If a small number of offenders are responsible for a hugely disproportionate share of crime, then locking up exactly that small group ought to be one of the most efficient ways to cut crime rates. That simple idea – identify the career criminals, separate them from society, and the rest more or less takes care of itself – has genuine appeal. It also turns out to be far harder to execute fairly than it sounds.
A landmark study by Marvin Wolfgang at the University of Pennsylvania is where the modern version of this idea really starts. Wolfgang tracked 9,945 men, all born in Philadelphia in 1945 and all residents of the same part of the city, from birth into adulthood. The results were striking. Of the men in the study, 627 had already been arrested five or more times by the time they turned 18 – and this small group, just seven percent of the total cohort, was responsible for roughly two-thirds of all the crime committed by the entire group. Fourteen of them were murderers. On average, each had served only about four years in prison.
Wolfgang later repeated the study with a cohort of men born in the same area in 1958 – 12,160 men this time, of whom 982 turned out to be repeat offenders. That, again, worked out to almost exactly seven percent. But the second cohort was noticeably more violent than the first: rape incidents had doubled, and murder and robbery had roughly tripled. These offenders also tended to start younger and evade arrest longer than the men in the earlier study.
On paper, identifying and incapacitating this group looks like a clean solution. In practice, the numbers get uncomfortable fast: scaling Wolfgang’s seven percent figure up to the entire United States implies roughly 1.22 million offenders who would need to be incarcerated at once – on top of a prison system that is already badly overcrowded.
How Much Crime Does One Career Criminal Actually Commit?
The RAND Corporation studied 2,190 career offenders directly and found the pattern held: a small minority accounted for a hugely disproportionate volume of crime, much of it never leading to arrest. Half of the burglars in the study had committed at least six burglaries within eighteen months; more than ten percent had committed over twenty. Half of the robbers had committed five armed robberies in a single year. On average, every ten offenders in the sample were connected to 87 separate incidents. Offenders dealing drugs averaged around 100 incidents a year. Locking up even one such offender for a single year, by RAND’s estimate, prevents roughly four crimes that would otherwise have occurred.
Two factors stood out as the clearest predictors of repeat offending: the age at which someone started committing crimes, and drug use. Roughly half of all serious crime in America is committed by offenders under 18. The heaviest offenders tend to start at thirteen or fourteen and keep offending well into their sixties – but there is also a well-documented “aging out” pattern: an offender who goes fifteen years without a repeat offense is unlikely to become a career criminal as an adult. RAND also estimated that a heavy heroin user spending roughly $50 a day on the drug typically supports that habit through repeated theft and robbery – and found that 242 heavy drug users, tracked over eleven years, had collectively committed nearly half a million crimes, or roughly 2,056 offenses each on average.
Seven Warning Signs
Based on this research, RAND identified seven characteristics associated with career offenders: a criminal sentence before age sixteen; time spent in a juvenile facility; heroin or barbiturate use before the current arrest; heroin use as a teenager; a conviction more than two years prior to the current arrest; a prior conviction specifically for burglary or robbery; and one to two years already spent in prison before the current arrest. Applied retroactively to 781 sentenced offenders in California, this seven-point profile held up reasonably well.
A separate line of research from Robert Martinson complicates the picture further. Of roughly 400,000 offenders held under special rehabilitation programs, only twelve percent were rearrested – yet Martinson found no rehabilitation program that reliably reduced repeat offending on its own. A century ago, nearly every convicted offender in America went to prison; today only about nine percent do. Martinson’s conclusion was that decades of well-intentioned experimentation had, in effect, thrown away a genuinely effective tool. If police, prosecutors, investigators, and judges coordinate well, the criminal justice system can become substantially more effective – fewer offenders in prison, and safer homes and streets, are not necessarily in tension with each other. Researcher Lisa Schifron has argued that a refined version of offender profiling, used carefully, could make identifying career criminals considerably easier over time.
Why “Just Lock Them Up” Isn’t So Simple
Not everyone agrees that incapacitating career offenders is workable, even in principle. The critic Stuart Hoff has pointed out the obvious practical problem: who, exactly, decides which prisoner stays locked up and which one gets released? Any error in that judgment cuts both ways – some people end up unjustly confined, while genuine career criminals walk free and reoffend. The idea of confining a select group of offenders for a fixed period specifically to strip them of their capacity to offend – rather than to punish a specific crime – has been debated within criminal justice circles since 1975, and the debate has never fully settled.
Building directly on Wolfgang’s research, the criminologists Shlomo and Reuel Shinnar developed a formula suggesting that if a typical career offender commits roughly ten crimes over ten years, incapacitating that offender for five years could cut crime by as much as 80 percent. That number proved highly controversial. Researchers including Stevens Clarke (1974), David Greenberg (1975), and Stephen Van Dine (1977) challenged the estimate directly and found the real effect closer to just four percent – a massive gap that, even accounting for the limitations of every study involved, pointed to the need for far more careful estimates before anyone acted on the original figure.
The book Sense and Nonsense About Crime lays out the hard questions this approach still has to answer: Can we measure the reduction in crime with real precision? Can we correctly identify career offenders and accurately predict their future criminal behavior? Would incapacitating them require building new prisons – and if so, where does that money come from? Can this be done without violating the constitution? And what are the side effects of the policy once it’s in place? One more question belongs on that list: what happens to the overall crime rate once you try to fold selective incapacitation into the wider system of punishment?
There is also a subtler statistical problem. The bigger the estimate of how much crime a “typical” career offender commits each year, the more dramatic the projected impact of locking them up looks – but that estimate is only an average. If a handful of extreme offenders commit a hundred crimes a year each, they can single-handedly inflate the average for the whole group, making an ordinary repeat offender look far more dangerous, on paper, than they actually are. On top of that, there still isn’t an agreed-upon definition of exactly what makes someone a “career criminal” in the first place.
Can We Actually Identify These People?
According to police in one American state, out of a thousand offenders, only 22 were flagged as high-frequency offenders – but police themselves openly admit that “high-frequency” is essentially a guess. They don’t know, among those 22, who will end up back in prison and who won’t, and they aren’t confident how many of the 22 were flagged based on false tips from other offenders looking to settle scores. By the police’s own account, they simply don’t have full confidence in who is actually who on that list.
The paperwork behind these judgments is shakier than it looks. When a defendant appears before an American judge, prosecutors present a list of alleged priors meant to shed light on the defendant’s history – but these lists are frequently incomplete. They record accusations, not necessarily convictions, and cases dropped through complicated legal proceedings often vanish from the record entirely, leaving juries working from an incomplete picture of who they’re actually dealing with. There are also offenders with a bad reputation who haven’t actually committed much of anything – simply people who are bad at getting away with what little they do try, and get caught every time.
Statisticians describe the two ways this kind of prediction fails as false negatives and false positives – offenders wrongly cleared, and offenders wrongly flagged. Even the best predictive models for career offending make both kinds of mistakes regularly; a prediction that’s right 60 percent of the time is already considered a strong result in this field. That is worth sitting with: a coin-flip-plus-a-bit is the realistic ceiling for identifying who among a group of offenders will keep offending and who won’t, and every model still leans partly on actuarial life-table assumptions that don’t always hold as people’s circumstances change.
The Price Tag — and the Deeper Problem
Van Dine argues that selective incapacitation is politically appealing precisely because it sounds so clean, while the real cost of implementing it is genuinely destructive. By his estimate, a mandatory five-year sentence applied under this model would grow the adult prison population by only five to six percent – from roughly 13,000 to 25,000 within five years in the jurisdiction he studied – while requiring $5 million to $120 million in new prison construction, in exchange for a 25 to 28 percent drop in crime. The open question is whether American society is willing to actually pay that price, and how anyone proposes to calculate the honest cost-benefit ratio behind it.
The approach also raises uncomfortable questions of its own. Should decisions about future dangerousness rest on factors like race, employment status, social class, or a person’s past personal history, rather than strictly on crimes actually committed? Predictions of this kind lean heavily on housing stability, marital status, employment, and family support – all factors that correlate with future offending in ways that are also deeply entangled with race and class. A parole board weighing an older offender’s release faces a very different calculation than one weighing a young first-time offender’s sentence, and it isn’t obvious that the same predictive standard should apply equally to both. In other words: is this kind of prediction actually fair, in a legal sense that would hold up to scrutiny?
There is a reasonable counterargument that courts and police already do a version of this informally, relying on accumulated professional judgment and experience – American courts already lean harder on offenders they consider dangerous or repeat-prone, particularly for serious crimes, and reform institutions already know that habitual violent offenders often get shuffled out of overcrowded facilities sooner than anyone would like. Courts already exercise this kind of practical judgment when handing down different types of sentences for offenders they consider more or less likely to reoffend. Formalizing that judgment into policy would, at minimum, widen the demand for genuine reform and raise real questions about the growing costs to both society and the system meant to serve it.
A Good Idea in Theory, a Hard One in Practice
The honest conclusion is that selective incapacitation is a genuinely appealing idea in theory and a frightening, difficult one to execute in practice, because predicting human behavior with scientific precision runs into serious limits. The criminal justice system still has not found a reliable method for identifying career criminals while also correctly clearing the much larger number of offenders who are not career criminals – and that gap is a real and painful problem for any justice system to sit with. A system that ends up locking away innocent people, or people who would never have gone on to become career offenders, is not something any society should be willing to accept. That, more than anything else, is why this idea – elegant as it looks on paper – has never been fully put into practice.
No Comments