A new analysis of America’s criminal-justice data challenges one of the most common explanations for the country’s recent decline in imprisonment. The shrinking prison population and narrowing racial gap, researchers report, were not driven by a single nationwide shift toward leniency. Instead, the forces behind the contraction of the carceral system varied sharply by offence. For violent and property crimes, falling levels of recorded crime appear to have done most of the work. For drug offences, by contrast, prison admissions declined even though there is little evidence that drug use itself fell. The study suggests that the United States is moving away from mass incarceration through a combination of reduced crime and selective changes in policing, prosecution and sentencing.
The findings come from an examination of recorded crimes, arrests and prison admissions between 2000 and 2019. Jason Robey and Michael Light used these linked stages of the criminal-justice process to decompose changes in imprisonment by race and offence category. Their approach distinguishes between how much crime is recorded, how often police make an arrest after a crime is recorded, and how likely an arrested person is to enter prison. That distinction is crucial because prison populations can change for very different reasons. Fewer offences can reduce incarceration without any change in official policy, while fewer arrests or shorter prison sentences can lower imprisonment even when the underlying behaviour remains stable.
This framework also helps explain the historical rise of mass incarceration. During the period when imprisonment expanded most dramatically, the researchers say, the growth was largely associated with increased punitiveness in policing and case processing. In practical terms, that means the criminal-justice system became more likely to respond to observed or suspected offending with arrest, prosecution and imprisonment. The number of people entering prison therefore reflected not only the amount of crime in society but also the increasingly severe consequences attached to it. The new analysis indicates that the later decline in imprisonment cannot simply be understood as the reverse of that earlier process. The mechanisms that reduced incarceration were often different from those that had produced its extraordinary expansion.
For violent and property offences, the strongest explanation for falling prison admissions was a reduction in recorded crime. As fewer violent and property offences entered official statistics, fewer cases moved through the subsequent stages of the justice system. The researchers find that these declines were especially important among Black Americans, including a substantial reduction in violence. This result is significant because it suggests that the narrowing of racial disparities in imprisonment was not caused solely by equal treatment after arrest or by sentencing reform. Changes in the level of recorded offending, particularly violent offending, played a central role in reducing the number of Black people exposed to the criminal process in the first place.
The study’s use of the word “recorded” is technically important. Official crime data do not capture every offence that occurs; they reflect incidents reported to or detected by authorities and then classified as crimes. A decline in recorded crime can therefore arise from a genuine reduction in offending, changes in reporting behaviour, enforcement priorities or the way agencies record incidents. The researchers’ decomposition does not claim that every fall in official crime represents a direct change in underlying behaviour. Rather, it identifies the stage at which the measurable reduction occurred. In the data examined, the largest contribution for violent and property imprisonment came from fewer recorded offences, not from a broad decline in the probability that an arrested person would be imprisoned.
Drug offences produced a markedly different pattern. The researchers found little evidence that declining drug use explains the fall in drug-related prison admissions. This does not mean that drug consumption was constant in every community or that no aspect of drug activity changed during the period. It means that the available evidence did not show a decline large enough to account for the scale of the reduction in imprisonment. Instead, the decline appears to have occurred later in the process, through reduced probabilities of arrest and imprisonment. Police became less likely to arrest people for drug-related offences, and people who were arrested became less likely to receive a prison sentence.
The racial implications of that shift are particularly striking. Among Black individuals arrested for drug-related offences, the probability of imprisonment was cut roughly in half between 2000 and 2019. That change means that two people with similar contact with the justice system could face very different incarceration risks depending on when their cases were processed. It also illustrates why prison admissions cannot be interpreted as a simple measure of drug use or drug offending. Admissions are the product of multiple decisions: whether authorities detect an offence, whether they make an arrest, whether prosecutors pursue a case and whether courts impose imprisonment. A change at any one of these points can substantially alter the prison population without a comparable change in the behaviour that initially drew official attention.
Together, the results offer a more complicated account of America’s retreat from mass incarceration than a story centred on one reform movement, one political party or one national change in public attitudes. Crime reduction was central to the decline in imprisonment for violent and property offences, with falling violence among Black Americans playing an especially important role in reducing racial disparities. For drug offences, however, institutional decisions appear to have mattered more than changing levels of use. The study therefore presents the contraction of the carceral state as a hybrid process: fewer recorded crimes reduced the flow of people into prison in some categories, while less aggressive enforcement and sentencing reduced it in others. Understanding that distinction may be essential for evaluating whether future reductions in incarceration reflect safer communities, fairer institutions or both.
Subject of Research: Changes in U.S. incarceration rates and racial disparities from 2000 to 2019, analyzed by race and offence category.
Article Title: Why America’s prisons are shrinking and racial gaps are closing
Article References: Robey, J., Light, M. Why America’s prisons are shrinking and racial gaps are closing. Nat Hum Behav (2026). https://doi.org/10.1038/s41562-026-02536-y
Image Credits: AI Generated
DOI: https://doi.org/10.1038/s41562-026-02536-y
Keywords: incarceration, mass incarceration, racial disparities, prison admissions, criminal justice, recorded crime, arrests, drug offences, violent crime, sentencing reform
