Alert eyes and DWIs: an indirect evaluation of a DWI witness reward program in Stockton, CA

We evaluate a "grassroots" anonymous reward program targeting drunken driving in Stockton, CA. The time-series cross-sectional data covers 19 years for Stockton and six other California cities. Exploiting interrupted time-series regression, Zellner's seemingly unrelated regression (SUR) framework, and bootstrapped standard errors, we test for an impact of this program on alcohol-related injury or fatality accidents, the proportion of all accidents involving alcohol, and the number of DWI arrests. In its first decade, the citizen reward program appears to have averted some 275 alcohol-related accidents for social cost savings of between $21,000 and $5.6 million. Further, possibly 4495 arrests were precluded, saving some $1-3 million in arrest-related costs. Incentivized public monitoring of driving-after-drinking may be an effective drunken driving abatement program though our exploratory findings need further confirmation.