The Effects of Mandatory Ignition Interlock Devices on DUI Recidivism

• In Event: Criminal Justice Interventions with Cross-Policy Impacts

Abstract

State efforts to combat drunk driving and the associated social harms rely heavily on sanctions for those convicted of driving under the influence (DUI), tighter definitions of DUI through lower blood alcohol content (BAC) thresholds for establishing guilt, and swift and certain responses to drunk driving arrests and convictions. Although many of the sanctions might be characterized as creating a specific deterrent to DUI, there are several sanctions that basically aim to address individual drivers by restricting driving privileges or employing technology to prevent driving while intoxicated. One such sanction is the requirement that arrested and/or convicted drivers install an ignition interlock device (IID) in their vehicles. IIDs are essentially individualized breathalyzers that prevent people who have been drinking from starting or operating a vehicle. Currently, 31 states and the District of Columbia require all drivers convicted of a DUI to install an IID for a period prior to full restoration of driving privileges. The remaining states either require IIDs following repeat-offense convictions, or permit judges in DUI cases the discretion to impose IID orders post conviction.   This study estimates the effect of IID on several measures of DUI recidivism including subsequent arrest, crashes, and crashes with injury.  We exploit the coinciding expiration of a geographically-limited pilot IID program and the introduction of a conviction offense limited statewide requirement for IID installations to identify exogenous variation in IID installations.   The two sources of policy variation create a 25-percentage point relative increase in IID installations for DUI arrests occurring in counties that were not part of the geographically limited pilot relative to counties that were.  We find significant relative decreases in various DUI recidivism measures coinciding with relative increases in IID installation.  Two-stage-least-squares estimates suggest reductions in two-year recidivism rates on the order 17 to 26 percent of the non-IID baseline.

Author

• Steven Raphael, University of California, Berkeley
  Presenting Author