'Suspending your licence to drink': insights from an innovative criminal justice program
There is a growing debate in the health services and criminal justice literatures about programs that combine frequent testing with swift, certain, and fair sanctions to reduce heavy substance use among criminal justice populations. South Dakota has undertaken a radical policy experiment to reduce alcohol consumption among individuals whose alcohol use leads them to repeatedly threaten public health and public safety. South Dakota's 24/7 Sobriety Program (hereafter, 24/7) requires that alcohol-involved arrestees submit to breathalyzer tests twice per day or wear a continuous alcohol monitoring device. Those testing positive or missing a test face an immediate, but brief, jail term (typically a night or two). In essense, their license to drink alcohol is suspended.
Since 2005, approximately 30,000 unique South Dakotans in 24/7 accumulated more than 5 million days without a detected alcohol violation. This is remarkable coverage for a state with approximately 650,000 adults. Perhaps more impressive is that the proportion of breath tests ordered and passed exceeds 99%, and this includes failures and no-shows in the denominator. A 2018 report on alcohol-impaired driving from the National Academies of Science, Engineering, and Medicine reviewed the evidence on 24/7, concluding that the program has been shown to be effective in some rural areas; however, the authors argue that an important limitation to the existing literature is the “reliance on aggregate analyses rather than individual-level data.” This paper seeks to fill this gap.
Our analysis is based on 16,513 individuals who were arrested for a second or third drunk driving offense in South Dakota from 2004 to 2011. We obtained the complete criminal history information for these individuals (including probation revocations) and determined whether they participated in 24/7 based on the program’s administrative records. To estimate the causal effect of 24/7 on the probability of being arrested or having probation revoked, we used program availability in a county as an instrument for individual participation. The results show that 24/7 participation had a large effect on criminal behavior: We estimate that the probability a 24/7 participant was rearrested or had probation revoked 12 months after being arrested for driving under the influence (DUI) was 13.7 percentage points (49 percent; p = 0.002) lower than that of non-participants. We also detected reductions at 24 and 36 months—13.8 percentage points (35 percent; p = 0.013) and 11.7 percentage points (26 percent; p = 0.009), respectively. These findings are robust to a number of alternative assumptions and specifications.