Government & Politics

How to Fix Overcrowded Jails

Solving the problem means putting convicted felons on the street. A scholar explains why that’s the right move.

March 18, 2016

| by Lee Simmons


A crowd of men behind bars

Illustration by Stefani Billings

The United States locks up more of its citizens than any other country — more than Cuba, Russia, or Iran, as a percentage of population, and far more than European nations. Since the 1970s, incarceration rates have soared, and many of our prisons and jails are now dangerously overcrowded. There’s a growing bipartisan consensus that change is needed; we simply have no place to put new prisoners. But what’s the solution?

At the local level, one way to relieve the pressure on jails is pretrial release, leaving defendants free until they’re convicted. It’s an intuitively appealing answer, since over 60% of jail inmates nationwide have not yet been convicted of any crime, and it costs taxpayers $17 billion a year to keep them behind bars. But there’s always a risk they might use their liberty to commit more crimes or flee, escaping justice altogether.


Research from Stanford Graduate School of Business suggests a better way: split sentencing, an arrangement whereby low-level felony sentences are divided between jail time and community supervision. “It turns out that this gives you the best trade-off between jail population and public safety,” says management science professor Lawrence Wein. In fact, compared with status quo policies, split sentencing not only alleviates jail congestion, it actually reduces recidivism.

An Operations Management Approach

In 2011 the U.S. Supreme Court ruled that conditions in California’s teeming prisons constituted cruel and unusual punishment and ordered the state to reduce its inmate population by 25%. So it did — in part by shuffling low-level felons (those convicted of nonviolent crimes) into county jails. Now the county jails are overwhelmed, and those inmates are deprived of the rehabilitation programs offered in prison. This maneuver, known as “realignment,” met the state’s legal obligation, but only by relocating the problem.

To find a real solution, Wein and Mericcan Usta , a recent Stanford PhD who is now a research scientist at GroupM in New York City, built a mathematical model of the Los Angeles County jail system — the country’s largest and one of its most troubled. Borrowing a framework from operations management, the model simulates the flow of inmates through the judicial process, from arrest through sentencing, custody, and eventual discharge. Then, using data on recidivism and flight risk, the researchers were able to test alternative policies.

“There’s always a trade-off,” Wein says. “If the only objective was to minimize crime, you’d lock up every offender until they were no longer a threat. But the cost would be immense. So the real question is, what policy gives you the best trade-off between the goals of protecting society and limiting inmate numbers?”

The two main tools courts have to reduce overcrowding are pretrial release and split sentencing. Until recently, the latter was rarely used in Los Angeles (or anywhere else), and some judges have resisted it. It’s easy to understand why: Split sentencing means putting convicted criminals on the street, while pretrial release frees defendants who, after all, may be found not guilty.

Wein and Usta examined different ways of implementing these policies: for instance, limiting pretrial release to misdemeanor cases or adopting split sentencing only for low-risk felons — those deemed unlikely to re-offend, using a risk-assessment tool commonly employed by correctional agencies. For every such permutation, they ran 1,000 simulations, each replicating a time span of more than five years.

A Counterintuitive Conclusion

The results are surprising: Split sentencing can accomplish any targeted reduction in jail population with a smaller increase in crime than would result from a policy of pretrial release alone, even though it means freeing convicted felons. What’s more, the optimal trade-off comes from extending split sentencing to all low-level felons, including those categorized as high risk.


There’s always a trade-off. If the only objective was to minimize crime, you’d lock up every offender until they were no longer a threat.
Lawrence Wein

“It’s counterintuitive, but there are two things driving this,” Wein says. “First, convicted felons now make up 45% of all the inmates in LA jails, and most of those are high-risk — which is mainly to say they’re young, since that’s the key determinant in risk assessments. So to really reduce crowding, you have to go where the numbers are.”

The other factor is more subtle. Because of the pressure to turn over jail beds, the wheels of justice move faster for defendants in custody than for those on pretrial release. In Los Angeles, someone detained on a low-level felony charge will have their case decided in 53 days, on average, from the time of arraignment; for a defendant out on bail, it takes 191 days. Those in detention are also more likely to accept a plea deal.

That means pretrial release saves only 53 days of jail time per inmate while adding 191 days of recidivism risk. In other words, it exposes the public to 3.6 days of increased crime risk for every day of jail time saved. For split sentencing, the risk ratio is 1:1 — every day of increased crime risk is one day off the jail rolls — a much more favorable trade-off, in terms of both public safety and cost to taxpayers.

“We’re not saying split sentencing should be used instead of pretrial release,” Wein adds. “But it’s the key lever. You can’t make a big dent in jail congestion with pretrial release alone, and certainly not at an acceptable level of risk.” Indeed, California’s legislature endorsed split sentencing in January 2015, and its use has rapidly increased since then — with negligible increase in crime to date.

In fact, compared with the status quo policy in Los Angeles in 2014, the model simulations show that a full implementation of split sentencing would not only reduce jail population by 20%, it would actually cut recidivism by 7% — no trade-off required.

Wein stresses that this is all contingent on having strong systems in place for outside supervision and support, including drug treatment. “California has actually done a good job on this, and one benefit of realignment is that by handling it at the local level, you can keep a person under supervision in their home community, which really helps. That’s probably why there hasn’t been an uptick in crime.”

While the study focuses on the Los Angeles jail system, the lessons are applicable anywhere. “Jails in other states might not have as many felons, so the magnitude of the benefits from split sentencing would differ,” Wein says. “But the same forces are in play. This is something courts should definitely be looking at.”

Lawrence Wein is the Jeffrey S. Skoll Professor of Management Science at Stanford Graduate School of Business.

For media inquiries, visit the Newsroom.

Explore More