The idea is to let the stoner (or, preferably, former stoner) out early, but keep the murderer in prison. That makes perfect sense, right?
Perhaps not. I have a friend who works for the Safer Foundation, a non-profit in Chicago that helps prisoners reenter society. He told me that the group tends to have success with people who have been convicted of violent crimes. They're often able to hold stable jobs and stay out of trouble -- provided, of course, that someone will hire them.
This isn't just a matter of anecdote. The U.S. Bureau of Justice Statistics tracked (moderate-sized PDF) whether 272,111 convicts released in 1994 were arrested again within the next three years. The report, released in 2002, found the recidivism rate for violent criminals was slightly lower than for non-violent criminals. Overall, 67.5% of those released were arrested again within three years. Among violent offenders, however, the rate was only 61.7%.
This was, of course, just one study (although at the time BJS described it as "the largest recidivism study ever conducted in the United States"). But, there are some good reasons to think the BJS report got it right.
First of all, violent offenders aren't always the most serious, hardened criminals. Someone who gets convicted of assault for getting in a bar fight easily could be a far less dangerous criminal than someone who deals drugs for a living.
Another possible explanation: My friend at the Safer Foundation says they have the most success with older clients. That makes sense -- young people commit most crimes. Serious violent criminals spend more time in prison than anyone else, so they tend to be older when they get out.
Perhaps, many of them have reached a point in their lives when they've mellowed. The BJS report showed that more than 70% of released prisoners in their 20s were arrested again, but only 45% of those older than 45 were rearrested.
I'm not suggesting, of course, that we let all of the violent criminals run free. You'd need to know the answer to some additional questions to know just who should be let out of prison early and who shouldn't.
The real question isn't just whether violent or non-violent offenders are more likely to commit crimes once they go free. Rather, it's also what effect letting them out a little bit early would have. Do shorter sentences result in higher recidivism rates? For all types of convicts or just some?
Even more importantly, to design a smart policy you'd have to think not only about who is likely to commit crimes once they're out, but also what sorts of crimes they're most likely to commit. There's a big difference between someone who gets arrested anew for a minor probation violation and for a serious felony. According to BJS, violent offenders when released were 30% more likely than non-violent offenders to be rearrested for a violent offense.
If I have a point, it's this: To do a good job determining just who should be let out of prison and when, you have to weigh a lot of different factors. It's complicated work that demands well-informed empirical analysis.
Excluding all violent offenders from early release programs is politically safe. But, a more nuanced, thoughtful approach might save money on corrections, while still keeping citizens safe.
