiveness and a restoration of most civil rights, and are - at least in the modern era - an inherently symbolic act.54 Clemency is an umbrella term that applies to both commutations and pardons. Because pardon and commutation decisions are often run by the same agencies, the 50-state comparison of the characteristics of pardon authorities published by the Rights Restoration Project55 can be helpful to people seeking to learn about commutation in their state.

The procedures are often very similar, but the outcomes vary greatly between the states. Typically, an incarcerated individual submits a petition to the governor’s office, who reviews the petition or forwards it to whatever board must make the initial recommendation. At that point, the petition is approved or denied based on whatever criteria that state uses.

There is not a comprehensive data source on the numbers of commutations granted across the 50 states, but it appears that clemency in general and commutation in particular are used far less than they have been in years past.56 Notable recent exceptions are former Illinois Gov. George Ryan (R), who in 2003 commuted the death sentences of all 167 individuals on death row to either life or a sentence of years, and Mike Huckabee (R), who as Arkansas governor issued 1,058 acts of clemency, many of them commutations and pardons to individuals with violent crimes.

Executives should consider using commutation in a broad, sweeping manner to remedy some of the extremes of the punitive turn that led to mass incarceration. Many executives have the power to shorten the sentences of large numbers of incarcerated individuals or to release them altogether. It will be tempting for governors to take caution from President Barack Obama’s methods,57 which were bogged down by bureaucratic, structural and political cautiousness. We suggest following the unique strategies of President Gerald Ford, who granted clemency to tens of thousands of men for evading the Vietnam War.58


If states are serious about reversing mass incarceration, they must be willing to leaven retribution with mercy and address the long sentences imposed during more punitive periods in their state’s history. This report provides state leaders with eight strategies to shorten overly long prison sentences. All that is left is the political will.

About the Prison Policy Initiative

The non-profit, non-partisan Prison Policy Initiative was founded in 2001 to expose the broader harm of mass criminalization and spark advocacy campaigns to create a more just society. The organization is known for its visual breakdown of mass incarceration in the U.S., as well as its data-rich analyses of how states vary in their use of punishment. The Prison Policy Initiative’s research is designed to reshape debates around mass incarceration by offering the “big picture” view of critical policy issues, such as probation and parole, pretrial detention, and reentry outcomes.

About the author

Jorge Renaud is a Senior Policy Analyst at the Prison Policy Initiative. He holds a Masters in Social Work from the University of Texas at Austin. His work and research is forever informed by the decades he spent in Texas prisons and his years as a community organizer in Texas, working with those most affected by incarceration.


This report would not have been possible without the expertise and input of many individuals. Laurie Jo Reynolds, Shaena Fazal, and Nora Demleitner offered crucial looks at parole systems during early drafts; and Alex Friedmann, Bernadette Rabuy, Eric Lotke, Janice Thompson, and Lois Ahrens all gave invaluable feedback. I am particularly indebted to Margaret Love for her work on commutations and pardons, Patricia Garin for providing leads on academic articles, to John Cooper of Safe and Just Michigan for keeping me updated about criminal justice reforms in that state, to Families Against Mandatory Minimums for their excellent work on compassionate release, and to Edward E. Rhine of the Robina Institute of Criminal Law and Criminal Justice, both for his scholarship there and for taking time to provide clarity about parole in all 50 states. Special thanks to Peter Wagner for offering much-needed clarity and shaping, to Wanda Bertram for editing, to Wendy Sawyer for visionary graphics, and to the rest of my colleagues at the Prison Policy Initiative.


  1. Ashley Nellis, Ph.D. “Still Life: America’s Increasing Use of Life and Long-Term Sentences.” The Sentencing Project.  ↩
  2. United States Department of Justice. Office of Justice Programs. Bureau of Justice Statistics. National Corrections Reporting Program, 1991-2015: Selected Variables. Ann Arbor, MI: Inter-university Consortium for Political and Social Research.  ↩
  3. This is not to ignore split sentencing, an alternative system that eliminates parole decisions entirely, whereby a defendant must both serve time in a jail or prison and then complete a period of community supervision on probation. This report focuses on shortening longer sentences, many for convictions that are not often eligible for split sentencing. Therefore, we will not address split sentencing here.  ↩
  4. The “earliest release date” is just that - the day at which someone can first leave confinement. Many factors can influence that - for instance, whether that date is set by statute or by a judge, and whether a state grants good time to incarcerated individuals. In states with “Truth in Sentencing” laws, one’s earliest release date is much closer to the expiration of the sentence, as such laws require an individual to serve up to 85 percent (or more) of a sentence before being considered for release; etc.  ↩
  5. See for example, the Robina Institute’s list, “Parole Boards with Indeterminate and Determinate Sentencing Structures” and a similar effort by the National Conference of State Legislatures, “Making Sense of Sentencing: State Systems and Policies” (p. 5).  ↩
  6. For example, some states make use of mandatory and/or discretionary parole periods, but exclude them for violent or sex-related offenses, meaning that people with the longest sentences will serve all, or almost all, of the maximum amount of time they can spend in prison. And some states grant the parole board discretion to deny individuals release for even these short terms of community supervision.  ↩
  7. In 2016, BJS reports that Illinois had 23,889 individuals who “entered” parole. Only 18 of those individuals were granted discretionary parole, as those individuals committed their crimes before 1978. Anyone convicted of crimes committed after 1978 in Illinois must serve a certain percentage of their sentence and is released under mandatory supervision, not discretionary parole. The same situation exists in Florida. In 2016, 6,110 people “entered” parole, with 34 of those being granted discretionary parole for crimes committed before 1983. The other 5,363 individuals were released under mandatory supervision. All of these individuals were subject to parole supervision and revocation, regardless of how they were released.  ↩

  8. In some states, the judge might impose a sentence of “10 to 30 years”; in other states, statutes might require the judge to impose a sentence of 30 years, with the unspoken understanding that parole would be an option after 10 years. This difference in statutory structure around how sentences are expressed is not relevant to our conceptual explanation of how felony sentences are served, as illustrated in Figure 2.  ↩

  9. Objective data is often static, consisting of an individual’s criminal history, times arrested and incarcerated, previous probation and parole history, age at first arrest, and increasingly the outcome of a validated risk assessment tool. But subjective criteria tends to win out, as states deny parole on such factors as “lack of insight or insufficient remorse” (MI) or on the “seriousness and nature of the offense” (TX) and demand that the release will not “depreciate the seriousness of the crime so as to undermine respect for the law,” (NY) despite the fact that the seriousness of the crime was taken into account while sentencing, overwhelmingly in negotiations between prosecutors and defense attorneys.  ↩

  10. Hollywood often portrays parole hearings as a meeting between a supplicant incarcerated person and stone-faced parole board members. Reality often is less cinematic and less accountable. In many states, a staffer is sent to interview the prospective parolee and the board then reviews the staffer’s reports. Incarcerated people can be eligible for parole many times, be denied and finally granted release, without ever having spoken to a voting member of the board.  ↩

  11. The Robina Institute conducted an invaluable 2016 survey of chairpersons of state parole boards, asking them to rank the factors they considered when reviewing an individual for release, finding that “Nature of the present offense” and “Severity of current offense” were, on average, ranked as the first and second most important factors. By contrast, participation in prison programming was ranked 6th and “inmate’s demeanor at hearing” and “inmate testimony” were ranked 13th and 14th. E. L. Ruhland, E. E. Rhine, J. P. Robey, & K. L. Mitchell. (2016). “ The Continuing Leverage of Releasing Authorities.” Robina Institute of Criminal Law and Criminal Justice, the University of Minnesota. P. 4.  ↩

  12. South Dakota Codified Laws 24-15-8(1).  ↩

  13. New Hampshire Admin. Code Rules, Par. 301.03  ↩

  14. Utah Code S 77-27-5.3(2)  ↩

  15. PB Rule 82.2 22.510.3.8 C(2)(s) (full text here).  ↩

  16. Edward E. Rhine is the former director of the Parole Release and Revocation Project at the Robina Institute of Criminal Law and Criminal Justice, University of Minnesota. He is currently a Lecturer in the Sociology Dept. at The Ohio State University and was awarded the Association of Paroling Authorities Vincent O’Leary Award for 2018. Joan Petersilia is the Faculty Co-Director of the Stanford Criminal Justice Center, a past recipient of the Stockholm Prize in Criminology, a former president of the American Society of Criminology and is a world-renowned expert on prison reentry issues. Kevin R. Reitz is the James Annenberg La Vea Land Grant Chair in Criminal Procedure Law at the University of Minnesota and co-author of The Challenge of Crime: Rethinking Our Response (Harvard University Press, 2003). He was also appointed by the American Law Institute to be Reporter for the first-ever revision of the Model Penal Code, focusing on sentencing and corrections provisions within the Code.

    Edward E. Rhine, Joan Petersilia, and Kevin R. Reitz, “The Future of Parole Release,” Crime and Justice 46 (2017): 279-338. In Reinventing American Criminal Justice. (eds.) Michael Tonry and Daniel S. Nagin. Vol. 46 of Crime and Justice: A Review of Research. Edited by Michael Tonry. Chicago: University of Chicago Press.  ↩

  17. New Mexico Gov. Susana Martinez fired two parole board members who complained that other members were not truly considering individuals serving life sentences who, by statute, were eligible for parole. Terrell, Steve. June 26, 2012. “Governor shakes up parole board amid dispute over option for lifers’ release.” Santa Fe New Mexican.  ↩

  18. Much more common is parole officials being fired after someone commits a crime. See Jonathan Saltzman, Jan. 14, 2011. “Five out as governor overhauls parole board.”The Boston Globe; and D. Stamm, L.A. Cahn and D Chang. Dec. 3, 2012. “Parole employees fired after death of cop: ‘We’re scapegoats.’” NBC 10.  ↩

  19. By contrast, Texas allows for 10 years to lapse between parole reviews for some individuals. See Texas Gov. Code, Title 4, Chapter 508.141. (g-1)  ↩

  20. Of these four states, only New Jersey and Michigan allow individuals in the parole process to view the evidence the Board uses to deny them parole, although that procedure seems to be fairly common in other states. See Robina Institute: Profiles in Parole Release and Revocation.  ↩

  21. In New Jersey, only individuals who are prohibited from parole entirely are ineligible for presumptive parole. New Jersey statute mandates that an incarcerated person “shall be released on parole at the time of parole eligibility.” See N.J. Rev. Stat. S 30:4-123.53.  ↩

  22. Aside from state guidelines on when someone should be released, each of these states use a “risk assessment” tool as a part of their process. Three of the states (Hawai’i, Michigan, and Mississippi) require the tool to be scientifically validated. Unfortunately, three of the four states also allow parole boards to overrule, for subjective reasons, a risk assessment finding that someone is “low risk” and should be paroled. (Michigan passed legislation that expressly prohibits denial of parole for subjective reasons.)  ↩

  23. The influential American Law Institute originally published the Model Penal Code in 1962 as a guide to federal and state statute-writing bodies, and the 2017 revisions includes a Second Look provision. At least two-thirds of American statutory-writing bodies use the Code as a blueprint and follow the Code’s suggested structures and vocabulary in their own criminal codes. Margaret Love, Dec. 13, 2016. “No Second Thoughts about Second Look Sentencing.” The ALI Advisor.  ↩

  24. Kevin R. Reitz, Cecelia Klingele. Model Penal Code: Sentencing, Proposed Final Draft. (2017). P. 565.  ↩

  25. In Maryland, the Revisory Power of the Court limits judges to reviewing sentences within the first five years. The Model Penal Code approach is much more comprehensive as the American Law Institute explains: “No provision closely similar to S305.6 exists in any American jurisdiction.” MPC, Art. 305.6, Comment: A. Scope.  ↩

  26. Frank Newport, Sept. 26, 2012. “Americans trust judicial branch most, legislative least.” Gallup.  ↩

  27. See the National Conference of State Legislatures’s helpful table, Good Time and Earned Time Policies for State Prison Inmates.  ↩

  28. Washington State, for example, has over 1,000 individuals waiting for jobs and classes within the state’s prisons system, both of which offer earned-time credits.  ↩

  29. Nathan James, “Risk and Needs Assessment in the Federal Prison System.” Congressional Research Service, 2018. P. 15.  ↩

  30. TX Gov. Code S 498.004  ↩

  31. AL Code S 14-9-41 (f)(1) and (2)  ↩

  32. Prior to the mid to late 1980s, individuals sentenced to life in the United States rarely served more than 15 to 20 years. To provide just two examples, according to The Sentencing Project, judges polled in 2002 by the Michigan State Bar believed that lifers eligible for parole would serve 20 years or less, and in New Mexico, only 10 years was necessary before eligibility for parole was reached until 1986. (M. Mauer, R. S. King, and M. C. Young. (2004) “The Meaning of ‘Life’: Long Prison Sentences in Context.”)

    And in Northern Europe, individuals rarely serve lengthy sentences, even for murder. A California State University study comparing time served by individuals sentenced to life in three Northern European countries found that in Denmark, the most common determinate sentences given for murder were between 12 to 16 years, and in Sweden, sentences for murder averaged between 10 to 18 years. (Doris Schartmueller, 2018. “How Long is Life? Comparing the Processes of Release for Life-Imprisoned Offenders in Denmark, Finland, and Sweden.”)  ↩

  33. Of course, the ex post facto clause of the constitution prohibits making a sentence more punitive, but nothing in the constitution prohibits, and common decency should require, that sentences that would be less harsh if imposed today be made less harsh if they are still being served today.  ↩

  34. Laws to make 2010’s “Fair Sentencing Act” retroactive — like the “Smarter Sentencing Act” — are proposed in each Congressional session but have not yet passed. The only positive news is that the U.S. Sentencing Commission - an agency of the federal judiciary - recommended in 2014 that federal judges allow a portion of those sentenced under the old law to petition for resentencing, which resulted in approximately 6,000 individuals being released from federal prison in 2015.  ↩

  35. For Florida’s plea against retroactivity see Florida Assistant Attorney General Bruce R. Jacob’s last comment in the brief submitted to the Supreme Court, quoted in chapter 10 of Anthony Lewis’ Gideon’s Trumpet: “Jacob ended with a cautionary plea. ‘If the Court should decide to overrule Betts,’ he said, ‘respondent respectfully requests that it be accomplished in such way as to prevent the new rule from operating retrospectively.’ In other words, the newly defined right to counsel should not apply to persons already in prison - presumably including Clarence Earl Gideon.” The Supreme Court made their decision apply to Gideon, and explicitly made Gideon retroactive in Burgett v. Texas in 1967.  ↩

  36. To our knowledge, there is not an accessible definitive count of the number of people ultimately released by Gideon. Anthony Lewis, in Chapter 13 of Gideon’s Trumpet, says that Florida had, by Jan. 1, 1964, outright released 976 as a result of the Gideon decision, with 500 in the courts and hundreds awaiting hearings. (In 1960, according to the Bureau of Justice Statistics, there were 7,703 persons incarcerated in state prisons in Florida.)  ↩

  37. Atkins v. Virginia, 2002.  ↩

  38. Those 83 individuals were scattered across the Death Rows of 20 states, with North Carolina reducing the sentences of 16 condemned individuals, Texas 12, and Pennsylvania 10. This likely doesn’t mean North Carolina had a higher percentage of intellectually disabled individuals on Death Row, but that North Carolina was perhaps more willing to revise its statute to reflect true disability. See Death Penalty Information Center.  ↩

  39. Roper v Simmons (2005) changed the sentences of all individuals who’d been convicted as juveniles from death to life in prison. At the time, there were 71 individuals on death row in 12 states, 29 of them in Texas, 13 in Alabama, with no other state holding more than five. See Death Penalty Information Center.  ↩

  40. Miller v Alabama (2012) made automatic life without parole for juvenile offenders unconstitutional, and this decision was made retroactive in Montgomery v. Louisiana (2016), making 2,300 individuals eligible for parole or review.  ↩

  41. Many correctional systems are, in fact, aware that supervision as it currently practiced often does more harm than good. For example, Massachusetts admits that supervision itself will result in “a higher likelihood of re-incarceration.” Massachusetts Department of Corrections, Prison Population Trends, 2016. P. 48.  ↩

  42. The Pew Charitable Trusts brought much-need attention to the rarely scrutinized, barely understood role that community supervision plays in mass incarceration in its Sept. 2018 report, “Probation and Parole Systems Marked by High Stakes, Missed Opportunities.”  ↩

  43. “Revocation rates are also affected by the policies and practices of the supervising agency and the idiosyncrasies of individual parole officers. If revocations are triggered by less serious forms of misconduct, for example, or if the standard of proof at revocation hearings is low, parolees are on average more likely to be returned to prison.” Mariel E. Alper. “By The Numbers: Parole Release and Revocation Across 50 States.” Robina Institute of Criminal Law and Criminal Justice, University of Minnesota. 2016. P. 6.  ↩

  44. C. C. McVey, E. E. Rhine, and C. V. Reynolds, Robina Institute of Crim. Law & Crim. Just., “Modernizing Parole Statutes: Guidance from Evidence-Based Practice.” 2018.  ↩

  45. V. Schiraldi and J. Arzu, Columbia University Justice Lab. “Less is More in New York: An Examination of the Impact of State Parole Violations on Prison and Jail Populations.” 2018.  ↩

  46. In Georgia, for example, between 2011 and 2016, at least 14 incarcerated individuals died while awaiting review and another 16 died awaiting release after their petitions were approved. “High costs, not human rights, forces Georgia to release its sickest prisoners.” Prison Legal News, Aug. 23, 2016.  ↩

  47. Gina Barton, April 18, 2018. “Release programs for sick and elderly prisoners could save millions. But states rarely use them.” Milwaukee Journal Sentinel.  ↩

  48. Some states allow only family and attorneys to recommend that someone be released on these grounds; others allow prison personnel to do so. Again - there is a huge difference between states, but for a truly comprehensive view of state policies, go to FAMM’s detailed breakdown on state practices in their report. Mary Price, May 2018. “Everywhere and Nowhere: Compassionate Release in the States.” Families Against Mandatory Minimums.  ↩

  49. One of the criteria that qualified an individual incarcerated in Hawai’i for consideration for release on medical grounds is if “The inmate is too ill or cognitively impaired to participate in rehabilitation and/or to be aware of punishment.”  ↩

  50. In a unique positive development, new guidelines for federal prisons issued by the U.S. Sentencing Commission do not require a short-term “terminal” prognosis, as an excellent article in Health Affairs points out. B. Williams, A. Rothman, and C. Ahalt. “For Seriously Ill Prisoners, Consider Evidence-Based Compassionate Release Policies.”  ↩

  51. The Marshall Project reviews the failures of the federal Bureau of Prisons to operate a functional compassionate release system, citing many examples of doctors giving prognoses to individuals of a few months to live, prison officials disagreeing, and the individuals dying shortly thereafter, once two days after prison officials proclaimed that person had 18 months left to live. Christie Thompson, March 6, 2018. “Old, sick, and dying in shackles.” The Marshall Project.  ↩

  52. Model Penal Code: Sentencing S 305.7. The American Law Institute.  ↩

  53. Families Against Mandatory Minimums (FAMM) has similar recommendations. Mary Price, May 2018. “Everywhere and nowhere: Compassionate release in the states.” Families Against Mandatory Minimums.  ↩

  54. In 1977, Gov. Michael Dukakis of Massachusetts pardoned famous anarchists Nicola Sacco and Bartolomeo Vanzetti who were executed by the state in 1927. In 2001, New York Governor George Pataki pardoned deceased comedian Lenny Bruce for a 1964 conviction; and in 2010, Florida Governor Charlie Crist pardoned the late The Doors frontman Jim Morrison for a 1969 conviction. And just in 2018, President Donald Trump pardoned Jack Johnson, the first African-American heavyweight champion, 105 years after he was convicted of violating the Mann Act. While pardons are still used by many states to restore the rights of individuals with more recent convictions, in many ways they are a political statement by the executive granting the pardon.  ↩

  55. See the Restoration of Rights Project, “Characteristics of Pardon Authorities.”  ↩

  56. Maggie Clark. Feb. 8, 2013. “Governor’s pardons becoming a rarity.” GOVERNING, the State and Localities.  ↩

  57. President Obama did pick up the pace at the very end of his presidency, commuting 1,715, but fell far short of the 10,000 he was aiming for. Most clemency efforts do not meet their goals because they keep prosecutors - who are institutionally interested in supporting their own convictions - in key decision making positions; are overly focused on redemption; or are simply too bureaucratic. On that last front, Obama made his problem worse: Rather than reduce the bureaucracy inherent in seven layers of government review, he added an additional layer, with a process by which non-government volunteers also reviewed applications.  ↩

  58. Mark Osler. Apr. 1, 2016. “Obama’s Clemency Problem.” The New York Times.  ↩

Read the footnotes


States vary widely in their use of long sentences, their release systems, and their appetites for reform. For state advocates, journalists and policymakers looking for more individualized information, we have compiled fact sheets for all 50 states. See your state:

Table 1. The table below provides the data behind most of the graphics in this report and the fact sheets above. In some places in the report or fact sheets, we have selected a common reference point that allows us to compare all 50 states (for example, 2005 as a starting point when measuring change over time). However, many states have data beginning at earlier or later points in the same data sources; this table includes the earliest available data for each state rather than the common reference points used elsewhere to compare states.

State Total individuals with at least 10 years in prison, 2015 Percent change, total individuals with at least 10 years in prison from Year Began* date through 2015 Year Began Percent of total state prison population who were individuals with at least 10 years in prison, 2015 Percent change, percent of total state prison population who were individuals with at least 10 years in prison from Year Began* through 2015 Total parole population, 2016 Total returns to incarceration, 2016 Number of individuals returned to prison for technical violations, no new offense, 2016 Percent of all returns to prison that were technical violations, no new offense, 2016 Does the state’s parole board determine release date? Percentage of individuals who were eligible for parole whose release was granted in 2014.
Alabama 4,853 39% 2007 18% 5% 8,150 473 58 12% Yes 34%
Alaska 331 -3% 2002 9% -5% 2,100 559 369 66% Yes 40%
Arizona 3,516 161% 2000 8% 2% 7,379 2,488 2,472 99% Not since 1994 11%
Arkansas 2,510 129% 2001 16% 11% 22,910 6,003 5,741 96% Yes 67%
California 33,629 75% 1999 26% 14% 86,053 Not since 1977 18%
Colorado 2,599 116% 2000 13% 6% 9,953 3,226 2,397 74% Yes 26%
Connecticut 1,506 103% 2000 14% 8% 2,939 878 0 0% Yes 67%
Delaware 727 57% 2009 14% 5% 425 8 6 75% Not since 1990 n/a
Florida 17,444 172% 1999 18% 8% 4,611 1,101 764 69% Not since 1983 0%
Georgia 8,667 186% 1999 17% 9% 24,413 2,561 22 1% Yes 56%
Hawaii n/a 0% 1,479 336 336 100% Yes 32%
Idaho 617 27% 2010 9% 2% 4,875 379 0 Yes 68%
Illinois 6,506 122% 1999 14% 7% 29,629 8,340 6,570 79% Not since 1978 0%
Indiana 3,183 87% 2002 12% 4% 9,420 2,374 1,985 84% Not since 1977 n/a
Iowa 1,089 -36% 2001 12% -11% 5,901 1,502 789 53% Yes 51%
Kansas 1,349 7% 2011 14% 0% 4,331 162 0 0% Not since 1993 29%
Kentucky 1,770 108% 2000 8% 3% 16,536 6,210 2,093 34% Yes 52%
Louisiana 6,915 74% 2002 20% 5% 31,187 4,067 1,232 30% Yes 42%
Maine 177 -1% 2012 8% -1% 21 1 0 0% Not since 1976 n/a
Maryland 4,010 27% 2000 20% 7% 10,887 1,009 516 51% Yes 40%
Massachusetts 1,761 8% 2009 19% 4% 1,995 490 408 83% Yes 63%
Michigan 9,656 60% 1999 19% 6% Yes 68%
Minnesota 805 223% 2000 8% 4% 6,810 3,182 2,883 91% Not since 1982 n/a
Mississippi 2,965 216% 2000 16% 10% 8,424 2,013 0 0% Yes n/a
Missouri 4,500 95% 2000 14% 6% 17,657 6,613 3,439 52% Yes n/a
Montana 294 16% 2010 16% 6% 1,092 228 206 90% Yes 25%
Nebraska 581 68% 2000 11% 2% 1,050 429 416 97% Yes 87%
Nevada 1,875 35% 2008 14% 3% 5,507 1,219 525 43% Yes 56%
New Hampshire 301 8% 2011 11% 0% 2,451 797 797 100% Yes 80%
New Jersey 3,058 44% 2000 14% 7% 15,180 1,485 1,387 93% Yes 33%
New Mexico 743 65% 2010 10% 4% 2,763 1,644 1,172 71% Not since 1979 n/a
New York 8,812 52% 1999 17% 9% 44,562 9,736 6,362 65% Yes n/a
North Carolina 6,025 107% 2000 17% 7% 11,744 1,328 397 30% Not since 1994 n/a
North Dakota 75 142% 2002 4% 2% 634 311 239 77% Yes n/a
Ohio 8,007 49% 1999 15% 3% 18,284 1,722 127 7% Not since 1996 7%
Oklahoma 3,989 36% 2000 14% 1% 2,116 24 10 42% Yes n/a
Oregon 1,718 87% 2001 12% 4% 24,077 2,906 2,041 70% Not since 1989 n/a
Pennsylvania 7,856 108% 1999 16% 5% 112,351 11,595 5302 46% Yes 58%
Rhode Island 262 114% 2000 10% 3% 441 49 41 84% Yes n/a
South Carolina 4,191 135% 2000 20% 11% 4,963 244 212 87% Yes n/a
South Dakota 425 7% 2013 12% 1% 2,673 834 583 70% Yes n/a
Tennessee 4,712 120% 2000 15% 6% 13,063 1,591 704 44% Yes n/a
Texas 23,233 37% 1999 15% 3% 111,892 7,142 1,273 18% Yes 36%
Utah 436 100% 2000 7% 3% 3,502 1,821 1,547 85% Yes n/a
Vermont 20 185% 2003 1% -3% 1,083 Yes n/a
Virginia 7,489 16% 2009 21% 2% 1,576 140 35 25% Not since 1995 n/a
Washington 2,474 148% 2000 14% 7% 11,131 1,728 760 44% Not sinc 1994 42%
West Virginia 936 47% 2006 15% 4% 3,123 330 272 82% Yes 36%
Wisconsin 3,610 315% 2000 16% 12% 20,241 5,424 3,280 60% Not since 2000 n/a
Wyoming 234 25% 2006 10% 1% 783 160 133 83% Yes 63%
MEAN 86% 57% 42%
Total individuals with at least 10 years in prison, 2015
Number of individuals in a state prison for at least 10 years at the end of 2015. Source: Bureau of Justice Statistics, National Corrections Reporting Program, 1991-2015: Selected Variables, Year-End Population. Ann Arbor, MI: Inter-University Consortium for Political and Social Research [distributor], 2018-03-02
Percent change, total individuals with at least 10 years in prison from Year Began* date through 2015
Change in the number of individuals held in a state prison for at least 10 years, beginning in 2005 through the end of 2015. Source: Bureau of Justice Statistics. National Corrections Reporting Program, 1991-2015: Selected Variables, Year-End Population. Ann Arbor, MI: Inter-university Consortium for Political and Social Research [distributor], 2018-03-02
Year Began
The year that each state in chart began collecting data as to the number of individuals who had served at least 10 years in prison. Source: Bureau of Justice Statistics. National Corrections Reporting Program, 1991-2015: Selected Variables, Year-End Population. Ann Arbor, MI: Inter-university Consortium for Political and Social Research [distributor], 2018-03-02
Percent of total state prison population who were individuals with at least 10 years in prison, 2015
Individuals with at least 10 years in prison as a percentage of the entire state prison population at the end of 2015. Source: Bureau of Justice Statistics. National Corrections Reporting Program, 1991-2015: Selected Variables, Year-End Population. Ann Arbor, MI: Inter-university Consortium for Political and Social Research [distributor], 2018-03-02
Percent change, percent of total state prison population who were individuals with at least 10 years in prison from Year Began* through 2015
Percent change of individuals with at least 10 years in prison as a percentage of the entire state prison population, as compared with that same population when that state began collecting that data. Source: Bureau of Justice Statistics. National Corrections Reporting Program, 1991-2015: Selected Variables, Year-End Population. Ann Arbor, MI: Inter-university Consortium for Political and Social Research [distributor], 2018-03-02
Total parole population, 2016
The number of adults on parole in a given state on Jan. 1, 2016. Source: Bureau of Justice Statistics, Probation and Parole in the United States, 2016, Appendix Table 5.
Total returns to incarceration, 2016
The number of individuals who were on parole in a given state and returned to prison for 1) a new offense, 2) a revocation of parole on a technical violation, no new offense; 3) for treatment, and 4) unknown/other. Source: Bureau of Justice Statistics, Probation and Parole in the United States, 2016, Appendix Table 7.
Number of individuals returned to prison for technical violations, no new offense, 2016
The number of individuals a state returned to prison for a violation of conditions of parole without convicting them of a new, separate offense. Source: Bureau of Justice Statistics, Probation and Parole in the United States, 2016, Appendix Table 7.
Percent of all returns to prison that were technical violations, no new offense, 2016
The individuals that a state returned to prison for a violation of conditions of parole as a percentage of all individuals on parole who were returned to prison. Source: Bureau of Justice Statistics, Probation and Parole in the United States, 2016.
Does the state’s parole board determine release date?
Whether or not a state has statutorily done away with indeterminate sentencing and does not have a parole board with the power to grant discretionary parole. Source: E. E. Rhine, A. Watts, and K. R. Reitz. (2018) “Parole Boards within Indeterminate and Determinate Sentencing Structures.” Robina Institute of Criminal Law and Criminal Justice. University of Minnesota.
Percentage of individuals who were eligible for parole whose release was granted in 2014.
The number of individuals who were granted release by a state‚Äôs parole board as a percentage of the number of incarcerated individuals in that state who were eligible for parole and reviewed for parole. Source: Mariel E. Alper. (2016) “By the Numbers: Parole Release and Revocation Across 50 States.” Robina Institute of Criminal Law and Criminal Justice. University of Minnesota.

Stay Informed

Get the latest updates:

Share on 𝕏 Donate


Not near you?
Invite us to your city, college or organization.