The November (ahem) issue of the Journal of Law and Economics just hit my emailbox. Among the articles is Randi Hjalmarsson, Juvenile Jails: A Path to the Straight and Narrow or to Hardened Criminality? Here is the abstract (emphasis added):

Juvenile justice systems throughout the United States have become increasingly punitive since the 1970s. Most states have passed legislation making it easier to transfer juveniles to the criminal courts. Supporters of this "get tough" movement argue, in part, that juvenile courts are ineffective in deterring young offenders. This claim, however, is based primarily on poorly designed evaluations that do not account for the nonrandom nature of sentencing. This paper demonstrates how the institutional features of the justice system can be exploited to identify causality when true random assignment is not feasible. In particular, I capitalize on discontinuities in punishment that arise in Washington State's juvenile sentencing guidelines to identify the effect of incarceration on the postrelease criminal behavior of juveniles. The results indicate that incarcerated individuals have lower propensities to be reconvicted of a crime. This deterrent effect is also observed for older, criminally experienced, and/or violent youths.

Note that "deterrent effect" here is specific deterrence -- the individual punished is less likely to do it again. This is different from a "general deterrent" effect, where punishing one person makes others less likely to do it.