There are, however, significant impediments to pursuing such an approach. For one thing, there’s no shortage of groups that benefit from the status quo. “Yes, you have privately operated prisons,” said Glenn Loury, a professor of social science and economics at Brown University. “But you also have corrections-officers unions … Self-interested behavior in the system is not limited to profit.”
“We have perverse incentives in what we call the criminal-justice system,” Levin added. He pointed to police officers, rewarded more for making arrests to solve the crimes that have already occurred, than to prevent those crimes from taking place, or to prosecutors rewarded for securing convictions, and not for reintegrating offenders into society.
Then there are the steep costs of rehabilitation programs, where a day of treatment can be more expensive than a day of prison. “The key is getting way from the obsession with the duration and focusing on the quality of the time,” Levin argued. Perhaps offering treatment for substance abuse or mental health problems costs more per day, “but overall, in the long-term, the person would be kept there for a much shorter period and so ultimately you would be saving money.”
Skeptics of data-driven approaches also point to cases in which models produce disparate outcomes, yielding harsher sentences for members of ethnic or racial minority groups. Levin acknowledged the concern, but argued that “tweaks and adjustments to ensure they don’t have a disparate impact” could solve the challenges of actuarial modeling.
The proliferation of data gathering and of local reform efforts have, Levin argued, started to acquire a momentum of their own. “It seems like there’s just a culture at the state level,” he said, as governments commission studies and issue reports. “When you build a budget around controlling the growth of a prison population, even though there’s people who attempt to derail it, almost always it ends up going through.”
Despite that optimism, shifting to an outcomes-based approach, though, still faces one key hurdle. It will—invariably—lead to the release of some individuals who reoffend. When that happens, there’s often a political backlash, as voters seek an infallible approach.
“We should tell the public: Nothing’s perfect,” Loury said. “We’re going to do better here than we would do if there weren’t this intervention.” He argued that leveling with the public could help change expectations, creating the political space for experimentation and reform. Public officials, Loury said, need to tell the public that no matter how effective a particular change, it’s “going to leave us with some risk; there is no zero-risk environment.”
But not all of those who have tried that in practice are convinced it’ll prove politically viable. “I would love to agree with you,” said Roberto Villasenor, a retired chief of the Tuscon police department. “Unfortunately my experience has been that that just doesn’t happen and doesn’t work.”