Automated Enforcement and Traffic Safety

Traffic safety poses a persistent challenge for society and public policy. Conventional law enforcement by human police is often cost-ineffective because of information asymmetry and negative externalities of unsafe driving behaviors. Automated enforcement, in the form of traffic cameras on the road, has gained prominence in recent decades, yet its effectiveness and underlying mechanisms remain debated. This study examines the impact of traffic cameras on road safety using longitudinal data from a metropolitan city in China. We distinguish between advanced cameras, which use machine learning to detect various traffic violations and constantly record video, and conventional cameras, which rely on triggered image capture for a limited number of violations. Using an event study design with staggered camera installations at road intersections, we observe a significant and sustained reduction in accidents near advanced cameras, compared with locations with no cameras or only conventional cameras. Further analysis identifies three key mechanisms driving the effects of advanced cameras: (i) automated detection effect—superior technical capabilities to automate violation detection; (ii) real-time recording effect—continuous monitoring and recording capability to augment accident cause identification; and (iii) driver learning effect—technology-enabled deterrence that increases driver awareness of these cameras and encourages behavioral adjustments to mitigate accident risks. This study contributes to information systems, transportation economics, and criminology, offering policy insights into the effective design and deployment of automated enforcement to improve traffic safety.

This paper was accepted by D. J. Wu, information systems.

Funding: Z. Cheng acknowledges the support of the Staff Research Fund at the London School of Economics and Political Science. Z. Dong acknowledges financial support from the National Natural Science Foundation of China (NSFC) Young Scientists Fund [Grant 72202247] and the Guangdong Basic and Applied Basic Research Foundation [Grant 2024A1515012806].

Supplemental Material: The online appendices and data files are available at https://doi.org/10.1287/mnsc.2023.00575.