This item has been corrected.
Last week, president Obama vowed to end mass incarceration, the imprisonment of 2.2 million Americans. He’s commuted the sentences of 46 drug offenders—but ending the practice will require a major policy change at the state and federal level. The sooner this is done, the better. Evidence from the last 40 years suggests the mass imprisonment policy was a tragic failure. Putting more people in prison not only ruined lives, it may have created more new crime than it prevented.
There are five times as many people in prison today—nearly 5% of the population will be imprisoned at some point—as there were in the 1970s. The increase in crime during the 1960s and ’70s motivated Americans to get tough on crime, which took several forms. The most striking of these was putting lots of people in prison. Imprisonment is supposed to reduce crime in two ways: it takes criminals off the street so they can’t commit new crimes (incapacitation) and it discourages would-be criminals from committing crime (deterrence).
But neither of these outcomes came to pass.
A new paper from University of Michigan economics professor Michael Mueller-Smith measures how much incapacitation reduced crime. He looked at court records from Harris County, Texas from 1980 to 2009. Mueller-Smith observed that in Harris County people charged with similar crimes received totally different sentences depending on the judge to whom they were randomly assigned. Mueller-Smith then tracked what happened to these prisoners. He estimated that each year in prison increases the odds that a prisoner would reoffend by 5.6% a quarter. Even people who went to prison for lesser crimes wound up committing more serious offenses subsequently, the more time they spent in prison. His conclusion: Any benefit from taking criminals out of the general population is more than off-set by the increase in crime from turning small offenders into career criminals.
High recidivism rates are not unique to Texas: Within 5 years of release more than 75% of prisoners are arrested again.
Why does prison turn people into career criminals?
Prison obliterates your earnings potential. Being a convicted felon disqualifies you from certain jobs, housing, or voting. Mueller-Smith estimates that each year in prison reduces the odds of post-release employment by 24% and increases the odds you’ll live on public assistance. Time in prison also lowers the odds you’ll get or stay married. Being in prison and out of the labor force degrades legitimate skills and exposes you to criminal skills and a criminal network. This makes crime a more attractive alternative upon release, even if you run a high risk of returning to prison.
You could argue prison is still worth it if long sentences discouraged people from committing crime in the first place. Mueller-Smith estimates a one-year prison sentence would only be worth it (in terms of prison cost and forgone economic potential) if it deterred at least 0.4 fewer rapes, 2.2 assaults, 2.5 robberies, 62 larcenies or prevented 4.8 people from becoming a habitual drug user. And the deterrent effect is not this powerful—not even close. There exists little evidence that the possibility of a long prison sentence is much of a crime deterrent at all.
If long prison sentences was truly a deterrent, juveniles would stop committing crimes when they turn 18 and face the possibility of more jail time. But they don’t. There are several reasons why: First, criminals, who are often young men, tend to be impulsive and to discount the future. The possibility of a longer prison sentence is too far away to pose as a deterrent. Second, even after they turn 18, there’s still a fair amount of uncertainty around the length of their sentence. The Harris County data show jail time often depends on the judge that’s assigned, the prosecutor, available evidence, and how busy the court is, which determines the odds that someone convicted ends up with a plea bargain. Longer sentences aren’t much of a deterrent because, even in our era of mass incarceration, the possibility of actually going to prison is too uncertain and too far in the future to meaningfully impact behavior.
The one exception seems to be when there’s a high degree of certainty around the punishment. The 3-strikes rule in California levies draconian sentences on 3-time offenders—with near certainty, even if the crimes are fairly minor. There is some evidence that if someone has two strikes, they are less likely to commit a third offense.
So if mass incarceration has not deterred crime, what could work?
Would-be criminals respond to incentives like everyone else. If they sense the cost and risk associated with crime has increased in a meaningful way, they are more likely to obey the law. That’s why an effective deterrent is increasing the perceived probability of being arrested. Studies show if arrests increase 10%, crime falls by 3% to 5%. The simplest way to increase the perceived probability is to put more police on the street. Florida State’s Jonathan Klick and George Mason economist Alex Tabarrok estimate that raising terror alert levels increased police presence by 50%. It is not clear if that meant less terrorism, but more police decreased auto-theft by 43% and burglary by 15%.
More police reduce crime, while longer sentences don’t, because of how criminals perceive risk. Most people commit crime thinking they won’t get caught. Seeing more police changes this and increases the perceived cost of committing crime.
A notorious example of this is the dramatic crime drop that occurred in New York City in the 1990s. There are many reasons crime fell, including a stronger economy that offered better, legal alternatives to crime. But a major factor was that the New York police force grew by 35%. There is little evidence that arresting people for misdemeanors prevented them from committing more serious crimes later, which was the philosophy at the time. But crime did fall because arrests increased for all kinds of crime.
The rest of America also experienced a big drop in crime during mass incarceration. But this was for myriad reasons: Crime fell because of a booming economy, changing trends in drug and alcohol use, an aging population, and other more effective police methods. It’s possible that crime would have been even lower if we hadn’t put so many people in prison, and turned a generation of young men into hardened criminals.
Correction: This post originally stated that 5% of the US population is in prison; 5% of the US population will be imprisoned at some point.