We know for a fact—as credible a fact as we get from the criminal justice system—that more than eight out of 10 prisoners in the United States are rearrested after release. These data suggest some inconvenient truths about our criminal justice system—inconvenient, that is, to putative "reformers" determined to decarcerate the United States.
First, the high reoffending rates suggest that those in prison are typically there for a good reason. Unless one believes that prisons are so criminogenic (meaning, that they encourage criminality) that inmates enter pure but leave corrupted, it is hard to explain the preponderance of post-release crime. In fact, there is little evidence that prison is criminogenic. If it were, then longer sentences would produce more crime. But a 2021 review of the relevant research found "no substantial evidence" that longer sentences have a criminogenic effect. In other words, there is little indication that longer sentences generate more crime after release, or that shorter sentences produce less.
Second, a relatively small number of people commit a large proportion of the serious crimes. Around 5% of the United States population is responsible for half of all crime. These "ravenous wolves" keep repeating their predations until the criminal justice system finally incapacitates them.
Third, rehabilitation within prison walls is largely ineffective. If it worked, we simply wouldn't be getting an 83% recidivism rate for discharged prisoners—which is what the data evince.
Lately, many analysts sidestep these truths and attack the very concept or use of recidivism as an appropriate criminal justice yardstick. One recent paper argued that recidivism is such a crude indicator that relying on it is akin to "insanity": "it does not lead to any new information that could be used to shape an effective response" to crime.
It is true that recidivism measures are imperfect. For instance, if a police force is deployed in increased numbers to a high-crime area, there will be more arrests and of course more recidivists from that neighborhood. If the neighborhood is relatively poor and black, there would be increased numbers of poor black recidivists.
Likewise, if there is a spike in a particular type of crime—say, selling drugs on the street—and a prosecutor increases the demand for incarceration of drug arrestees, that too could generate more recidivism.
In both of the above hypothetical scenarios, one could argue that it was policy changes—as opposed to changes in human behavior—that produced more recidivism. But so what? The various "imperfections" in measuring recidivism do not render its penological use inappropriate. Would it be "insanity" to ignore the crime history of poor black recidivists or drug dealers simply because law enforcement policy changes increased their chances of apprehension and conviction? That is the equivalent of telling a judge that you shouldn't have to pay a fine for speeding because the cops never used to put a patrol car on that street before. In fact, recidivism provides an excellent prediction of post-release behavior and should be used to shape the response to misconduct.
Moreover, virtually all of the criticisms of recidivism apply to measures of crime itself. After all, the definition of any specific crime varies from state to state, and enforcement differs from police force to police force and from prosecutor to prosecutor—to say nothing of conviction and sentencing variations among judges. Should we therefore scrap crime itself as a measure of social health or local habitability, because of these ineluctable inconsistencies?
The answer is obvious. Humans can only make policy based on the best information available. We cannot wait for data perfection.
The best information on recidivism is that it is a very good predictor of criminal behavior. While we know that the overwhelming majority of released prisoners reoffend, it also turns out that those who have no prior arrests or convictions before their imprisonment are most likely to be crime-free upon release. The U.S. Sentencing Commission once examined the crime histories of federal prisoners and found that those without prior criminal histories had a mere 6.8% recidivism rate.
Likewise, a new study of New York State's bail reforms showed that the reforms—surprisingly—did not cause more people to be arrested as a result of the reduction in jailings. However, this was mainly true for non-repeaters. Defendants without a pending case were less likely to be rearrested after bail reform. But those with a pending case were 6% more likely to be rearrested after the reforms. In short, freeing recidivists is risky and judges should be able to take defendants' arrest and conviction history into account when determining whether to release or jail them.
Decarcerationists want to discredit recidivism because they don't like its implications. But recidivism remains the gold standard for criminal justice outcomes. Consideration of recidivism does not preclude consideration of other measures, too. We should, for instance, look at addiction reduction rates for drug intervention programs and employment rates for vocational training programs. But if these interventions are targeting prisoners or former prisoners, we will always want to know if the subjects have been arrested again.
Barry Latzer is emeritus professor at New York's John Jay College of Criminal Justice and author of The Myth of Overpunishment: A Defense of the American Justice System and a Proposal to Reduce Incarceration While Protecting the Public (Republic Books).
The views expressed in this article are the writer's own.
Uncommon Knowledge
Newsweek is committed to challenging conventional wisdom and finding connections in the search for common ground.
Newsweek is committed to challenging conventional wisdom and finding connections in the search for common ground.
About the writer
To read how Newsweek uses AI as a newsroom tool, Click here.
