Blogged on to post on the WaPo story on NY, incarceration, and crime rates and saw that Kim had stolen my thunder. I agree that it's a succinct and thorough summary of what should be conventional wisdom as anyone fashions informed corrections sentencing policy. There are a couple of points that need to be added, though, even regarding Professor Blumstein's reference to the 25% reduction in crime associated with prisons in the 90s.
That 25% is a good rule of thumb, but other studies have the impact between high single digits and around 40%. The 25% was obtained by two different scholars employing two different methodologies. It's probably as good a number as we'll get. But please keep in mind a couple of provisos whenever you hear it cited. One, much of that 25% might have been obtained with other methods (as the overall theme of the article notes) that would have been more cost-effective, that is, freeing resources for other areas of criminal justice that have been starved as we funnel our funds into prisons. That means that those other methods could have been, more importantly, preventative, that is, without having to have a victim to start with.
The other thing to remember is that these estimates of "crime reduced by prison" generally refer to the crimes the offenders would have committed had they been out but weren't because the offenders were behind bars. But if prisons are "crimogenic" for some offenders, then we don't know how many more crimes they commit once they are released that they wouldn't have committed had we found another punishment and different supervision for them, especially since most offenders age out of criminal behavior to some extent as they reach the ages at which these guys have served their terms. I'm not aware of any good studies to research this question directly(which would be extraordinarily hard to design simply by their nature), but if someone else has seen anything, please let us know. In the meantime, studies of matched juveniles who get tracked for later criminality after either probation or prison (Petersilia's probation work and the Sampson/Laub life-course studies) indicate that putting people in prison generally gets you more criminals later. (The Washington State Institute for Public Policy, of course, also has some work indicating this as well.)
The WaPo story is important because it reveals the "justice premium" that I've mentioned here. We don't have to lock up all the people we do to have an impact on crime. I personally believe that most crime reduction comes because of cultural changes and waves and that government is only effective on the margins anyway, but talk about a difficult study to design. We lock people up more for reasons having to do with us and enforcing norms regardless of cost than for why we say we're doing it. Advocates on both sides of the "prisons" issue need to acknowledge and address that before we can ever really hope to get the spiraling expenditures under control before they absorb anything else we can do in criminal justice.
That's what I would have said if Kim hadn't beaten me to the post. Hope your Thanksgiving was good. Be back Monday.