Inmates at Devens federal prison in Massachusetts in 2015. (Nikki Kahn/The Washington Post)

In an important op-ed in the New York Times last week, Marc Morjé Howard argued in favor of parole — not just for drug offenders but also for violent offenders. It isn’t an easy argument to make. But it’s past time to start making it.

Prisoners whose sentences include a range of years — such as 15 to 25 years, or 25 years to life — can apply to their state’s parole board for discretionary parole, but they almost always face repeated denials and are sent back to wither away behind bars despite evidence of rehabilitation. (Inmates who have served their maximum sentence are released on what is called mandatory parole.)

Rejection is usually based on the “nature of the crime,” rather than an evaluation of a person’s transformation and accomplishments since they committed it. The deeper reason for the rejection of discretionary parole requests is simple: fear. Politicians and parole board members are terrified that a parolee will commit a new crime that attracts negative media attention.

There are several theories behind why we incarcerate people who commit crimes. One reason is to protect society from those who would hurt others. But as Howard points out, the vast majority of even violent criminals “age out” of crime. If you were to draw a curve of at what age most crimes are committed, that curve starts sloping downward from the early 20s on.

But are prisoners who have served long sentences for violent crimes genuinely capable of reforming and not reoffending? The evidence says yes. In fact, only about 1 percent of people convicted of homicide are arrested for homicide again after their release. Moreover, a recent “natural experiment” in Maryland is very telling. In 2012, the state’s highest court decided that Maryland juries in the 1970s had been given faulty instructions. Some defendants were retried, but many others accepted plea bargains for time served and were released. As a result, about 150 people who had been deemed the “worst of the worst” have been let out of prison — and none has committed a new crime or even violated parole.

As Howard concedes, there will always be a small population of prisoners who aren’t capable of rehabilitation. But the vast majority are, particularly among those in their mid-30s and older.

We also incarcerate people as a deterrent — so others won’t commit similar crimes. But the research of Mark Kleiman (and others) suggests that to serve as an effective deterrent, punishment needs to be swift and certain. What we have is the opposite. We hand out severe punishments, but we do so inconsistently, and with long lags between the crime itself and the imposition of a sentence. (Kleiman argues — persuasively — that implementing less severe punishment more consistently and more swiftly is a formula for both less crime and less incarceration.)

Incarceration can also be used for rehabilitation. This in fact was the original intent behind incarceration as punishment in the United States, starting with the first modern prison in the country, Philadelphia’s Eastern State Penitentiary. That facility originally taught prisoners a trade so that they’d have a marketable skill when they were released. Some practices were badly misguided — for example, the prison also forced inmates to spend long stretches in solitary confinement, not as a punishment, but so they could repair their relationship with God. But the intent was repair, not punishment for punishment’s sake.

Retribution is the dominant paradigm driving incarceration today. We want to punish criminals. We want them to suffer. We create hostile prison environments rife with violence, racial resentment and rape. This is what politicians have run on and demagogued since the Warren Court.

If punitiveness is your primary objective, you probably aren’t going to find Howard’s op-ed very persuasive. It’s a pretty primitive way of thinking about crime and punishment, but also an understandable one. If a murder victim didn’t get to experience life and freedom in middle age, why should that victim’s killer?

But there are costs to a system that runs on retribution. There are the tangible costs of warehousing and providing health care for a large and growing prison population that you intend to keep behind bars until death. Those costs are massive and growing and will soon be overwhelming. There are also unseen costs in the unrealized potential of people capable of redemption. There’s an enormous population of prisoners capable of and willing to contribute to society. In fact, it seems intuitive that redeemed people who feel they owe a debt to the community will work harder than the rest of us to pay it off, and ask for less in return. If your aim is to make people who take a life suffer for the rest of their own lives, then yes, keeping a reformed man in his 40s imprisoned until death certainly accomplishes that. But it also punishes the people he might have helped, provided a service to or even inspired with his rehabilitation. Vengeance is also expensive. Howard estimates taxpayers pay more than $30,000 per year per prisoner. That’s already a lot, but that figure will go up as a prisoner ages. All of which means that as prisoners grow less and less likely to reoffend, our determination to keep them behind bars gets more and more expensive.

Of course, paroling violent offenders isn’t without costs. Our system is far from perfect, which means that inevitably someone incapable of rehabilitation will convince a parole board that they’ve changed, and they’ll commit another violent crime. As Howard points out, the data suggest this is extraordinarily rare. That’s in part due to how few people convicted of violent crimes are paroled. But as Howard also points out, of those who have been, a very small percentage go on to commit violent crimes. This, even in a society that hasn’t been particularly kind or welcoming to those with a criminal record. But paroling more people convicted of violent crimes will inevitably, at some point, somewhere down the line, produce a repeat offender. The data overwhelmingly suggest that such incidents will be rare enough to be drastically overwhelmed by the benefits of a more generous and forgiving parole policy. But those rare incidents will be easy to exploit. Advocates should be prepared for them.

In the end, this is a question of what sort of society we want to be. We can be a punitive society that believes in retribution, no matter the costs. We can be a society that believes in redemption, regardless of cost. Or we can be a society of people who strive for a rational, data-driven system that will never be perfect, but that will strive to protect us from truly dangerous people while also recognizing that, as the attorney and activist Bryan Stevenson puts it, “each of us is more than the worst thing we’ve ever done.”