Association for Public Policy Analysis & Management

Greg Midgette (Presenter), University of Maryland

Much of the criminal deterrence literature focuses on changes in policing and sanction severity, but these neither guarantee certainty nor celerity of punishment for a violation. This paper presents an individual-level analysis of a large-scale effort to dramatically increase both certainty and celerity of sanction. South Dakota’s 24/7 Sobriety Program (hereinafter 24/7) requires that alcohol-involved offenders abstain from alcohol and be tested for alcohol multiple times per day. Those failing or missing a test are subject to a swift, certain, and moderate sanction, typically a night or two in jail. Using criminal-history information for 20,243 individuals arrested for a second or third offense for driving under the influence of alcohol, this paper estimates the effect of 24/7 participation on re-arrest for any offense or having probation revoked. Exploiting variation in timing of county adoption in an instrumental variables bivariate probit model, we estimate that relative to non-24/7 participants, 24/7 reduces arrests and revocations by 13.7 percentage points (49 percent) 12 months after DUI arrest. We also detected reductions at 24 and 36 months—13.8 percentage points (35 percent) and 11.7 percentage points (26 percent), respectively. The implications of these results extend beyond reducing heavy alcohol use and alcohol-related crime; they provide evidence that it is possible to create a credible deterrent threat on a large scale by prioritizing both certainty and celerity of sanction.