Can haze warning policy reduce traffic accidents: evidence from China

Haze pollution may decrease drivers' driving performance by worsening their physical and psychological states. This paper explores the effects of haze warning policy on traffic accidents for the first time. We use the daily-city traffic accident data from 2016 to 2019 in China and construct Regression Discontinuity (RD) strategies based on the warning signal thresholds of PM2.5 concentration for estimations. The results show that one yellow warning and one orange warning can reduce traffic accidents by 8.8% and 13.1% on that day respectively, while the red warning does not work significantly possibly due to the self-perceived channel rather than the warning-signal channel. We also find that the effects may vary among different groups of drivers, vehicles, and roads. Our results prove that the haze warning policy is a non-negligible tool to reduce traffic accidents, which is useful to policy-making both related to haze pollution regulation and transportation safety.