The headline was easy to write. The DEA has released the full transcript of its cannabis rescheduling hearing, 2.533 pages across eleven days, and in it the agency defends its own proposal: cannabis no longer belongs in Schedule I. Chief Administrative Law Judge Derek Julius is preparing his recommendation, the final briefs have been filed since mid August, and the DEA Administrator will make the ultimate call.
What makes the transcript worth reading is not the headline. It is how quickly some of the arguments came apart once anyone asked a follow up question. And it is that the exact points which wobbled in Washington happen to be unresolved in Germany right now.
Four passages that are easy to miss
Ed Wood of DUID Victim Voices testified on cannabis and road safety. Under cross examination he conceded that if consumption rises, researchers should expect to find THC in more drivers regardless of whether cannabis caused any crash. Detection is not impairment. On their own, in his words, that data proved nothing. Shortly afterwards the same witness presented a projection of additional traffic fatalities resulting from rescheduling, and had to acknowledge that there is no good government data supporting it. His estimate instead rested on comparing fatality rates between states with and without a commercial cannabis market.
The same witness produced a second awkward moment. He referred to statements made by an earlier government witness, even though a sequestration order was in place precisely to prevent that. The judge restated that witnesses were not supposed to read or listen to prior testimony. It decides nothing. But in an administrative proceeding where the integrity of the record determines its evidentiary weight, it is an unhelpful passage, all the more so because the DEA later moved to strike that testimony.
The economically heaviest argument from the opposing side came from the National Drug and Alcohol Screening Association: seven to nine million safety sensitive workers would fall out of federal drug testing if cannabis moves to Schedule III. Cross examination turned that into something else. The organisation’s executive director confirmed that employers can still test under their own company policies. And the Department of Transportation has since made clear that a state licensed prescription is no excuse for a positive test for pilots or commercial drivers.
The fourth point is the strangest one. The DEA argued against the DEA. Pharmacologist Luli Akinfiresoye testified that her section had applied the old five part test for currently accepted medical use rather than the newer two part standard the Office of Legal Counsel considers sufficient. This was not her personal view, she said, but a decision taken together with colleagues in her section. Their conclusion was that cannabis fails the test, and she disagreed with the contrary assessment from the Department of Health and Human Services. All while lawyers from the same agency argued the exact opposite in the same room.
Why this is more than a footnote
You can read that as a collection of anecdotes. I read it differently. The proceeding is set up as an administrative hearing, but what it mainly produces is material for what comes next. Several lawsuits against the reclassification are already running, consolidated by a federal appeals court, filed by states, by prohibitionist organisations and by a pharmaceutical company. Their strongest argument is not medical but structural: the April order creates a hybrid status Congress never authorised. Cannabis formally sits in Schedule III while keeping Schedule I and II style controls, meaning quotas, registration requirements and import and export permits.
In that setting, having the agency’s own scientific section on the record applying a different standard and reaching a different conclusion than its own legal team is not a detail. It is a brief the plaintiffs no longer have to write themselves.
Which is why I am careful about the timeline. First the judge’s recommendation, then the Administrator’s decision, a proposed rule before year end, a final rule realistically in 2027, and litigation running over all of it the entire time. Anyone who has been pricing dates in this industry for eight years prices these conservatively.
For trade, almost nothing changes
This is the most common misreading I am running into at the moment. The April order explicitly added state licensed medical marijuana to the list of substances requiring import and export permits. The DEA actively made sure that Schedule III would not create a looser trade regime, because that would have put the United States out of step with Article 31 of the 1961 Single Convention. Every shipment needs an export permit, and it is only granted against a matching import permit from the destination country, within national quotas. This is exactly the system under which Canada, Portugal, Australia and Israel have been shipping to Germany for years.
In theory that opens a route for US product into Europe for the first time. In practice EU GMP stands in the way, and operators built around the internal market of a single US state will not clear that bar quickly. Anyone competing for flower volume in Germany does not need to worry about California yet.
Where the effect is immediate is capital. The end of 280E improves American operators’ cash flows straight away, the first listings on major exchanges have happened, and Trulieve spun off its medical business and took it to the NYSE in June. For Europe this cuts both ways. The uncomfortable side: institutional money that looked at Europe partly because the US market was simply not investable for many funds now has a home market again. The comfortable side: better capitalised US groups go shopping, and the targets sit where certified European capacity and market share already exist. We have seen that pattern in the transactions of recent months.
The point where Germany should look more closely
Of all four passages, the road safety testimony is the one that connects most directly to our own debate. Because we are having the same argument, with the roles reversed.
Until 22 August 2024 Germany applied a limit of 1.0 nanograms of THC per millilitre of blood serum. That number was never an impairment threshold. It was essentially the lower boundary of what a laboratory can reliably measure. Regular consumers remain above it for days after use without being impaired in any meaningful sense. That was the criticism, and it is why an interdisciplinary expert group convened by the transport ministry recommended a value tied to documented, driving relevant impairment. Since then 3.5 ng/ml has been written into section 24a of the Road Traffic Act, carrying a fine starting at 500 euros and a driving ban. Novice drivers and everyone under 21 remain subject to an absolute prohibition, as does mixed use with alcohol. And the limit only governs the administrative offence layer. Add erratic driving, observable impairment or a crash and you are into sections 316 and 315c of the criminal code, whatever the measured value. Patients on a prescribed therapy are exempt from the administrative offence when using the medicine as directed, but not from the rules on actual unfitness to drive.
The fight over that number has been running in a loop ever since, and both camps have a data problem. The TÜV association put it plainly back in 2024: it is scientifically established that cannabis raises crash risk, but nobody can currently say how many crashes in Germany actually result from it. Road safety associations nevertheless keep calling for a tighter limit, on the expectation that legalisation would put more intoxicated drivers on the road.
That expectation has now been tested. The Centre for Interdisciplinary Addiction Research at the University Medical Centre Hamburg-Eppendorf published a study in The Lancet Regional Health Europe in January 2026 comparing Germany eight months after the Consumer Cannabis Act against Austria as a control group, using a difference in differences design. The result: among monthly consumers, the share of trips under the influence of cannabis was, if anything, slightly lower than before, and no statistically significant effect of legalisation could be identified. That is consistent with what the EKOCAN evaluation programme has reported in its interim findings.
And here is the part I find most telling. In June 2026 the federal government introduced a bill amending the Road Traffic Accident Statistics Act, parliamentary document 21/6558, for the purpose of covering the statistical recording of THC concentration in blood serum in accidents at all. So we have been arguing about a threshold for more than two years and are only now starting to collect the data that would let anyone eventually verify it.
That is precisely the gap that opened up under oath in Washington. A witness for the prohibitionist side confirmed in cross examination that THC detection and impairment are two different things, and that the causal link is not supported by good government data. None of that means cannabis behind the wheel is harmless. It means the numbers driving policy on both sides of the Atlantic are considerably softer than the confidence of the demands suggests. Anyone in Germany calling for a stricter limit should be able to say which impairment they are regulating and how they measure it. Anyone defending the current value should be willing to sit through the evaluation that is only now becoming possible.
One more thing worth noting is what is missing from the drug commissioner’s correction list of 10 August. Hendrik Streeck names four areas: the difficulty for law enforcement of telling legal from illegal supply apart, tighter oversight of the medical cannabis market, stricter control of home cultivation and much earlier prevention work. Road traffic is not on it. That either signals that the evidence no longer supports the political heat, or it is a topic being saved for later. I would bet on the former, and I would use that quiet period rather than waste it.
What remains
In Washington the argument is about whether cannabis has an accepted medical use at all. Germany answered that question in 2017 and took the plant out of narcotics law in 2024. Since 30 July 2026, however, dried flower is no longer part of the statutory health insurance entitlement under section 31 paragraph 6 of the Social Code Book V. An estimated 65,000 publicly insured patients are affected, and no grandfathering was provided. Standardised extracts, dronabinol and nabilone remain reimbursable, as a rule only after a six month therapeutic trial with an approved finished medicinal product. In parallel, the amendment to the Medical Cannabis Act, with its telemedicine restrictions and mail order ban, is still sitting in the health committee because the coalition cannot agree.
So the United States is moving towards a model Germany is currently narrowing. Both movements turn on the same question, namely what counts as sufficient evidence for which application. And in both countries the answer is being given politically rather than scientifically.
One side effect of the American process should not be underestimated. Schedule III materially lowers the barriers to clinical research in the United States. The evidence that produces over the next three to five years lands in exactly the segment German policymakers have just designated as the preferred reimbursement route, namely standardised extracts and finished medicines. That is the most relevant long term consequence of this whole process, and it has nothing to do with trade.
What I will be watching over the coming months: the judge’s recommendation, the Administrator’s decision, the proposed rule before year end, and the appeals court proceedings. And in Berlin, whether the second and third readings of the Medical Cannabis Act amendment are called this year, and whether the accident statistics bill gets through. It sounds like administrative housekeeping. It will still shape the European market of 2027 more than any single headline out of Washington.
Sources
- Full DEA hearing transcript, 2,533 pages, eleven hearing days, released 25 August 2026 (dea.gov/NPRM2026)
- Acting Attorney General’s order of 23 April 2026 and the amendment to 21 CFR 1312.30 adding import and export permit requirements
- The Lancet Regional Health Europe, Schranz et al., January 2026, on driving under the influence of cannabis after partial legalisation (UKE, Centre for Interdisciplinary Addiction Research)
- German Bundestag, BT-Drs. 21/6558, draft second act amending the Road Traffic Accident Statistics Act, referred 25 June 2026
- German Bundestag, BT-Drs. 21/3061, draft first act amending the Medical Cannabis Act
- ABDA, new rules for cannabis medicines effective 30 July 2026 (revised section 31 paragraph 6 SGB V, GKV contribution rate stabilisation act)
- TÜV association, position paper on changing the THC limit in road traffic
- Süddeutsche Zeitung, interview with Hendrik Streeck, 10 August 2026
