If there’s one thing the United States is good at, it’s putting people in prison.

The number of people incarcerated in America has quadrupled in the last three decades. At last count, the country housed 1.5 million inmates.

At a time when states are still struggling to balance their budgets post-recession and increase funding on education and health care, the U.S. will spend $40 billion on incarceration this year alone, according to the National Conference of State Legislators.

There are many broad societal arguments to be made about changing sentencing guidelines, but take emotions out of it—as economists often do—and it may be easier to see, from a dollars and cents perspective, which prisoners it might make the most economic sense to set free.

To cope with the cost, you could argue—as some have—that the country should consider releasing a portion of the prison population (or not put them there in the first place) to reduce overcrowding and to save money. But it’s not clear that letting people out of prison will necessarily help states' bottom line. That’s because there are also very calculable costs of crime to society.

Estimates differ on exactly how much each crime costs, but one recent study added up a number of factors, including the cost of a victim's medical care, property loss, and lost earnings; spending on police protection, legal services, and incarceration; crime career costs (money an individual loses by engaging in crime rather than in legal activities); and added expenses for intangible costs. It found that murder costs society about $8.9 million per offense, rape costs about $241,000 per offense, aggravated assault costs $107,000, and robbery costs $42,300. (For estimates for arson, embezzlement, vandalism and other costs, take a look here).  

So what’s a state to do if it wants to save money on correction costs, but also prevent crime and the costs that come with it? California was ordered to reduce its prison population by 30,000 in a 2011 Supreme Court decision that found the state’s overcrowded prisons violated the Eighth Amendment’s ban on cruel and unusual punishment—an order that Justice Alito wrote was “gambling with the safety of the people of California.” Various news stories suggest the state hasn't quite figured out the best way to release criminals without incurring big costs to society. Freeing those believed to be the least risky has led to the elimination of incarceration for people who were found guilty of child abuse and domestic violence. One man who was freed after violating parole then went and raped and murdered his grandmother, which is certainly a big cost to society by any measure.


The Cost of Crime to Society

Figures in 2008 dollars (Data: The Cost of Crime to Society, Drug and Alcohol Dependence Volume 108)

It raises the question: Is it possible to take a different approach to incarceration, and save money at the same time?

One Penn economist, David Abrams, has tackled this question, in a paper called “The Imprisoner’s Dilemma: A Cost-Benefit Approach to Incarceration,” in which he looks into the most cost-effective ways to balance the price of incarceration with the costs of crime to society.

One important question to start with: How does prison reduce the cost of crime to society in the first place? Criminologists have a few ideas on this, but they basically boil down to this:

  1. A fear of prison and long prison sentences keep people from performing crimes (this is called general deterrence).
  2. The experience of prison makes people who have served time want to avoid it, and thus not perform any more crimes (specific deterrence).
  3. When people are in prison, they’re not doing any criminal damage in the outside world (unless they’re in gangs or have smuggled cellphones).

The concerns with reducing sentences and letting people out of prison, then, are obvious: If sentences are shorter, more people will commit crimes because they’re not afraid of going to jail, and if more people are on the streets rather than in prison, they’ll be free to engage in more criminal activity.

But, Abrams wondered, is there a way to figure out which policies are most effective in reducing both the cost of crime to society, and the cost of punishment? He tested a few policy prescriptions and found that some could have a net benefit to society, when the costs and benefits are balanced.

Let’s look at the simplest one first: A one-time prisoner release. A release of inmates with the least amount of time on their sentences—the oldest, or those who have committed the least-serious crimes—could produce big net benefits, he found. Early release of certain prisoners convicted of assault, who are unlikely to reoffend, produced a net benefit of $35,138 per year per released prisoner. Releasing certain prisoners convicted of robbery, burglary, and motor-vehicle theft produced a similar net benefit.

This is assuming that there is some way to identify the prisoners who are likely to be the least-severe offenders, he said. (You can play around with this nifty calculator to measure the monetary benefits of releasing different numbers of prisoners.)

France had experimented with mass-prisoner releases until President Nicholas Sarkozy put a stop to it in 2007 because he thought it an inefficient way to manage prison populations—some studies imply the releases didn't reduce crime, though. Italy also issued collective pardons for decades to reduce overcrowding, though offenders could go back into prison, and serve longer terms, if they committed crimes once pardoned, an effort that proved somewhat effective at lowering recidivism.

A second policy prescription Abrams investigates: The idea of reclassifying crimes so that some lower-level offenses don’t lead to prison time. Change the rules so that stealing $1,500 no longer leads to incarceration, but $4,500 does, for example, and you’ll be locking up fewer people, while still getting the most costly criminals off the street. This is relevant in the American criminal justice system because so many people are incarcerated for parole violations; taking away the possibility of prison time for these crimes could save a lot of money, and potentially not lead to a significant cost to society.

Abrams finds that reclassifying crimes has a net benefit to society, too, in crimes including assault, burglary and motor-vehicle theft. Reclassifying crimes in robbery, usually a violent crime, actually will cost society about $1,000 per prisoner, though.

“If we think we can do a good job of segregating out the severe crimes and criminals, then we almost certainly should,” he told me. “That’s an important policy implication.”

Of course, it can be difficult to figure out how to reclassify crimes: In California, as part of a “prison realignment,” counties tried to let the least-risky convicts out of prison, but soon found that many of these people soon committed more offenses.

That leads to the third, and most counterintuitive policy prescription—increasing sentence lengths. It sounds more expensive, since people will stay in prison for longer stretches, at a larger cost to taxpayers. But longer sentences could deter more people from committing crimes, since they know that lots of jail time awaits them. It could also keep more serious criminals, who would otherwise potentially be committing crimes, off the streets, preventing them from costing society more money.

Abrams calculated the costs and benefits to society if all sentences were increased by 10 percent, using four different estimates on the cost of crime. Abrams finds that increasing sentences also has a net benefit to society, for many crimes. The net benefits of extending sentences for assault can be around $10,000 per prisoner, robbery can be around $3,500 per prisoner, motor-vehicle theft can be around $2,860, for example.

“This is certainly counterintuitive—a lot of people think that we have a prison population that is too large. I think we probably have a prison population that’s too large,” he said.

That fact that all three of these policy ideas produced a net benefit to society implies that, from an economist’s perspective, America's doing something wrong in the way it imprisons people. How to fix it, though, is another question.