Recidivism is the pattern of reoffending after a person has already been through the criminal justice system, and in the United States it is strikingly common: Bureau of Justice Statistics data shows about 66% of released state prisoners are rearrested within three years and roughly 82% within ten.1Bureau of Justice Statistics. Recidivism of Prisoners Released in 24 States in 2008: A 10-Year Follow-Up Period (2008-2018) Those numbers shape sentencing law, prison programming, and parole policy, and they explain why a second or third offense is treated so differently from a first.
What Counts as Recidivism
In everyday usage, recidivism means a person punished for a crime goes on to commit another one. The legal system is more specific. Depending on the jurisdiction, someone may be classified as a recidivist when they are rearrested, reconvicted, or returned to a correctional facility. Some jurisdictions require a formal conviction before applying the label. Others count a new arrest as sufficient for statistical purposes.
A central mechanism in recidivism law is the look-back period, which defines the window of time used to evaluate someone’s criminal history. Habitual offender statutes typically ask whether a person has committed multiple felonies within a defined span of years. If they have, the sentencing consequences for a new offense escalate sharply.
At the federal level, the three-strikes provision at 18 U.S.C. § 3559(c) mandates life imprisonment for anyone convicted of a “serious violent felony” who has two or more prior convictions for serious violent felonies or serious drug offenses.2Office of the Law Revision Counsel. 18 U.S.C. 3559 – Sentencing Classification of Offenses The statute defines serious violent felonies broadly, including murder, robbery, kidnapping, arson, carjacking, aggravated sexual abuse, and any offense carrying a maximum sentence of ten or more years that involves the use or threatened use of physical force. Many states run parallel habitual offender laws with their own thresholds and penalties.
How Common Reoffending Actually Is
Researchers measure recidivism by following a release cohort, a group of people who left prison or started community supervision around the same time, and counting how many are rearrested, reconvicted, or returned to custody within a set number of years. The Bureau of Justice Statistics has used five-year, nine-year, and ten-year follow-up windows in its major studies.3Bureau of Justice Statistics. Recidivism Patterns Explorer
Among prisoners released across 24 states in 2008, about 66% were arrested within three years and 82% within ten. Roughly 61% returned to prison within the decade, either for a new sentence or a supervision violation.1Bureau of Justice Statistics. Recidivism of Prisoners Released in 24 States in 2008: A 10-Year Follow-Up Period (2008-2018) An earlier study of prisoners released in 2005 found that 44% were arrested within the very first year, and 82% of all arrests during a nine-year follow-up occurred within the first three years after release.4Bureau of Justice Statistics. 2018 Update on Prisoner Recidivism: A 9-Year Follow-up Period (2005-2014) The concentration of risk in the first year or two is why reentry support during that window matters so much.
Rates vary significantly by original offense. Over ten years, people originally incarcerated for property crimes had the highest rearrest rate at nearly 87%, followed by public order offenses at about 82%, drug offenses at roughly 81%, and violent offenses at approximately 77%.1Bureau of Justice Statistics. Recidivism of Prisoners Released in 24 States in 2008: A 10-Year Follow-Up Period (2008-2018) The fact that people convicted of violent offenses actually reoffend at a lower rate than those convicted of property crimes surprises most readers, but the pattern has been consistent across multiple BJS studies.
One caveat matters for interpreting any of these figures: they are rearrest numbers, not conviction numbers. Being arrested does not mean a person committed a new crime. Some arrests do not lead to charges, and some charges do not lead to convictions. Rearrest is used because it captures the broadest picture of justice-system contact, but the headline percentages overstate the share of people who are actually convicted of new offenses.
Technical Violations Versus New Offenses
Not all recidivism involves a new crime. The system distinguishes between two kinds of conduct that can send someone back to prison.
Technical violations happen when a person on supervised release breaks the rules of their supervision without committing a new offense. Failing a drug test, missing a meeting with a probation officer, or leaving a restricted geographic area are common examples. These violations can trigger a revocation hearing where a judge decides whether to return the person to custody.5United States Courts. Just the Facts: Revocations for Failure to Comply with Supervision Conditions and Sentencing Outcomes Revocation is not automatic. Judges can impose shorter sanctions like a brief jail stay, electronic monitoring, or extended supervision.
New criminal offenses are a separate category. A new charge starts its own legal process with its own potential sentence, and a new felony conviction on top of an existing record pushes the person deeper into habitual offender territory. A significant share of people counted as recidivists return to prison for technical violations rather than new criminal conduct, which is worth remembering when reading any national rearrest or return-to-custody figure.
How Prior Convictions Change a Sentence
Criminal history is one of the most important factors in determining a sentence. Federal law requires courts to consider “the history and characteristics of the defendant” and “the need to protect the public from further crimes” when imposing any sentence.6Office of the Law Revision Counsel. 18 U.S.C. 3553 – Imposition of a Sentence In practice, a person with prior convictions almost always faces a longer sentence than a first-time offender convicted of the same crime.
The federal mechanism is the presentence investigation report, prepared by a probation officer after a guilty plea or conviction. The report calculates a criminal history category under the U.S. Sentencing Guidelines based on prior convictions and identifies the applicable sentencing range.7United States Courts. Presentence Investigations A person in the highest criminal history category can face a guideline range two or three times longer than someone in the lowest category for the identical offense.
At the extreme end sits the federal three-strikes rule, which mandates life imprisonment for a third serious violent felony conviction.2Office of the Law Revision Counsel. 18 U.S.C. 3559 – Sentencing Classification of Offenses Most states have their own habitual offender statutes. Some require three qualifying felonies, others two. Penalties range from mandatory minimums to life sentences.
What Drives Reoffending
Recidivism is rarely a matter of pure individual choice. Researchers identify a set of factors called criminogenic needs that are directly tied to the likelihood of a new offense. When someone leaves prison without stable housing, without a job, and without a support network, the statistical odds of a new arrest are high regardless of intent.
Employment is one of the heaviest barriers. Many incarcerated people have limited work history and gaps that are hard to explain to employers. Low educational attainment compounds the problem. Housing instability makes it nearly impossible to hold a job, comply with supervision, or maintain the kind of daily routine that keeps people out of trouble. Substance use history adds further risk, particularly when treatment is unavailable after release.
Age at first offense is a reliable predictor. People who enter the system young tend to face more persistent challenges throughout adulthood, partly because incarceration disrupts the education and social development that normally happens during those years. Antisocial peer networks, family instability, and untreated mental health conditions add layers of risk, and when several of these factors overlap, as they frequently do, the probability of a new offense climbs sharply.
The legal consequences of a conviction compound these pressures. Federal law bars people with certain felony convictions from working at FDIC-insured banks, and many states impose their own licensing restrictions on fields like healthcare, education, and commercial driving. Federal law allows public housing authorities to deny housing based on criminal history. Federal student aid can be restricted for drug convictions, and people with felony drug convictions can face bans from SNAP and TANF benefits in many states. The Fair Chance to Compete for Jobs Act now prohibits federal agencies from asking about criminal history until after making a conditional job offer, with exceptions for security-clearance, law enforcement, and sensitive national security positions.8Federal Register. Fair Chance to Compete for Jobs Many states and cities have adopted similar “ban-the-box” laws for private employers, with wide variation in coverage. Stacked together, these barriers weight the structural deck toward reoffending.
What Actually Reduces Recidivism
The strongest evidence for reducing reoffending points to correctional education. A RAND Corporation meta-analysis of 30 years of research found that inmates who participated in education programs had 43% lower odds of returning to prison, a 13-percentage-point drop in the recidivism rate.9RAND Corporation. Evaluating the Effectiveness of Correctional Education GED programs, vocational training, and post-secondary education all showed benefits, and the study found the programs cost-effective because savings on reincarceration outweighed program costs.
Cognitive behavioral therapy has a more mixed record. Earlier meta-analyses suggested it could cut reoffending by about a quarter, from roughly 40% in control groups to around 30% among participants. A 2021 meta-analysis in Lancet Psychiatry that examined only rigorous randomized controlled trials of programs delivered during incarceration found essentially no measurable effect. The gap likely reflects differences in implementation quality rather than proving CBT useless, but it is a reminder that offering a program is not the same as offering a well-run one.
The First Step Act of 2018 built the federal system around this evidence. The law requires the Bureau of Prisons to provide evidence-based programs designed to lower reoffending risk, and it lets eligible inmates earn time credits toward early transfer to home confinement or supervised release by participating.10Office of the Law Revision Counsel. 18 U.S.C. 3632 – Development of Risk and Needs Assessment System11United States Sentencing Commission. First Step Act Earned Time Credits
The system runs on a risk assessment tool called PATTERN (Prisoner Assessment Tool Targeting Estimated Risk and Needs), which classifies each federal inmate as minimum, low, medium, or high risk for both general and violent reoffending. PATTERN weighs fixed factors like age, criminal history score, and whether the current offense was violent alongside changeable factors like disciplinary infractions, program completion, education level, and drug treatment participation. The changeable factors are the point. They give inmates a concrete path to lower their classification by earning a GED, completing vocational training, or finishing substance abuse treatment. To qualify for earned time credits toward early release, a person’s general and violent risk scores must both fall in the minimum or low range, and the statute requires periodic reassessment so someone who starts at high risk can work their way down.12Office of the Law Revision Counsel. 18 U.S.C. 3632 – Development of Risk and Needs Assessment System
What the research consistently shows is that in-prison programming alone is not enough. Jurisdictions seeing the most meaningful reductions in recidivism combine prison-based programming with real post-release support for housing, employment, and supervision. The structural barriers waiting on the other side of the prison gates do most of their work after someone walks out, and that is where the next arrest usually begins.