When Germany’s Cannabis Act took effect on 1 April 2024, the loudest objection was easy to state: if adults can legally hold and grow cannabis, more people will use it. Two and a half years on, the most detailed data series yet built to test that claim has been published. It draws on 23 survey waves and almost 36,000 respondents, spread evenly across the two years before the reform and the two years after. The headline is short: in these numbers, there is no sign of an increase.
The study, led by Daniel Kotz of Heinrich Heine University Düsseldorf and Jakob Manthey of University Medical Center Hamburg-Eppendorf, appeared in September 2026 in the International Journal of Drug Policy (DOI: 10.1016/j.drugpo.2026.105501). It is a short report, a compact analysis with a deliberately narrow question, and it sits inside the government-funded evaluation of the Consumer Cannabis Act, in which several of the authors take part.
Key Takeaways
Between April 2022 and April 2026, 6.4 percent of 35,957 respondents aged 14 to 64 said they had used cannabis in the past twelve months. The trend in that figure stayed essentially flat on both sides of 1 April 2024. Neither the population as a whole nor men, women, 14- to 24-year-olds or 25- to 64-year-olds showed a statistically meaningful change in trajectory after the reform, and the share of users consuming daily or almost daily did not move either. The authors are equally clear about what the data cannot show: the design has no control group, effects of legalisation elsewhere have taken years to build, and household surveys miss the people most at risk. Their recommendation is not reassurance but continued monitoring across multiple data sources.
Where the Data Come From
The source is DEBRA, the German Study on Tobacco Use, a rolling national survey that has run since 2016. It tracks tobacco and nicotine use and, since April 2022, cannabis. Every wave, roughly 1,500 people are interviewed face to face at home using a computer-assisted questionnaire. About 40 percent of respondents are recruited through random stratified sampling; the rest come through quota sampling so that the sample matches the German population on age, gender, household size, education and region. The analysis weighted the data on those same variables.
The researchers used waves 36 to 58, running from April/May 2022 to March/April 2026. Twelve waves fall before the Cannabis Act came into force, eleven after. Because the survey stopped interviewing people over 64 in January 2025, the whole analysis is limited to the 14-to-64 age range.
The survey question was plain: have you ever used cannabis? Anyone who answered that they had used it daily or almost daily, at least weekly, at least monthly, or less than monthly in the past twelve months counted as a past-12-month user. Within that group, daily or almost daily use served as the marker of heavy consumption.
One detail matters for how much weight the findings deserve. The analysis plan was published on the Open Science Framework before the data were examined (osf.io/w92v5), along with the statistical code. Pre-registration is a public commitment that stops researchers from adjusting their method until a preferred result emerges. The raw data themselves are not open.
How You Look for a Bend in a Trend
Picture the prevalence line as a road. The question is whether it changes direction at a particular junction, here the first full survey wave after 1 April 2024. The tool for that is a piecewise logistic regression: the line is allowed one slope before the junction and another after it, and the test asks whether the two slopes differ. A random effect for each wave absorbs the extra noise that comes from small per-wave samples.
The team chose not to test for a sudden jump on the date itself. Their reasoning is that a law rarely shifts the behaviour of a whole population overnight, and that a jump test is prone to mistaking random noise for a real step. As a check, they ran placebo tests that moved the assumed break point two waves earlier and two waves later. If a false date produces a bend as convincing as the real one, the real one is suspect.
The same analysis was run for the total sample, for men and women separately, for 14- to 24-year-olds and 25- to 64-year-olds, and for heavy use within the group of past-12-month users. That last subgroup, 2,302 people, was too small to be split further by age or gender.
The Results in Detail
Of 35,957 respondents, 2,302, or 6.4 percent, reported past-12-month cannabis use, roughly 100 users per wave. Among them were 1,644 men and 658 women; 750 were aged 14 to 24 and 1,549 were aged 25 to 64. Within the user group, 270 people, or 11.7 percent, reported daily or almost daily use.
The numbers that answer the research question are the coefficients for the change in slope after the reform. None reached statistical significance; every p-value was 0.15 or higher. In the total population, the estimated change in slope was 0.001 with a p-value of 0.978, which is as close to zero as survey data get. For men the figure was 0.020 (p = 0.633), for women minus 0.039 (p = 0.462). For 14- to 24-year-olds it was 0.086 (p = 0.234), for 25- to 64-year-olds minus 0.034 (p = 0.301). For heavy use within the user group, the change was 0.125 (p = 0.154).
Trends before the reform were also flat almost everywhere. The single exception was the share of heavy users among past-12-month users, which was drifting downwards before April 2024 (coefficient minus 0.091, p = 0.04). The placebo tests confirmed the overall picture. A post-hoc model that did allow for a level shift on the reform date changed nothing of substance, apart from nudging the p-value for the pre-reform slope in the 25-to-64 group from 0.071 to 0.028.
The raw curves are not perfectly still. Among men, prevalence rose around the third quarter of 2023, fell back, and rose again from mid-2024. The 14-to-24 group showed the same shape more sharply. Heavy use fluctuated more than any other measure and drifted slightly down over the period. But with around 100 users per wave, swings of that size are what random variation looks like, and none of them lines up cleanly with the reform date.
How Strong Is This Evidence?
The authors list their limitations with unusual directness, and they deserve to be read as carefully as the results. First and most important: this is an observational study with no control group. There is no second Germany without a Cannabis Act to compare against. A flat line therefore does not prove the law had no effect; it shows that no effect stands out from the other forces that move cannabis use anyway.
Second, the Consumer Cannabis Act and the Medical Cannabis Act took effect on the same day, and their effects cannot be separated. The medical law opened access to cannabis flower on prescription via telemedicine for almost any indication, and Germany has since become Europe’s largest commercial cannabis market, with around 200 tonnes of medical cannabis available in 2025. Some of that consumption may appear in the survey as use, and the data cannot tell which.
Third, household surveys under-represent marginalised groups, precisely the people most likely to develop problems with cannabis. Fourth, absolute prevalence is probably underestimated, because not everyone tells an interviewer in their living room the truth about drug use. Fifth, legalisation might have made people more willing to admit use. That would have produced an artificial rise, not hidden one, so it does not fit the pattern observed.
Sixth, the twelve-month question itself blurs the boundary. Someone interviewed in the first year after the reform may have been describing use that took place before 1 April 2024. The authors judge this a minor problem: around 80 percent of users in their sample had also used cannabis in the past 30 days, and the second year of follow-up lies entirely after the reform.
On the strengths side stands the density of the data. Twenty-three measurement points give a far finer picture than the usual national survey conducted every three years. The authors calculate that just ten waves of 1,500 people would detect a rise from 5 to 7 percent with more than 90 percent power. An increase of that magnitude would not have gone unnoticed.
What Other German Data Show
The finding matches what other sources have reported. The Epidemiological Survey of Substance Abuse, run every three years, documented a steady climb in past-12-month prevalence among 18- to 59-year-olds from 5.1 percent in 2012 to 10.0 percent in 2021, according to an analysis by Olderbak and colleagues (DOI: 10.1111/add.16356), all of it under full prohibition. For 18- to 64-year-olds, a team led by Eva Hoch reported in Deutsches Ärzteblatt International (DOI: 10.3238/arztebl.m2025.0161) a prevalence of 8.8 percent in 2021 and 9.8 percent in 2024, a difference that was not statistically significant. The rise was confined to men (10.7 to 12.3 percent); among women it was essentially flat (6.8 to 7.1 percent).
Young adults tell a similar story. Orth and colleagues, writing in the Journal of Health Monitoring (DOI: 10.25646/14216), found that among 18- to 25-year-old men prevalence rose from 26.9 percent in 2023 to 31.6 percent in 2025, while among women it dipped from 19.4 to 18.8 percent, both continuing a trend that dates back to 2008. Heavy use has been remarkably stable for decades: about 10 percent of users since 1995, with an outlier of 15.7 percent in 2021, and no significant difference between 2021 (17.3 percent) and 2024 (16.2 percent).
Two other lines of evidence point the same way. A difference-in-differences analysis using Austria as a control (DOI: 10.1016/j.lanepe.2026.101593) found no short-term rise in driving under the influence of cannabis, and wastewater analyses from 15 German cities, reported in the second interim report of the EKOCAN evaluation (DOI: 10.25592/uhhfdm.18530), show no jump after the reform. For context on how German figures sit within the wider picture, see our breakdown of the latest UNODC report on global cannabis use.
Why Two Years Is Not the Final Word
Germany’s model is structurally different from Canada or the US states with commercial retail. In those jurisdictions, prevalence has risen modestly after legalisation, as both a systematic review and a review of reviews (DOI: 10.1016/S2468-2667(24)00299-8) have documented. An analysis of US college students (DOI: 10.1111/add.14939) further suggests that effects accumulate, with the largest changes appearing around five years after a reform. Germany has no shops, no advertising and no price competition. Legal access runs through a plant on the balcony or through non-profit cultivation associations, which have only been able to apply for licences since July 2024 and have been slow to get off the ground.
That is why the authors phrase their conclusion cautiously: the legal routes to cannabis have probably not yet fully unfolded, and the increase observed since 2012 may continue at a gentler slope. For the political debate, this cuts both ways. Anyone arguing to roll back the law on the grounds of rising consumption has, so far, no support in this series. Anyone claiming the law has been proven harmless is stretching the data just as far. On the market side, meanwhile, the weight is shifting visibly towards medical cannabis, as our analysis of the BfArM import figures for the first quarter of 2026 shows.
What This Means for the Evaluation
The Consumer Cannabis Act mandates a four-year scientific evaluation, and DEBRA is one of its pillars. The series continues; the next waves will show whether the flat line holds. What the study also makes visible is research infrastructure: a population survey with monthly cadence, a pre-registered analysis plan and open code. Why that kind of foundation matters more for cannabis science than questions of legal scheduling is something we explored in the context of US marijuana rescheduling and the research funding barrier. And what Germany’s model lacks compared with commercial markets, namely any fiscal dimension at all, is laid out in the Yale Budget Lab estimate of federal marijuana tax revenue. For how Germany’s approach to non-intoxicating cannabis products compares with its neighbours, our comparison of CBD regulation across nine European countries fills in the picture.
The honest summary is this: two years after partial legalisation, no measurably larger share of people in Germany use cannabis than before, and the proportion of daily users has not changed. Whether that holds once cultivation associations operate nationwide and the medical market keeps expanding is an open question, and the study frames it as one.
FAQ
Not measurably, according to the DEBRA survey of 35,957 people interviewed between April 2022 and April 2026. Past-12-month prevalence across the whole period was 6.4 percent, and the trend showed no statistically significant change of direction after 1 April 2024 in any age or gender group. Other German surveys reach a similar conclusion while documenting a long-term rise that began well before the reform.
Among 14- to 24-year-olds, the study found no significant change in trend, although the raw curve swings more than in older groups because the subsample is smaller (6,103 people). A separate survey of 18- to 25-year-olds recorded a rise among men and a slight fall among women between 2023 and 2025, both continuing a trend since 2008.
Under the Consumer Cannabis Act, adults aged 18 and over may carry up to 25 grams of dried cannabis in public, keep up to 50 grams at home and grow up to three plants for personal use. Since 1 July 2024, non-profit cultivation associations can apply for licences to grow and distribute cannabis to their members. Commercial sale remains prohibited, as does supply to minors, use near schools and playgrounds, and advertising. Medical cannabis has been available on prescription since 2017 and is now governed by the separate Medical Cannabis Act. Germany’s rules apply only in Germany; possession and use remain criminal offences in most other jurisdictions, and travellers should not assume any equivalence.
No, and the authors say so explicitly. Without a control group, no cause-and-effect claim is possible; the effects of the consumer and medical laws cannot be separated; and experience from commercial markets shows that changes in use can take several years to appear. The study shows that there was no detectable increase in the first two years, and no more than that.
Legal Disclaimer
This article is provided for informational purposes only and reports on published scientific research and the legal situation in Germany. It does not constitute legal advice, medical advice, or a recommendation to acquire, cultivate or consume cannabis. Cannabis remains a controlled substance in most countries, and laws on possession, cultivation and use differ substantially between and within jurisdictions. Readers are responsible for knowing and complying with the law where they live. For questions about your own legal or health situation, consult a qualified professional.
