States Can Shorten Probation and Protect Public Safety

This report was updated in April 2021 to reclassify probation sentences in Georgia.

Overview

More than 3.5 million, or 1 in 72, adults were on probation in the United States at the end of 2018—the most recent year for which U.S. Bureau of Justice Statistics (BJS) data is available—more than triple the number in 1980.¹ Nationwide, on any given day, more people are on probation than in prisons and jails and on parole combined.

At its best, probation—court-ordered correctional supervision in the community—gives people the opportunity to remain with their families, maintain employment, and access services that can reduce their likelihood of reoffending while serving their sentences. But, as previous research by The Pew Charitable Trusts has shown, the growth and size of this population have overloaded local and state agencies and stretched their resources thin, weakening their ability to provide the best return on taxpayers’ public safety investments, support rehabilitation, and ensure a measure of accountability.² One key factor driving the size of the probation population is how long individuals remain on supervision.

A growing list of high-quality studies have shown that long probation sentences are not associated with lower rates of recidivism and are more likely than shorter ones to lead to technical violations—noncompliance with one or more supervision rules, such as missing appointments or testing positive for drug use. Recent research from the Council of State Governments has found that such violations contribute significantly to state incarceration rates and correctional costs: More than 1 in 10 state prison admissions are the result of technical violations of probation rather than convictions for a new crime.

To date, the average length of probation has not been well documented, because data on individual terms has been lacking.³ To begin addressing this gap and help criminal justice stakeholders better understand how long people spend on probation—as well as the effects of term length on individual recidivism outcomes— The Pew Charitable Trusts conducted an in-depth analysis of BJS data from 2000 through 2018. Additionally, Maxarth LLC examined Oregon and South Carolina data to quantify the potential to reduce probation lengths without increasing re-offending in those states, and the National Conference of State Legislatures (NCSL) reviewed probation sentencing statutes in all 50 states. This report provides a first-of-its-kind national and state-level portrait of the average length of probation and explores whether term lengths can be safely reduced and what options are available for state policymakers looking to improve their system’s outcomes.

Key findings:

• Nationwide, the average probation term is just under two years, with substantial variation across states. Average terms range from nine months in Kansas to nearly five years (59 months) in Hawaii.
• Although the average total time spent on probation declined nationally by roughly 22 days, or about 3%, between 2000 and 2018, it increased in 28 states. In 14 states, average total time on probation rose by more than a third over that span. Across the 250,000 people who exited probation in those 14 states in 2018, that increase adds up to almost 300,000 additional years spent under supervision, in the aggregate.
• Length of time spent under supervision is a key factor in the size of the probation population. Fifteen of the 22 states that cut their average probation terms between 2000 and 2018 also saw declines in their probation populations. Further, in three of those 15 states, the supervision populations fell even though the number of people entering probation increased. These findings demonstrate that shortening supervision terms is a critical component of efforts to reduce probation populations, allowing agencies to redirect public safety resources to provide the greatest benefit—particularly toward the early months and individuals at highest risk of re-offending.
• Many people on supervision serve longer terms than are necessary for public safety. An analysis of data from Oregon and South Carolina showed that, among people who were on probation for a year without being arrested, more than 90% could have spent less time on supervision without an impact on recidivism (as measured by re-arrests). Had these individuals served the shortest supervision terms needed to minimize re-offending, the average probation length in South Carolina would have been shortened from 26 to 18 months and in Oregon from 24 to 14 months, without an associated increase in arrests. These reductions would have cut the two states’ average daily populations (ADPs) on supervision by 32% and 44%, respectively, with the declines driven largely by people whose probation terms could be reduced by two or more years. Oregon and South Carolina were chosen for this analysis because of their ability and willingness to provide person-level arrest and probation data and because their average supervision term lengths during the study period (24 and 26.2 months, respectively) were close to the national average of 22.4 months. Although these states may not be representative of the U.S. as a whole, the findings are nevertheless sufficiently compelling to support efforts by other jurisdictions to review their own data and evaluate whether their probation terms are longer than is necessary to protect public safety.
• Maximum allowable probation sentences vary substantially across states, and few states have statutes that provide for early release. Judges often make decisions about when or whether to terminate supervision early. Therefore, how long people serve on probation may frequently be determined by jurisdiction-specific factors, such as judicial philosophy and culture. This analysis also identified a set of approaches that policymakers can consider as they seek to shift the focus of probation from length of time to achievement of goals related to risk reduction (e.g., rehabilitative programs, drug treatment, and employment), which can in turn safely shorten probation terms, lower recidivism rates, and bolster public safety. These include:
  • Implementing goal-based instead of time-based supervision. Under this approach, probation term lengths are determined by the time necessary for a person to meet obligations and complete programming.
  • Instituting earned time credits. These and other early discharge programs allow people on probation to reduce their supervision term through regular compliance with rules and requirements.
  • Creating mandatory periodic reviews of probation. Administrative or court assessment policies require supervision officers or judges to regularly evaluate people’s progress toward changing their behavior and gauge their readiness to be released, possibly earlier than initially planned, from probation.

No national standard exists for how long probation should be for any given case. Rather, the findings of this and other research suggest that probation should be only long enough to meet its basic objectives of providing accountability proportional to the underlying criminal offense, connecting people to needed treatment and services, and enabling individuals to complete programs such as cognitive behavioral therapy and counseling that have been shown to reduce the risk of re-offending.⁴

Research indicates that people are at the highest risk of re-offending early in their probation terms; for example, among people on felony probation in Oregon who were re-arrested within three years of entering probation, 69% were arrested in the first year.⁵ Further, studies show that after the first year, many supervision provisions, such as reporting requirements and community-based services, have little effect on the likelihood of re-arrest, so keeping probation terms short and prioritizing resources for the early stages of supervision can help improve success rates among people on probation, reduce officer caseloads, and protect public safety.

Although probation was originally conceived as an alternative to incarceration, criminal justice officials, policymakers, and other stakeholders increasingly acknowledge that keeping people on probation longer than is needed to deliver public safety benefits carries unnecessary and unproductive costs and wastes scarce resources.⁶ This report aims to help state and local leaders better understand and address the critical issue of probation length by providing essential data and offering policies and practices that can improve outcomes for probation departments and the people they supervise across the U.S.

Probation lengths across the U.S.

Probation lengths in 2018

The national average probation term length in 2018 was 22.4 months. Analysis of 2018 average lengths shows significant variation among states: Average probation lengths ranged from just nine months in Kansas to 59 months, or close to five years, in Hawaii. (See Figure 1.)

Change in probation lengths since 2000

Although the 2018 national average term of 22.4 months is about three weeks shorter than the 2000 figure of 23.2 months, average length of probation increased in more than half of the states (28) from 2000 to 2018, with average terms more than doubling in three of those states.⁹ (See Table 1.)

Mean probation terms rose by more than a third in 14 states between 2000 and 2018, growing by an average of 14 months. When multiplied by the quarter-million people who exited supervision in 2018 in those states, this increase represents nearly 300,000 additional years spent on supervision (292,716).

Table 1

Average Probation Terms and Change in Length Varied Widely Across States

Probation lengths and rate by state, 2000 and 2018

Note: Average probation terms are rounded to nearest month; percentage changes were calculated using unrounded figures.

Sources: Bureau of Justice Statistics, “Annual Probation Survey and Annual Parole Survey” (2000-2018)

© 2020 The Pew Charitable Trusts

Nine of the 10 states with the longest probation lengths in 2018 increased their average probation terms from 2000, while 9 of the 10 states with the shortest lengths in 2018 either shortened or approximately maintained their average term since 2000. In the 28 states where average probation length grew between 2000 and 2018, the mean rise was 42%, but increases ranged from 0.3% in Nevada to 109% in Idaho. (See Figure 2.) Three states—Idaho, Kentucky, and New Jersey—experienced increases of 100% or more. Among states with lower average terms, the mean decline was 25%, with decreases ranging from -4% in Vermont to -49% in Oregon and Georgia.

Probation population and term lengths

Because the number of people on probation is a function of how many enter supervision and how long they stay, changes in the length of terms affect probation populations. From 2000 to 2018, probation populations fell in 15 of the 22 states that reduced the lengths of their supervision terms. And three of those 15 saw their populations decrease even though the number of people entering probation increased. Further, of the 27 states where populations were higher in 2018 than in 2000, about three-fourths (20) also had longer term lengths, and in 12 of these states, those population increases occurred despite declining admissions. (See Appendix Table A.1 for state-level data.)

This data shows that cutting the length of supervision can play an important role in shrinking probation populations. And reducing the number of people on supervision can allow agencies to direct resources where they can have the biggest impact on public safety. For example, research has shown that when officers have smaller caseloads and implement evidence-based practices, it can reduce recidivism.¹⁰

The causes of growth or decline in states’ probation term lengths are complex. New judges may have different judicial philosophies than their predecessors about appropriate probation sentence length, how to address technical violations, and whether to grant early discharge. State reform efforts may change the offenses that are eligible for probation, policies that enable people to earn time off their probation for compliance, or limits on the circumstances under which someone’s supervision could be revoked. Supervision agencies may adopt revised rules that curtail or extend supervision, or patterns of offending may shift in a given state. In any given state, multiple factors probably interact, but an analysis of those dynamics is beyond the scope of this study.

Data limitations leave unanswered questions

Notably, BJS’ figures lack sufficient detail to fully illuminate the length of probation in the states. For instance, when terms are viewed in the aggregate, large numbers of people with short probation terms—those of less than a year—could obscure the smaller number who remain on probation for decades.¹¹ Further, because the statistics do not break out probation lengths by successful or unsuccessful completion, they do not reveal whether a drop in a state’s average term length is the result of reforms that allow people who remain crime-free to be safely released early or of high rates of violations leading to early termination of supervision and incarceration.

Aggregate data also does not show the significant variability in probation length across and even within jurisdictions. People may spend different amounts of time on probation for the same offense based on their locations, the judges who sentence them, or the supervising agencies’ practices. For example, a study in Minnesota found that two districts in the state had average probation terms for felony drug offenses of three years and nine years, respectively, but such differences are not evident in state averages.¹²

In addition, although the BJS figures do not provide information on probation lengths by sex or race, other evidence indicates that people of color in the probation system face multiple disadvantages that could affect length of stay, such as at sentencing and in sanctioning of violations. For instance, research conducted in several jurisdictions showed that African Americans were more likely than Whites to have their supervision revoked— even when accounting for factors known to increase a person’s risk for re-offending—which could result in shorter probation terms. On the other hand, another study reports that African Americans and Latinos received longer probation sentences than Whites with similar criminal history profiles, possibly yielding longer periods of supervision for people of color.¹³ Given these potentially offsetting outcomes, more data is needed to accurately gauge the nature and extent of disparities in probation lengths among different groups.

Finally, the BJS data also does not allow for comparisons of probation lengths by offense type or reveal when probation lengths are not aligned with the severity of the underlying crimes.¹⁴ For example, in Minnesota, the average probation sentence for a felony in 2014 and 2015 was 66 months, or about 5 1/2 years. But when broken down by offense type, the average terms were 4 1/2 years for violent offenses and 6 1/2 years for drug offenses.¹⁵ Although information is not currently available nationally to analyze racial disparities in supervision terms or the factors that drive probation length, states and localities could improve their understanding of these issues in their jurisdictions by gathering and analyzing their own data.

Probation length and recidivism

Research has found that longer terms of probation are not correlated with lower rates of re-offending and are more likely than shorter terms to result in technical violations, which studies have shown are a key driver of state incarceration rates and costs.¹⁶ To better understand the relationship between length of supervision and recidivism, Pew engaged an independent research firm, Maxarth LLC, to examine the relationship between longer probation terms and re-offending.

For this analysis, Maxarth researchers looked at data for people placed on probation between July 2009 and June 2013 in South Carolina (18,293 individuals) and between January 2010 and December 2014 in Oregon (41,224 individuals). Maxarth used these states because they were able to provide high quality data linking probation, arrests, and criminal history across several years. Further, data from the first year of observation shows that average probation term lengths in both states (24 months in Oregon and 26.2 months in South Carolina) were close to the 2018 national average of 22.4 months. (See the appendix for more information.) Because most re-offending occurs early in supervision, the researchers confined the study to people who had completed a year of their probation terms without being arrested—and so would be good candidates for shorter terms—and investigated whether these individuals could have served less time on supervision without jeopardizing public safety.

The Maxarth researchers created a statistical profile of each person using data points such as criminal history and age that are known to be linked to re-offending. The information that Maxarth examined included individuals’ criminal histories before supervision and for at least three years after starting probation, as well as demographics, risk assessment scores, and discharge information.

The researchers matched people who had been discharged from probation with individuals with similar profiles who were still on probation and used those matches to compare recidivism outcomes, as measured by re-arrests, at 12, 18, 24, and 30 months from the start of probation. They then used that information to determine how many people could have served a shorter term of probation and how much less time they could have spent under supervision.

Most people who were arrest-free for their first year on probation could have safely spent less time on supervision

The analysis found that in both states, 9 in 10 people who completed their first year on probation without being arrested could have spent at least three fewer months under supervision than they actually did with no negative impact on recidivism, and many could have served much less time.¹⁷ More than half (53%) of people on probation in the study in Oregon and about 3 in 5 (61%) of those in South Carolina could have served terms that were shorter by a year or more. (See Figure 3.) And these results were not confined only to people convicted of low-level offenses.

Had all eligible individuals served these reduced terms, both states would have had significantly shorter average probation terms over the study period without increasing rates of re-arrest. South Carolina’s average length of supervision would have dropped by almost a third (31%), and Oregon’s would have fallen by about 42%. (See Figure 4.)

Shortening time on supervision would have also reduced states’ probation ADP. South Carolina’s ADP during the study period was approximately 14,719 people, but this analysis suggests that figure could have been reduced by about 32% to 10,071 without increasing recidivism. Similarly, the ADP in Oregon was about 24,323 but could have been reduced to 13,697, approximately a 44% decrease. (See Figure 5.)

Although the people in South Carolina and Oregon whose probation terms could have been shortened by three years or more make up small shares of the total number who could have served reduced terms (20% and 11%, respectively), they account for the largest portion of the potential drop in ADP in both states: almost half (44%) of Oregon’s decrease, and about a third (33%) of the decline in South Carolina. (See Figure 6.)

States should analyze data to determine potential probation reductions and policies

This analysis suggests that many people who are arrest-free for their first year on probation could have served shorter terms with no increase in recidivism, reducing states’ probation populations. However, a deeper examination of which characteristics make people good candidates for shorter terms (offense type, for example) was beyond the scope of this research. By gathering and analyzing their own data, states can assemble information on their probation populations that they can then use to help identify policy and practice changes that could safely shorten term lengths.

State probation statutes

State laws significantly influence probation duration. Sentencing statutes outline term lengths for different offenses. Additionally, laws govern the circumstances under which terms can be extended and any mechanisms, such as earned time credits and mandatory reviews, that allow for early discharge. To better understand the statutory landscape across the country and how state laws affect probation length, Pew collaborated with the NCSL to compile the first 50-state analysis of maximum felony and misdemeanor probation terms,¹⁸ as well as statutory mechanisms for early discharge from probation. (See Table 2.)

Notes: “Max” means the probation term cannot exceed the statutory maximum sentence of incarceration for the same crime, and “none” means that state law prescribes no maximum, leaving it at the judge’s discretion. Statutes are current through 2017 legislative sessions and were collected and verified before subsequent legislative sessions. Indiana’s maximum misdemeanor probation term can be up to two years if the court finds that the use or abuse of alcohol or drugs contributed to the offense, and Minnesota’s six-year misdemeanor maximum applies to a few offenses. In Wisconsin, the maximum sentence for a misdemeanor can be increased if an individual is convicted of multiple misdemeanors.

Source: National Conference of State Legislatures, 50 State Review of 2017 Statutes

© 2020 The Pew Charitable Trusts

Maximum felony probation lengths

At the time of NCSL’s statutory review, 43 states had statutory maximums for felony probation terms, and the remaining seven states leave the length of probation to the discretion of the judge. (See Figure 7.) Nineteen states had a five-year maximum felony term, and in 13 the felony probation could not exceed the statutory maximum sentence of incarceration for the same crime. Alaska, Hawaii, and Texas had the longest felony probation maximum term at 10 years. The remaining eight states set their maximum terms at two, three, four, or seven years.

Maximum misdemeanor probation lengths

Forty-five states had statutory maximums for misdemeanor probation. (See Figure 8.) Fifteen had a two-year maximum;¹⁹ nine states had a five-year maximum, and in another 10, the length could not exceed the maximum statutory sentence of incarceration for the same offense. The remaining 11 states had one-, three-, six-, and seven-year maximums for misdemeanors.

Extensions of probation can exceed statutory maximum sentences, driving up term lengths

In many states, probation sentences can be extended or reinstated, sometimes even beyond statutory maximums. Typically, this occurs because a person fails to follow the rules of supervision, including payment of probation fees.²⁰ In Utah, for example, the court can order the entire probation term to start over because of a violation.²¹ And a study in Arizona found that more than half of people on probation for drug offenses had their terms reinstated at least once, mostly for technical violations.²²

Some state statutes explicitly prescribe that the maximum term includes the original probation sentence and any extensions. However, others allow extensions past the maximum, sometimes up to the statutory sentence of incarceration for the same offense (which can be longer than the maximum probation term); one such state is Oklahoma, where the maximum felony probation term is two years, but the average length of time on probation is three because terms can be extended up to the maximum statutory sentence for the underlying offense.²³ In some states, probation can be extended indefinitely under certain circumstances, which commonly include failure to pay restitution or child support.²⁴

Early discharge offers opportunity to shorten probation terms

NCSL also reviewed statutes relating to early termination of probation. Early discharge, sometimes referred to as early termination or early release, is an incentive offered to people to promote compliance with the rules and provisions of supervision that supports and encourages behavior change while reducing caseloads so probation agencies can concentrate their resources on individuals at higher risk of re-offending. This analysis included statutes that provide for the court’s authority to discharge people from probation early, petitions or recommendations for early termination, earned time credits, and automatic or mandated review for early discharge. (See Figure 9.)

Court authority to terminate probation early

Most states have statutes related to the court’s authority to modify probation terms, whether by terminating, extending, or reducing the amount of time or by changing the conditions. Forty states have statutes, many of which are broadly written, that explicitly allow the court to end probation early. For example, Arkansas law indicates that the “court may discharge defendant from probation at any time.”²⁵ On the other hand, Connecticut statute states the court can terminate someone’s probation “for good cause shown”—usually meaning the person’s conduct—while in Vermont the court must deem that an early discharge serves “the ends of justice.”²⁶ For other states, the conditions under which a person’s probation can be ended are more specific. In Texas, for instance, a judge can terminate probation only after the individual “has satisfactorily completed one-third of the original community supervision period or two years of community supervision, whichever is less.”²⁷

In Montana, reviews are scheduled based on an individual’s risk and needs.²⁸ A conditional discharge review is conducted at nine months for people at low risk to re-offend, 12 months for medium risk, 18 months for moderate risk, and 24 months for high risk.

Petition or recommendation for early discharge

Twenty-two states have statutes that allow a designated entity to submit a petition or recommendation to the court to have an individual’s probation terminated early, but who can initiate this action varies by state. In 21 states, the department of corrections or the probation department, supervisor, or officer can petition the court for early termination, though some of these states require that the individual first complete a specified amount of time on probation. For example, in Michigan, people convicted of felonies must serve half of their probationary periods before they can be recommended for early termination.²⁹ In addition, some states, including New Mexico, New York, and Ohio, allow local governments to adopt their own policies allowing probation officers to recommend early termination.

Further, in 10 states, people on probation can petition for their own early termination. In Colorado this applies only to people convicted of impaired driving, whereas in Idaho anyone on probation can request termination at any time.³⁰ New York and Ohio also permit local governments to enact policies allowing individuals on probation to petition for early termination. In Rhode Island, the probation department can give certificates to people who have been on probation for three years, had no technical violations, and completed all obligations, including paying restitution and fines, and those individuals can then present the certificate to the superior court for consideration of early termination.³¹ It is not clear how often these statutory early discharge mechanisms are used.

Mandated review for early discharge

Ten states have statutes requiring corrections authorities to periodically review individuals’ eligibility and readiness for early discharge for probation. In several states—Alabama, Alaska, Connecticut, Kansas (if low-risk), Kentucky, and Texas (after half of the sentence or two years of supervision, whichever is more)—these evaluations generally occur after one to two years on probation, but state provisions vary and some offenses are excluded.³² In North Carolina, individuals sentenced to more than three years of probation are subject to a mandatory review at the three-year point.³³

Earned time credits

Earned time credits, sometimes referred to as earned credits or earned compliance credits, are a process by which people secure time off their sentences through good behavior, including general compliance or completion of programming. Research indicates that offering people the opportunity to reduce their sentences via earned time credits encourages compliance and increases successful outcomes without compromising public safety, especially if the incentive is perceived as significant, such as receiving 15 to 30 days off a sentence for each month of compliance.³⁴

Sixteen states have statutes that allow for earned time credits on probation, but they vary considerably with respect to the amount of time by which the sentence can be reduced, eligible offenses, and requirements. For example, in South Dakota, only individuals who have been convicted of a felony and sentenced to at least six months of probation are eligible for earned time credits, while other states allow people sentenced for misdemeanors to earn credits.³⁵

A handful of states, including Alaska, Arkansas, Delaware, Missouri, and Utah, have “30 for 30” policies, in which the person on probation gets 30 days of credit for 30 days of compliance with probation conditions.³⁶ Arizona, Maryland, and South Carolina have “20 for 30” policies, and Nevada gives 10 days of credit for 30 days of compliance.³⁷

Many statutes are limited in scope. In Illinois, an individual can receive earned time credits only by obtaining increasing levels of educational degrees, and in Indiana earned time is available only to people on home confinement as a condition of probation.³⁸ Although Maryland does have an earned time policy, the state’s practice typically has been not to end probation but instead put people on “abatement,” meaning they are not required to regularly report to a probation officer.³⁹ However, violations can result in a return to active supervision.

Statutes do not always predict probation lengths

The mere presence or absence of statutes related to probation length is not necessarily a good predictor of actual terms of probation because of variation in how statutes are crafted and implemented across jurisdictions and the discretion often written into laws. For example, a comparison of the 10 states with the longest probation lengths and the 10 with the shortest finds that half of the states in both groups have statutes that provide for earned time or early discharge via judicial review, most have felony maximum terms, and almost all provide the court with the authority to terminate probation.

This finding suggests that reducing statutory maximums and allowing early termination alone may not reduce probation terms, particularly if state law has exceptions to maximum terms that keep many people on probation longer. For example, a study in Maricopa County, Arizona, found that although the state’s credit policy enabled certain individuals on probation to earn credit at a rate of 20 days for every 30 days in compliance, the requirement that people stay current with financial obligations, including court-ordered fees, restitution, fines, and surcharges, was preventing some people from being discharged.⁴⁰

On the other hand, an evaluation of Missouri’s 30 for 30 earned time policy found that it effectively reduced the state’s supervision population and officer caseloads, cutting supervision terms by 14 months on average without negatively affecting recidivism.⁴¹

Discretion

A key factor affecting the ways in which states implement probation length statutes is individual choice in decision-making. In some cases, discretion on the part of judges, prosecutors, and probation agencies can help shorten terms while protecting public safety by allowing decision-makers to consider each individual’s circumstances and behavior while on probation. For example, in 2010 the New York City Department of Probation began actively recommending early termination for people on probation who met eligibility requirements, and as a result, the department went from discharging 3% of people through early discharge in 2007 to 17% in 2012.⁴² Under this change, individuals discharged early had a lower felony re-arrest rate than those who stayed on probation for their full terms.⁴³

However, discretion also leaves room for disparities between jurisdictions or even courtrooms.⁴⁴ Judicial or regional discretion can affect probation terms at various points, including initial sentencing, when a judge or probation officer sets the requirements for a person on supervision, when decisions are made about how to address supervision violations, and when considering or recommending early termination.

Recommendations

Policymakers can take important steps toward improving supervision efficiency and outcomes by adopting innovative and proven solutions and sound administration to address terms of probation, including:

• Goal-based supervision, which prioritizes outcomes as opposed to time-based supervision models that focus on enforcing rules over a set period. Under goal-based systems, the sooner people achieve their goals, the sooner supervision ends, encouraging positive behavior change.
• Earned compliance credits, which promote positive behavior, encourage compliance, increase successful supervision outcomes, and reduce caseloads by allowing people who follow the rules and complete needed programs to earn early discharge. Effective earned compliance programs establish and communicate clear expectations and conditions to the individual on supervision from the outset.
• Automatic review of supervision, which ensures that states use specific guidelines and definable behavioral goals to determine eligibility for early termination and should also involve documentation and monitoring of termination decisions to ensure fairness.⁴⁵

These policies serve a dual mission, incentivizing law-abiding behavior and concentrating supervision on individuals who fail to comply with court- or agency-ordered rules. Jurisdictions seeking to increase successful completion of supervision; decrease revocations, particularly those involving incarceration; and safely reduce the number of people on probation are gradually moving in this direction, but much more progress is within reach.

Conclusion

One in 72 adults in the U.S. are on probation. Because of the size of this population, state and local agencies tasked with administering community supervision often struggle to provide alternatives to incarceration that are effective in reducing re-offending. This analysis indicates that a key driver of this large population is long probation terms. The findings show that many people are kept on probation for months or years, even when doing so no longer delivers public safety benefits, stretches caseloads, strains resources, and needlessly burdens people on supervision with conditions and fees.

To be effective, probation systems should prioritize resources for the period during which a person is most at risk to re-offend, typically the first 12 to 18 months; employ policies and practices shown to reduce re-offending; and terminate supervision when additional probation will deliver no further public safety benefit. This approach can enable agencies to promote community safety, make the best use of limited public resources, and improve long-term outcomes for people on supervision.

Appendix A: Supplemental data

Table A.1

Percentage Change in Probation Population, Average Term Length, and Entries by State, 2000 and 2018

Note: Data is for 2000 and 2018 and does not consider impacts of fluctuations in admissions and length of stay in intervening years.

Sources: Bureau of Justice Statistics, “Annual Probation Survey and Annual Parole Survey” (2000-2018)

© 2020 The Pew Charitable Trusts

Appendix B: Methodology

Much of the data in this report comes from the federal Bureau of Justice Statistics’ Annual Probation and Annual Parole surveys, which are the only community corrections dataset covering all 50 states, the District of Columbia, and the federal system. The 2018 survey was sent to 454 agencies, including 40 “central” state and Washington, D.C., agencies; 415 “separate” state, county, or court agencies; and the federal system.⁴⁶ Federal and District data was not used in this analysis.

Some data issues do exist, including changes in reporting methods from year to year and state to state; differences in the handling of people with multiple concurrent probation sentences; varied approaches to counting people who enter and exit probation more than once a year; missing data; and inconsistencies across states regarding inclusion of supervision programs such as drug courts, administrative caseloads, or private supervision in their probation numbers.

The probation-lengths analysis calculated the mean length of time on probation as the inverse of the exit rate. Because of the small growth rate in probation in recent years, this method yields sufficient estimates.⁴⁷ BJS indicates that, given a relatively “stationary” population and small growth rate, the inverse of the exit rate performs well compared with other methods.⁴⁸

Oregon and South Carolina propensity score matching

The states’ probation population compositions differed somewhat: Oregon had a higher proportion of people on probation for felony charges; South Carolina had more people successfully exit; and, because the states use different risk assessment instruments, the distributions and meaning of assessed risk levels (low, moderate, high) are inconsistent.

Table B.1

Probation Lengths and Early Termination Study Background Data

Note: Data is for 2000 and 2018 and does not consider impacts of fluctuations in admissions and length of stay in intervening years.

Source: Analysis by Maxarth LLC of data from: the South Carolina Department of Probation, Parole and Pardon Services, 2009-13; South Carolina State Law Enforcement Division; Oregon Department of Corrections, 2010-14; and Oregon Criminal Justice Commission

© 2020 The Pew Charitable Trusts

The propensity score models included attributes related to recidivism as measured by re-arrest so that the groups were balanced on relevant features (sex, race, current charge category, current charge class, risk level, criminal history, and sentence imposed on controlling charge). A propensity score (the output from a logistic regression model) was computed for each individual. Then, the closest matching sample member from the opposite group was matched to each member in the treatment group; that is, the group that is still being supervised.

Because those serving longer probation terms constituted the treatment group, the propensity scores were used to find individuals who have served shorter probation terms but were otherwise similar in every other respect. The individuals in the control group—those who were successfully discharged—provided the potential outcome, i.e., what could reasonably be expected had a matched individual served a shorter length of time on probation. Comparing the actual and the potential outcome for the treatment group then allowed Maxarth to identify the number of people currently on probation who had the same or fewer arrests (over a fixed follow-up window) as similar individuals who were already discharged. This subsample of people currently on probation was identified as the number who could have safely been discharged earlier.

The control group therefore was defined as anyone who had been successfully discharged from probation by the cut-point month, while the treatment group was defined as anyone not released from probation by then (i.e., was still being supervised beyond that month). For example, a 12-month cut-point would mean the treatment group consisted of those still on probation 12 months after intake, and the control group comprised those who have successfully been discharged by month 12 after intake. Similar treatment and control indicators were defined for the 18-, 24-, and 30-month cut-points.

The data used to measure public safety and construct potential outcomes for the treatment group was defined as the number of arrests from the cut-point to the end of 36 months after intake. As a result, the same number of follow-up months was used for the treatment and control groups around the same cut-point. However, although the actual and potential outcomes for the matched groups were consistent, outcomes cannot be compared across cut-points because the follow-up periods were different. For example, the number of re-arrests after the 12-month cut-point included all arrest events recorded between months 13 and 36. But for the 24-month cut-point, the outcome included only arrests between months 25 and 36, and for the 30-month cut-point, it included only the last six months of the follow-up period (months 31 to 36).

Estimated propensity scores were used to develop a matched comparison sample for each treatment sample. The matched sample was the one with the closest propensity score—the probability of being in the treatment group—to the current treatment sample. In addition, matching was done with replacement—i.e., the same control sample could be matched to multiple treatment observations.

The subsample of individuals who were still on probation and were potential candidates for early discharge were those for whom additional probation either made no difference (i.e., the outcome under continued supervision was the same as that of the matched control sample) or made things worse (i.e., the outcome under further supervision was worse than if the person had been discharged earlier).

The actual early discharge date is harder to infer from this analysis and probably depends on additional state-specific factors. For purposes of conducting simulations, however, the cut-point was used as the early discharge date for all eligible individuals. Moreover, the earliest discharge date was selected if an individual qualified for more than one such date. For example, if an individual qualified for early discharge using both the 24-month and 18-month cut-points, then the early discharge date was set at the end of 18 months post-intake.

Endnotes

1. D. Kaeble, “Probation and Parole in the United States, 2017-18” (Bureau of Justice Statistics, 2020), https://www.bjs.gov/content/pub/pdf/ppus1718.pdf. The total population on supervision in the U.S. at the end of 2018 was approximately 4.4 million, with about 900,000 people on parole or other forms of post-release supervision.
2. The Pew Charitable Trusts, “Probation and Parole Systems Marked by High Stakes, Missed Opportunities: 1 in 55 Adults Is Under Community Supervision” (2018), https://www.pewtrusts.org/en/research-and-analysis/issue-briefs/2018/09/probation-and-parole-systems-marked-by-high-stakes-missed-opportunities.
3. M.S. Phelps, “Mass Probation: Toward a More Robust Theory of State Variation in Punishment,” Punishment & Society 19, no. 1 (2017): 53-73, https://journals.sagepub.com/doi/abs/10.1177/1462474516649174.
4. M.W. Lipsey, “Rehabilitation Programs for Adult Offenders: A Meta-Analysis in Support of Guidelines for Effective Practice” (National Criminal Justice Reference Service, 2019), https://www.ncjrs.gov/pdffiles1/nij/grants/252504.pdf.
5. J. Austin, “Reducing America’s Correctional Populations: A Strategic Plan,” Justice Research and Policy 12, no. 1 (2010): 9-40, https://journals.sagepub.com/doi/abs/10.3818/JRP.12.1.2010.9; National Research Council, “Parole, Desistance From Crime, and Community Integration” (2008), https://www.nap.edu/catalog/11988/parole-desistance-from-crime-and-community-integration; R. Rosenfeld, J. Wallman, and R. Fornango, “The Contributions of Ex-Prisoners to Crime Rates,” in Prisoner Reentry and Crime in America, eds. J. Travis and C. Visher (New York, NY: Cambridge University Press, 2005); Statistical Analysis Center, “Recidivism Dashboard,” State of Oregon Criminal Justice Commission, accessed Oct. 1, 2019, https://www.oregon.gov/CJC/SAC/Pages/Recidivism-dashboard.aspx.
6. The Pew Charitable Trusts, “Policy Reforms Can Strengthen Community Supervision: A Framework to Improve Probation and Parole” (2020), https://www.pewtrusts.org/en/research-and-analysis/reports/2020/04/policy-reforms-can-strengthen-community-supervision.
7. As with other national criminal justice data, reporting may vary from state to state in data quality and data included in the numbers provided to BJS. For example, in some states the entire misdemeanor probation population, those in drug and other specialty courts, or those on private probation may not be included in reported probation statistics.
8. E. Patterson and S. Preston, “Estimating Mean Length of Stay in Prison: Methods and Applications,” Journal of Quantitative Criminology 24, no. 1 (2008): 33-49, https://www.researchgate.net/publication/225646782_Estimating_Mean_Length_of_Stay_in_Prison_Methods_and_Applications.
9. This analysis excludes the federal and District of Columbia probation populations.
10. S.K. Jalbert and W. Rhodes, “Reduced Caseloads Improve Probation Outcomes,” Journal of Crime and Justice 35, no. 2 (2012), 221-238, https://doi.org/10.1080/0735648X.2012.679875.
11. Phelps, “Mass Probation.”
12. K.L. Mitchell, “It’s Time to Rethink Probation Lengths in Minnesota,” News and Views From Robina (blog), Robina Institute of Criminal Law and Criminal Justice, Jan. 13, 2017, https://robinainstitute.umn.edu/news-views/it%E2%80%99s-time-rethink-probation-lengths-minnesota.
13. J. Jannetta et al., “Examining Racial and Ethnic Disparities in Probation Revocation: Summary Findings and Implications From a Multisite Study” (The Urban Institute, 2014), https://www.urban.org/sites/default/files/publication/22746/413174-Examining-Racial-and-Ethnic-Disparities-in-Probation-Revocation.PDF; E. Owens, E.M. Kerrison, and B.S. Da Silveira, “Examining Racial Disparities in Criminal Case Outcomes Among Indigent Defendants in San Francisco” (2017), https://www.law.upenn.edu/live/files/6792-examining-racial-disparities-may-2017-summary.
14. Mitchell, “Rethink Probation Lengths.”
15. Ibid.
16. K.D. Morgan, “Factors Associated With Probation Outcome,” Journal of Criminal Justice 22, no. 4 (1994): 341-53, https://www.sciencedirect.com/science/article/pii/0047235294900817; D.E. Olson and A.J. Lurigio, “Predicting Probation Outcomes: Factors Associated With Probation Rearrest, Revocations, and Technical Violations During Supervision,” Justice Research and Policy 2, no. 1 (2000): 73-86, https://journals.sagepub.com/doi/abs/10.3818/JRP.2.1.2000.73; B. Sims and M. Jones, “Predicting Success or Failure on Probation: Factors Associated With Felony Probation Outcomes,” Crime & Delinquency 43, no. 3 (1997): 314-27, https://journals.sagepub.com/doi/abs/10.1177/0011128797043003005; The Council of State Governments, “Confined and Costly: How Supervision Violations Are Filling Prisons and Burdening Budgets” (2019), https://csgjusticecenter.org/publications/confined-costly/.
17. This analysis did not explore potential probation term reductions for people with arrests in the first year of supervision, but at least some individuals in this group likely could also have had their terms reduced.
18. In some states, statutes concerning maximum probation terms exclude certain offenses, such as sex offenses, effectively allowing much longer maximums, including life probation terms. For example, in New York certain drug felonies have a lifetime probation maximum, and terms for sex offenses that were excluded from the statutory review ranged from 10 years to life. Statutes are current as of the end of the 2017 legislative sessions.
19. Wisconsin Stat. 973.09(2)(a)2, https://docs.legis.wisconsin.gov/statutes/statutes/973/09/2/a. The two-year maximum sentence can be increased if an individual is convicted of multiple offenses.
20. And, in some places, individuals can have their supervision revoked and be sent to jail or prison for nonpayment of fees. See K.L. Mitchell et al., “Profiles in Probation Revocation: Examining the Legal Framework in 21 States” (Robina Institute of Criminal Law and Criminal Justice, 2014), https://robinainstitute.umn.edu/publications/profiles-probation-revocation-examining-legal-framework-21-states.
21. Utah Code § 77-18-1(12)(E)(Ii), https://le.utah.gov/xcode/Title77/Chapter18/77-18-S1.html; J. Bronson and E.A. Carson, “Prisoners in 2017” (2019), https://www.bjs.gov/content/pub/pdf/p17.pdf.
22. N. Rodriguez and V.J. Webb, “Probation Violations, Revocations, and Imprisonment: The Decisions of Probation Officers, Prosecutors, and Judges Pre- and Post-Mandatory Drug Treatment,” Criminal Justice Policy Review 18, no. 1 (2007): 3-30, https://journals.sagepub.com/doi/abs/10.1177/0887403406292956.
23. Oklahoma Stat. Tit. § 22-991a(E), http://oklegal.onenet.net/oklegal-cgi/get_statute?99/Title.22/22-991A.html.
24. No maximum for outstanding restitution: Kentucky Rev. Stat. § 533.020(4), https://apps.legislature.ky.gov/law/statutes/statute.aspx?id=20076; Utah Code § 77-18-1(10)(B) (2019), https://le.utah.gov/xcode/Title77/Chapter18/77-18-S1.html?v=C77-18-S1_2019051420190514; Wisconsin Stat. § 973.09(3), https://docs.legis.wisconsin.gov/statutes/statutes/973/09. No maximum sentence for offenses related to failure to pay child support: Mississippi Code § 47-7-37; North Dakota Cent. Code § 12.1-32-06.1-5, https://www.legis.nd.gov/cencode/t12-1c32.pdf.
25. Arkansas Code Ann. § 16-93-314, https://law.justia.com/codes/arkansas/2017/title-16/subtitle-6/chapter-93/subchapter-3/section-16-93-314/.
26. Connecticut Gen. Stat. § 53a-33, https://www.cga.ct.gov/current/pub/chap_952.htm#sec_53a-33; Vermont Stat. Ann. Tit. 28 § 251, https://legislature.vermont.gov/statutes/section/28/005/00251.
27. Texas CCP, Art. 42a.701 (a), https://statutes.capitol.texas.gov/Docs/CR/htm/CR.42A.htm.
28. Montana Code § 46-23-1011 (6), https://leg.mt.gov/bills/mca/title_0460/chapter_0230/part_0100/section_0110/0460-0230-0100-0110.html; Rodriguez and Webb, “Probation Violations, Revocations, and Imprisonment.”
29. Michigan Comp. Laws § 771.2, http://www.legislature.mi.gov/(S(u40cehaecp4srr0apj1a5rtz))/mileg.aspx?page=GetObject&objectname=mcl-771-2.
30. Colorado Rev. Stat. 42-4-1307(7), https://codes.findlaw.com/co/title-42-vehicles-and-traffic/co-rev-st-sect-42-4-1307.html; United States Courts, “Post-Conviction Supervision—Judicial Business 2018,” accessed June 26, 2019, https://www.uscourts.gov/statistics-reports/post-conviction-supervision-judicial-business-2018; T.C. Pratt, “A Self-Control/Life-Course Theory of Criminal Behavior,” European Journal of Criminology 13, no. 1 (2015): 129-46, https://doi.org/10.1177/1477370815587771.
31. Rhode Island Superior Court Rules, Early Termination of Probation and Parole, 35©, http://www.doc.ri.gov/community-corrections/probation-parole/early-termination.php.
32. Alabama Code § 15-22-54(B), http://alisondb.legislature.state.al.us/alison/CodeOfAlabama/1975/15-22-54.htm; Alaska Stat. 12.55.090(G) (2019), http://www.akleg.gov/basis/statutes.asp#12.55.090; Connecticut Gen. Stat. § 53a-29(G), https://www.cga.ct.gov/current/pub/chap_952.htm#sec_53a-29; Kansas Stat. § 21-6608 (D), https://www.ksrevisor.org/statutes/chapters/ch21/021_066_0008.html; Kentucky Rev. Stat. § 439.552, https://apps.legislature.ky.gov/law/statutes/statute.aspx?id=39574; Texas CCP, Art. 42a.701 (B), https://statutes.capitol.texas.gov/Docs/CR/htm/CR.42A.htm.
33. North Carolina Gen. Stat. § 15a-1342 (D), https://www.ncleg.net/EnactedLegislation/Statutes/HTML/BySection/Chapter_15A/GS_15A-1342.html.
34. The Pew Charitable Trusts, “Missouri Policy Shortens Probation and Parole Terms, Protects Public Safety” (2016), https://www.pewtrusts.org/en/research-and-analysis/issue-briefs/2016/08/missouri-policy-shortens-probation-and-parole-terms-protects-public-safety; E.J. Wodahl, B.E. Garland, and T.J. Mowen, “Understanding the Perceived Value of Incentives in Community Supervision,” Corrections 2, no. 3 (2017): 165-88, https://doi.org/10.1080/23774657.2017.1291314.
35. South Dakota Codified Laws § 23a-48-15 Et Seq., https://sdlegislature.gov/Statutes/Codified_Laws/DisplayStatute.aspx?Type=Statute&Statute=23A-48-16.
36. Alaska Stat. 33.05.020 (2019), http://www.akleg.gov/basis/statutes.asp#33.05.020; Arkansas Code Ann. § 16-90-1303, https://law.justia.com/codes/arkansas/2012/title-16/subtitle-6/chapter-90/subchapter-13/section-16-90-1303; Delaware Code Tit. 11,§ 4383, http://delcode.delaware.gov/title11/c043/sc08/index.shtml; Utah Code § 64-13-21 (7), https://le.utah.gov/xcode/Title64/Chapter13/64-13-S21.html?v=C64-13-S21_2019051420190514. In 2018, Utah passed House Bill 291, ending the existing earned compliance credit program and requiring the Utah Sentencing Commission to draft supervision length guidelines for individuals on probation and parole. These guidelines are meant to increase equity in criminal supervision lengths, respond to public comment, relate the length of supervision to an individual’s progress, take into account an individual’s risk of offending again, relate the length of supervision to the amount of time a person has been under supervision, and enhance the discretion of sentencing judges while preserving the role of the Board of Pardons and Parole.
37. Arizona Rev. Stat. § 13-924 (2013), https://www.azleg.gov/viewdocument/?docName=https://www.azleg.gov/ars/13/00924.htm; Maryland Code, Correctional Services, § 6-117 (2013), https://govt.westlaw.com/mdc/Document/NB63B5A2064B511E690DBA4A79E125EF7?transitionType=Default&contextData=%28sc.Default%29; South Carolina Code Ann. § 24-21-280 (D), https://www.scstatehouse.gov/code/t24c021.php; Nevada Rev. Stat. § 176a.500 (5), https://www.leg.state.nv.us/NRS/NRS-176A.html#NRS176ASec500.
38. Illinois Comp. Stat. 730, 5/5-6-2 (2013), http://www.ilga.gov/legislation/ilcs/fulltext.asp?DocName=073000050K5-6-2; Delaware Code Tit. 11, § 4383. Indiana Code 35-38-2.5-5, https://www.in.gov/idoc/files/IC_35-38-2.5___Home_Detention.pdf. Depending on the offense, in Indiana, an individual could also earn good time credit for time confined while awaiting trial or sentencing or for time served on pretrial home detention (Ind. Code § 35-50-6-3.1.).
39. Maryland Code, Correctional Services, § 6-117.
40. Arizona Rev. Stat. § 13-924; M.L. Griffin, J.R. Hepburn, and K. Ginsburg, “The Effects of Earned Time Credit on Successful Probation Outcomes: An Evaluation of Maricopa County, Arizona’s Experience” (2013).
41. The Pew Charitable Trusts, “Missouri Policy Shortens Probation and Parole Terms.”
42. M.P. Jacobson et al., “Less Is More: How Reducing Probation Populations Can Improve Outcomes” (Harvard Kennedy School, (2017), https://www.hks.harvard.edu/centers/wiener/programs/criminaljustice/research-publications/executive-sessions/executive-session-on-community-corrections/publications/less-is-more-how-reducing-probation-populations-can-improve-outcomes.
43. Ibid.
44. M. Jones and J.J. Kerbs, “Probation and Parole Officers and Discretionary Decision-Making: Response to Technical and Criminal Violations,” Federal Probation 71, no. 1 (2007), https://www.uscourts.gov/federal-probation-journal/2007/06/probation-and-paroleofficers-and-discretionary-decision-making.
45. S. Torres, “Early Termination: Outdated Concept in an Era of Punitiveness,” Federal Probation 63, no. 1 (1999): 35-41, https://www.uscourts.gov/federal-probation-journal/1999/06/early-termination-outdated-concept-era-punitiveness; J.J. Norris, “The Earned Release Revolution: Early Assessments and State-Level Strategies,” Marquette Law Review 95 (2011): 1551, https://scholarship.law.marquette.edu/mulr/vol95/iss4/15/.
46. D. Kaeble, “Probation and Parole in the United States, 2017-2018” (Bureau of Justice Statistics, 2020), https://www.bjs.gov/content/pub/pdf/ppus1718.pdf.
47. Patterson and Preston, “Estimating Mean Length of Stay”; E.J. Herberman and T.P. Bonczar, “Probation and Parole in the United States, 2013” (Bureau of Justice Statistics, 2015), https://www.bjs.gov/content/pub/pdf/ppus13.pdf.
48. E.J. Herberman and T.P. Bonczar, “Probation and Parole in the United States, 2013” (Bureau of Justice Statistics, 2015). “Stationary” indicates roughly consistent numbers of admissions and releases, leading to a mainly stable probation population.