Evaluation of the causal impact of recreational marijuana legalisation on traffic safety in the US
本文采用增强合成控制法，评估美国部分州娱乐用大麻合法化对交通死亡率的因果影响，发现合法化后死亡率持续上升，且与零售可得性相关，驳斥了否认二者关联的观点。
Since the legalisation of recreational marijuana in certain US states, traffic fatalities involving drivers testing positive for marijuana have markedly increased, thereby prompting the need to understand how this policy change affects road safety. While marijuana is well-known to impair driving, determining if its recreational use directly causes more traffic fatalities remains contentious due to challenges in roadside impairment testing. Additional challenges arise because (i) Simulations may not accurately replicate driver impairment and road conditions, (ii) Estimation based on observational data must adjust for (unobserved) confounding factors, requiring an innovative model to generate causal inference, and (iii) The dynamic, evolving nature of the process requires capturing temporal relationships. This paper contributes by employing a rigorous study design based on an augmented synthetic control method to assess the causal impact of recreational marijuana legalisation on traffic fatalities. It identifies a consistent but lagged pattern of increased fatality rates in several states post-legalisation, with the effect primarily linked to the drug's retail availability. These findings disprove any prevailing conjectures that dismiss the link between recreational marijuana use and fatal traffic crashes, highlighting the need for informed policy responses.