nonacademicresearch.org Editorial
The United States incarcerates approximately 2 million people — more than any other country in absolute terms and among the highest rates per capita globally. The central justification for incarceration beyond punishment is that it reduces future crime, either by deterring would-be offenders or by incapacitating people who would otherwise reoffend. The empirical evidence on recidivism — reoffending after release — does not support the view that current incarceration practices reduce reoffending. Longer sentences do not produce lower reoffense rates, and some evidence suggests incarceration increases recidivism risk relative to community-based alternatives.
The United States incarcerates a higher proportion of its population than any other population — roughly 2 million people at any given time. The stated purpose of incarceration includes deterrence, incapacitation, rehabilitation, and retribution. The empirical evidence on whether increasing incarceration reduces crime — and whether prison rehabilitates or increases recidivism — is more complicated than policy debates typically acknowledge.
U.S. incarceration rates quadrupled between 1973 and 2009, driven by mandatory minimum sentencing, truth-in-sentencing policies, the War on Drugs, and associated legislative changes. The U.S. rate of approximately 650 per 100,000 residents is five to ten times higher than comparable Western European nations.
This expansion occurred against a backdrop of falling crime rates, which began declining from their early 1990s peaks. Some policy advocates attributed the crime decline to increased incarceration; others argued the causes were independent (demographic shifts, the decline of crack cocaine markets, economic conditions, policing strategy changes). Distinguishing these explanations requires methods that isolate the specific contribution of incarceration.
Early work on this question (Levitt, 1996, American Economic Review) used prison litigation — overcrowding lawsuits that forced states to release prisoners independently of crime trends — as a natural experiment to estimate the crime-reducing effect of incarceration. Levitt found that each prisoner released was associated with approximately fifteen additional crimes per year, suggesting substantial incapacitation effects.
However, a crucial finding from subsequent research is that returns to incarceration are not constant. Liedka et al. (2006, Criminology) analyzed state-level panel data from 1972 to 2000 and found evidence for declining marginal returns: each additional prisoner produced less crime reduction as incarceration rates rose, with some evidence that at high incarceration levels, additional imprisonment might even increase crime (through criminogenic effects of prison and disruption of social networks). They estimated that the U.S. had likely already passed the point of net crime-reducing incarceration in most states.
A National Academy of Sciences review (Travis, Western, & Redburn, 2014, The Growth of Incarceration in the United States) concluded that the crime-reducing effect of additional incarceration in the United States was "small" and that continued growth in imprisonment would likely have "no effect or would cause increases in crime." The report attributed approximately 25% of the crime decline from 1990 to 2010 to the rise in incarceration — a meaningful but not dominant contribution.
The evidence on whether imprisonment reduces reoffending after release is largely negative. BJS data show that approximately two-thirds of released prisoners are rearrested within three years of release, and more than three-quarters within five years (Alper et al., 2018, BJS Special Report).
Nagin et al. (2009, Criminology) reviewed the experimental and quasi-experimental literature on the effect of imprisonment versus alternative sanctions on subsequent offending and concluded that imprisonment generally does not have a deterrent effect on the sanctioned individual (specific deterrence). If anything, some studies find that imprisonment increases subsequent offending relative to community-based sanctions — the "school of crime" effect — particularly for first-time or young offenders.
Mueller-Smith (2015, working paper, University of Michigan) exploited random assignment to judges with different sentencing tendencies in Harris County, Texas, to estimate the causal effect of incarceration on recidivism. He found that incarceration substantially increased subsequent criminal activity — felony incarceration for defendants who would otherwise have received probation increased recidivism rates by 50 to 60 percentage points over a four-year period, driven by labor market disruption and social network effects.
The evidence on what actually reduces reoffending is more encouraging for structured programs. A meta-analysis by Lipsey (2009, The Prison Journal) of 548 studies of juvenile justice interventions found that certain program types were consistently effective at reducing recidivism: therapeutic approaches (counseling, skill building, restorative justice), mentoring programs, and educational and vocational training. Purely punitive programs — boot camps, scared straight interventions, intensive supervision without services — were either ineffective or counterproductive.
For adults, cognitive-behavioral therapy (CBT) delivered in prison or community settings has the strongest evidence base. A meta-analysis by Landenberger and Lipsey (2005, Journal of Experimental Criminology) found that CBT reduced recidivism by approximately 30% on average across 58 studies. Substance abuse treatment for drug-involved offenders is also among the most consistently effective interventions.
Education programs in prison — General Equivalency Degree programs, college programs — have been associated with significant recidivism reductions in randomized and quasi-experimental studies. A RAND Corporation study by Davis et al. (2013) found that inmates who participated in correctional education programs were 43% less likely to return to prison than those who did not.
Incapacitation effects are real — people who are incarcerated cannot commit crimes in the community during their sentence. The debate is about whether the marginal prisoner being incarcerated at current U.S. rates produces net crime reduction or, as some researchers argue, net crime increase through criminogenic prison effects after release.
Some researchers also argue that crime rates are primarily driven by forces — demographic age structure, drug market cycles, economic conditions — that are largely independent of criminal justice policy, limiting the potential effect of any incarceration policy change.
The evidence does not support the proposition that increasing incarceration at current U.S. levels significantly reduces crime. Returns to incarceration appear to have diminishing marginal returns, and current rates likely exceed the crime-minimizing point. Recidivism rates after prison are high, and evidence suggests imprisonment often increases rather than reduces subsequent offending — particularly for first-time offenders and those incarcerated for short periods.
Evidence-based alternatives — CBT, educational programming, substance abuse treatment, community supervision with services — reduce recidivism at lower cost than incarceration. The current U.S. incarceration system is expensive (approximately $35,000 per prisoner per year federally), criminogenic in important respects, and not achieving its stated rehabilitative goals for most offenders.
nonacademicresearch.org Editorial (2026). Does Incarceration Reduce Recidivism? What the Evidence Shows. nonacademicresearch.org. nar:i4b9ymbt5gjjv15zmm
@misc{98mztda7,
title = {Does Incarceration Reduce Recidivism? What the Evidence Shows},
author = {nonacademicresearch.org Editorial},
year = {2026},
howpublished = {nonacademicresearch.org},
note = {nar:i4b9ymbt5gjjv15zmm},
}Temporary identifier. This paper carries a temporary nar:* identifier valid for citation within the independent research community. A permanent DOI will be minted via DataCite once the platform completes nonprofit registration.
Log in to join the discussion.
Loading…