“Truth in Sentencing, Incentives and Recidivism”

The title of this post is the title of this notable recent empirical paper authored by David Macdonald available via SSRN. (Hat tip to this recent episode of the podcast Probable Causation for highlighting the paper and bringing it to my attention.)   Here is its abstract:

Truth in Sentencing laws eliminate discretion in prison release.  This decreases the incentive for rehabilitative effort among prisoners.  I use a regression discontinuity design to exploit a change in these incentives created by the introduction of TIS in Arizona.  Before prison, I find that sentences were reduced by 20% for TIS offenders.  Further, I find that rule infractions increased by 22% to 55% and education enrolment fell by 24%.  After release, I find offenders were 4.8 p.p. more likely to reoffend.  I further find that recidivism and infractions effects are largest among drug and violent offenders.  Finally, I show that the reduction in sentences resulted in a broad equalization of time served at the cutoff, which indicates that the removal of early-release incentives by TIS was the main mechanism driving results.