# Short-term effects of cannabis legalisation in Germany on driving under the influence of cannabis: a difference-in-differences analysis using Austria as a control

## TL;DR

This study found no significant short-term increase in cannabis use or driving under the influence of cannabis in Germany after legalisation, compared to Austria.

## Contribution

The study provides empirical evidence on the short-term effects of cannabis legalisation on driving under the influence using a difference-in-differences approach.

## Key findings

- Cannabis use increased slightly in Germany, but not significantly compared to Austria.
- Driving under the influence of cannabis decreased slightly among frequent users in Germany.
- Driving under the influence of cannabis combined with other drugs was most common among weekly users.

## Abstract

In April 2024, Germany legalised adult cannabis possession and cultivation, and in August 2024 established legal THC-limits for driving. This study aimed to examine short-term impacts on (1) cannabis use and (2) driving under the influence of cannabis (DUIC), and (3) investigates the extent of DUIC involving cannabis combined with alcohol or other drugs (DUIC(+)).

Data came from two cross-sectional population surveys in Germany and Austria (control) before (t0: Nov–Dec 2023) and after legalisation (t1: Nov 2024–Jan 2025). We assessed 12-month cannabis use among adults aged 18–64 (Germany: nt0 = 6670, nt1 = 9692; Austria: nt0 = 2132, nt1 = 2102) and DUIC among at least monthly cannabis users (excluding medical use; Germany: nt0 = 393, nt1 = 589; Austria: nt0 = 86, nt1 = 92) using a difference-in-differences (DiD) approach. For t1, we compared the proportion of DUIC(+) and cannabis-only DUIC(−) episodes among all DUIC episodes by use frequency.

In Germany, cannabis use rose from 12·1% to 14·4%, but this trend did not significantly differ from Austria (DiD-effect: OR = 1·18, 95% CI 0·95–1·48, p = 0·141, weighted). Among at least monthly users, DUIC decreased slightly from 28·5% to 26·8% (unweighted), with no significant difference compared with Austria (DiD-effect: aOR = 0·68, 95% CI 0·27–1·68, p = 0·408). Results held across sensitivity analyses including additional confounders and negative controls. At t1, DUIC(+) accounted for 21·5% of episodes. DUIC(−) was most common among daily users, DUIC(+) among weekly users.

Eight months after legalisation, no significant short-term effects on cannabis use or DUIC were observed. DUIC(+), associated with higher traffic risk, was most common among weekly users. A comprehensive evaluation of the cannabis reform requires further monitoring of DUIC and traffic data.

Federal Highway and Transport Research Institute (FE 82.0816/2023).