If we want to get our disgraceful incarceration rate back to our own historical level - let alone the lower levels enjoyed by other economically advanced democracies - we have to reduce the prison-plus-jail headcount by about 80%. You read that right: line up five prisoners, and let four of them out.

There are innocent people in prison, and guilty people who didn’t do anything seriously wrong and who wouldn’t threaten public safety if released. But you can’t get from where we are to where we need to be just by letting those people go. More than half of today’s prison inmates are serving time for crimes of violence. So if we’re not content with mass incarceration - as we shouldn’t be - then we have to release some seriously guilty people.

The good news is that you don’t need to lock someone up to control that person’s behavior. We’ve learned that from swift-certain fair community corrections programs such as HOPE in Honolulu, Sobriety 24/7 in South Dakota, and the Swift-and-Certain program now managing 17,000 probationers and parolees in the State of Washington.

The logical next step is to apply the same idea to people now serving prison time. Since prison is expensive, that means you can afford what would otherwise look like expensive interventions, including supported work and supported housing, without breaking the budget.  And the “graduated re-entry” approach solves the hardest problem of all: managing the transition back from prison to the community by making it a slow process rather than a discrete leap from confinement to freedom and from being fed, clothed, and housed at public expense to being on your own.

This VOX essay explores some of the options. We can’t claim now to know what will work. But it should be obvious to everyone that business as usual is not an acceptable choice.

Update More thoughts from Ed Kilgore and BooMan.

Second update Alex Tabarrok at Marginal Revolution thinks this  is a return to the halfway-house idea and worries about excessive supervision.

Halfway houses were a good idea, but problems with staffing, siting (NIMBY), and management mean that they can’t possibly be built on anything like the relevant scale. Moreover, living in a closed facility - while obviously better than living on a cellblock - isn’t very good preparation for living free.

As to over-supervision, recall the first rule of policy analysis: “Compared to What?” This is about managing people who would otherwise be in prison.