Study Finds No Link Between Cannabis Legalization and Traffic Deaths

Study Design and Scope

Researchers used a difference-in-differences event study design with multiple imputation methods to analyze traffic fatality data across 24 states that legalized recreational cannabis between 2012 and 2024. The study controlled for state-level demographic variables, economic indicators, and temporal trends to isolate the effect of legalization from confounding factors.

The analysis spanned nearly two decades of data, comparing fatality rates in states before and after legalization against control states that maintained prohibition. Imputation-based methodology addressed missing data points and allowed researchers to model counterfactual scenarios.

Key variables included:

• Traffic fatality rates per 100,000 population
• Legalization implementation dates
• State-level alcohol consumption patterns
• Vehicle miles traveled
• Demographic composition shifts

Core Findings

The study found no statistically significant increase in traffic fatality rates following recreational cannabis legalization in any of the 24 states examined. Point estimates varied across states and time periods. But confidence intervals consistently crossed zero, indicating no reliable causal relationship.

This finding held across multiple sensitivity analyses. Researchers tested different lag structures to account for delayed policy effects, alternative control group selections, and varying definitions of the post-legalization period. None produced evidence of increased fatalities.

The null result contradicts earlier studies that suggested short-term increases in some states. The longer observation window and more sophisticated statistical methods likely explain the divergence.

Policy Implications

These findings undercut a primary argument opponents use against legalization in states still debating recreational access. Traffic safety concerns have stalled legalization efforts in Ohio, Pennsylvania, and Minnesota in recent sessions, with lawmakers citing fears of impaired driving epidemics.

Yet the study doesn't address enforcement gaps. Roadside THC testing remains unreliable. Per-se impairment thresholds vary wildly across legal states. Colorado uses 5 ng/mL whole blood; Washington uses the same threshold but applies it differently; California has no numeric limit.

That enforcement inconsistency complicates the safety picture. A lack of measurable fatality increases doesn't mean impaired driving isn't occurring—it may mean detection and attribution remain poor. For full background on this debate, see the CannIntel topic hub on cannabis and traffic safety.

What the Study Doesn't Settle

Researchers examined population-level fatality rates, not individual crash causation or THC-positive driver prevalence. Those metrics tell different stories. THC-positive fatalities have risen in most legal states, but so have THC-positive rates in the general driving population, making causal inference difficult.

The study also didn't parse differences in state regulatory models. States with stricter DUI enforcement, lower potency caps, or delayed retail rollouts may show different patterns than those with immediate access to high-THC concentrates. The aggregate analysis smooths over those variables.

Researchers noted that substitution effects—cannabis replacing alcohol—could mask any cannabis-related risk increase. If legalization reduces drunk driving fatalities while adding cannabis-impaired crashes, the net effect might be neutral or even protective.

Enforcement approaches will likely remain inconsistent. Until roadside testing technology catches up to policy, the traffic safety debate will hinge more on perception than data.