A Better Way to Reform Prisons
Prison systems in the U.S. are bloated, inefficient, and wasteful. We can get a much better return on the estimated $40 billion we invest annually in prisons by instituting reforms based on what’s been shown to work. In a recent American Enterprise Institute report, I lay out a plan for evidence-based prison reform that 1) significantly expands the delivery of effective programming, 2) further reduces our reliance and spending on prison, and 3) places greater emphasis on the use of validated risk assessments to help prison systems make better programming and downsizing decisions.
The nation’s imprisonment rate has fallen by more than 10 percent over the last decade. Because we’ve overused prison, this decline is a step in the right direction. But prison downsizing alone will not improve public safety unless it’s accompanied by an increase in effective programming resources for prisoners, probationers, and parolees.
Decades of research have shown there are effective interventions that reduce recidivism by targeting known risk factors for reoffending. Examples include substance abuse treatment, cognitive-behavioral therapy, sex offender treatment, and education and employment programs. Research has also shown, however, that five-year rearrest rates for U.S. prisoners are near 80 percent.
One reason why recidivism rates are so high is that many prisoners are idle in prison, due to their own choice, overcrowding, or a lack of programming resources. When prisoners are “warehoused”, it diminishes their chances for success in landing a job and desisting from crime after they get released. A recent study found that warehousing increased the likelihood of recidivism by 13 percent.
How can we deliver more programming that’s been proven to be effective without increasing the costs? After all, programming costs money, and a lack of money in state budgets is a big reason why a number of states have recently downsized their prison populations. Further reducing the use of prison is necessary to not only lower the costs, but also to free up the physical space needed within prisons to provide more programming. Prison populations can be reduced by 1) decreasing the number of persons (re)entering prison and 2) shortening the lengths of stay for those admitted to prison.
When individuals enter prison, it should be long enough to participate in effective programming, which usually lasts between 3 and 9 months. The best way to safely reduce prison admissions would be to restrict probation and parole violators (about two-thirds of all prison admissions) to the more serious offenders who are, as it is, more likely to get longer revocations. The less serious violators, who are more likely to get warehoused due to their relatively brief stays in prison, should remain in the community. We can achieve better public safety outcomes by reallocating the decarceration “savings” to provide more programming resources for all probation and parole violators—those revoked to prison as well as those who stay in the community.
But if it’s necessary to extend the minimum length of stay in prison to at least five months for rehabilitation purposes, the same holds true for limiting how long most inmates should be imprisoned. Because inmates with longer sentences tend to be warehoused for much of their imprisonment, the average sentence length (5 years) is ample time to participate in multiple effective interventions. Shortening confinement periods for more inmates with longer sentences would generate decarceration “savings” that, once again, should be reinvested to ramp up the delivery of prison programming.
Downsizing the prison population to increase programming would require correctional systems to make more decisions relating to program placement and recidivism risk. Improving the quality of these decisions would require more extensive use of validated risk assessment instruments. Concerns have been raised about whether algorithms and “big data” are being used to worsen racial and ethnic disparities. What critics have failed to point out, however, is that the alternative would be reliance on professional judgment, which doesn’t perform very well in predicting future behavior—and recidivism is no exception. To be sure, the design and use of actuarial risk assessments can, and should, improve. But it’s also important to recognize that research has long shown that statistical prediction is the superior approach.
Implementing evidence-based prison reform would require a shift from punishment to rehabilitation in both our ideology and practice. Make no mistake, this would be no small feat. One enduring school of thought has been that if we make prison so horrible, it will motivate inmates to desist from crime. Increasing the misery of the prison experience may satisfy the impulse for retribution, but it doesn’t lead to an efficient use of taxpayer dollars. Indeed, the evidence has long shown that punitive strategies are costly and, ultimately, ineffective in promoting desistance from crime. Instead, if we want prisons to be leaner, more cost-effective, and more successful in reducing recidivism, we need reform based on what’s been shown to work.
Grant Duwe is an Academic Adviser to the American Enterprise Institute (AEI) for Criminal Justice Reform and Research Director for the Minnesota Department of Corrections.