Five myths about the U.S. prison system

Criminal justice policy has secured a prominent place in the Democratic presidential primary race, with several candidates offering plans to reform the way Americans punish. Reinventing how we handle crime is one of the most important civil rights challenges of our time, yet much of the debate about criminal justice is built on misconceptions that often push reform in inadequate and even counterproductive directions.

• Myth No. 1: U.S. prisons are full of nonviolent drug offenders.

Asked recently about voting rights for felons, Sen. Cory Booker of New Jersey, one of the Democratic presidential candidates, claimed that "we locked up more people for marijuana in 2017 than all the violent crimes combined." Rep. Alexandria Ocasio-Cortez, D-N.Y., has echoed that view, suggesting on Twitter that the prison system is defined by nonviolent people "stopped w/ a dime bag."

But the simple truth is that, at a minimum, 55% of those in state prison have been convicted of a violent crime — and more than half of these people, or nearly 30% of the total prison population, have been found guilty of murder, manslaughter, rape or sexual assault, according to the Bureau of Justice Statistics. Slightly less than 15% are incarcerated for drug crimes, even though most Americans believe the figure to be about 50%. (Drugs play a bigger role in the federal prison system, but that holds only about 10% of all prisoners; most incarcerated people are in state prison.)

The share of those in state prison for committing violence is even greater than 55%, however. Prisoners are classified by the most serious offense for which they are convicted, not arrested or charged. So if someone is arrested for a violent crime but ends up pleading guilty to a drug charge, his crime is classified as a nonviolent drug offense, even if the underlying incident — like a domestic violence case in which the victim won't testify — is the reason the prosecutor sought prison time.

• Myth No. 2: Private prisons drive mass incarceration.

When people try to explain how the United States ended up with nearly 25% of the world's prisoners, they often point to firms that directly profit from incarceration by running prisons or by providing services to public facilities. At a recent presidential campaign event, Sen. Elizabeth Warren, D-Mass., blamed private prisons for mass incarceration ("We need to get rid of for-profit, private prisons"). Sen. Bernie Sanders, I-Vt., focused on private prisons in his 2016 presidential bid and is doing so again ("The private prison racket has got to end").

Second, of the roughly $50 billion we spend on prisons, about two-thirds, or $30 billion, is spent on wages and benefits for public-sector employees. In comparison, private prison firms collectively earn a few billion in revenue and (more important for their incentives to lobby) about $300 million in profits — just 1% of the public-sector wage bill. So public-sector correctional officer unions have a reason to lobby against reforms that would reduce inmate populations, especially since prisons often provide some of the only well-paying jobs in the rural communities where they are located.

• Myth No. 3: Long sentences are causing our prison population to age.

It is true that the average age of prisoners has increased in recent years, as has the number of inmates who are older than 55 or even 65. The common explanation is "ultralong sentences," as the Economist put it, or "the fallout of an era of long sentences," as a Nation article said — people serving very lengthy stints are growing old and dying in prison.

The reality is more complicated. Among prisoners who are over 65, fully half have served fewer than 10 years, which means they were admitted in their mid- to late-50s. Much of the aging of the prison population plausibly comes not from long sentences but from our inability to address late-age violence properly. Age-related illnesses like Alzheimer's and frontotemporal dementia can lead to violent behavior, and far too often we rely on police and jails, not public health services, to manage it. One study also suggests that extensive drug use when younger has led to higher levels of crime among a cohort of older Americans.

There is a problem with how we think about the other half of older inmates as well, those who were admitted when younger and are growing old behind bars. Even as we increasingly understand the problems and costs with people aging in prison, we often restrict reform efforts aimed at them, such as compassionate-release programs, to those convicted of nonviolent crimes — even though nearly 90% are serving time for violent crimes, frequently murder or manslaughter.

"Once a criminal, always a criminal?" was the headline on a CBS News article about recidivism rates. And in its 2011 decision in Brown vs. Plata, the Supreme Court pointed to recidivism statistics to claim that California's prisons don't rehabilitate but instead "produce additional criminal behavior."

Perhaps the single most important statistic for evaluating new criminal justice policies is their effect on recidivism. So it would be great if recidivism statistics actually measured reoffending. They don't. Formal recidivism numbers measure something like "the number of people rearrested" or "the number of people readmitted to prison." Such data both understates and overstates actual reoffending, glossing over a critical debate about what recidivism even means.

First, the undercounting: We can't measure reoffending that no one detects. No one voluntarily reports crimes they committed that the police did not observe or that victims did not report. This can exacerbate racial disparities, too: Considering the communities most heavily policed, reoffending by whites is much less likely to be detected than that by blacks.

The overcounting issue is less immediately obvious but arguably far more important. If someone shifts from committing one robbery per day to one per month, he could eventually be reconvicted and classified as a recidivist. That the amount of harm he caused fell by about 95% is immaterial. Ours is a binary metric; any one failure counts as recidivism. It's a metric inconsistent with what we know about the rocky reality of desistance.

• Myth No. 5: Not sending someone to prison saves about $35,000 a year.

It's easy to see where the number comes from: We spend about $50 billion to lock up 1.4 million people per year in state prisons. That's about $35,000 per person in prison per year. Yet that is not what we save when we release a single inmate from prison.

Recall that about two-thirds of prison spending is wages: If we don't cut payroll in proportion to inmate population, then the savings per prisoner will be much less. Other fixed costs, like water and heating bills, likewise do not decline steadily with the number of prisoners. The real savings of reducing the prison population by one is often about $4,000 to $16,000, not $35,000.

Now, if we release enough people that we can close an entire prison wing or facility, perhaps we can save something closer to the estimated average. But states often keep mostly empty facilities open, and when they do close institutions, they often relocate most of the correctional officers to other facilities.