Selling it legally: the United States, Canada, and the UK's medical-only route

In 2012 voters in Colorado and Washington chose to legalise cannabis for adult, non-medical use. Six years later Canada did the same for an entire country. On 1 November 2018, a fortnight after the Canadian change, the United Kingdom took a much smaller step: it allowed specialist doctors to prescribe cannabis-based medicines, while leaving every other use a crime.

Two of these jurisdictions opened licensed, profit-making markets; the third opened a narrow medical door and kept the rest of the house locked. Both camps in the cannabis debate point to them. Supporters of reform cite falling arrests and the absence of a teenage surge. Opponents cite rising daily use and accidental poisonings of small children. Both sets of claims are partly right. The useful question is which outcomes the evidence supports, and how firmly.

This is the fourth article in our series on prohibition and legalisation. The first sets out the different kinds of regime, the second weighs the record of criminal prohibition, and the third looks at decriminalisation and non-commercial models. Here we turn to the most liberal model, commercial legalisation, and to the UK's medical-only route. The final article takes up the health evidence and the regulatory toolkit in more depth.

Three countries, three answers

"Legal" and "illegal" are too crude as categories. The policy literature distinguishes several regimes, and evidence from one does not transfer automatically to another (Kilmer, 2017; Hall and Lynskey, 2020). The three countries in this article sit in different places on that spectrum.

The defining feature of commercial legalisation is in the supply column: the market is run by private firms that make money from selling cannabis. A licensed shop that checks identification is a different thing from a street dealer. A company with a marketing budget and an interest in selling more is a different thing again. Keeping those two features apart is one of the main tasks of this article.

How to read the evidence

The review this series is based on weights evidence by how well a study can separate the effect of a policy from everything else changing at the same time. Four hazards recur in this literature.

• Pre-existing trends. Cannabis use was already rising in many places before reform. A rise after a law changes is not, by itself, evidence that the law caused it.
• Reporting effects. When use becomes legal, people become more willing to admit to it in surveys. Part of an apparent increase may be more honest answers.
• Potency drift. The product changed over the decades, so "use" in one decade is not the same exposure as "use" in another.
• Short follow-up. Most commercial markets are less than a decade old. Harms that build up over years of heavy use, such as dependence or psychosis, may not yet be visible.

These hazards cut both ways. They weaken claims that reform has been harmless, and they equally weaken claims that it has been disastrous.

Several study designs appear below.

Difference-in-differences. Suppose youth cannabis use falls by two points in states that legalised and by three points in similar states that did not. The before-and-after change in the legalising states (minus two) is misleading, because use was falling everywhere. The difference between the two differences (plus one) is a better estimate of what the law did. The method assumes the two groups would have moved in parallel without the reform. It is "quasi-experimental": not a randomised trial, but an attempt to approximate one with a comparison group.

Odds ratios. Odds are the chance that something happens divided by the chance that it does not. An odds ratio of 1 means no difference; 0.92 means the odds are 8% lower. When an outcome is fairly common, an 8% fall in the odds is not exactly 8% fewer people, but it points the same way.

Confidence intervals and "no association". Every estimate carries uncertainty, usually shown as a confidence interval: the range of values reasonably compatible with the data. A finding of no association generally means the best estimate is close to zero and the plausible range includes both small rises and small falls. It does not prove the effect is exactly zero.

Interrupted time series. Researchers follow an outcome, such as hospital visits, over a long period and ask whether its level or slope breaks when a policy changes. A trend that carries on through the reform without a kink suggests the reform was not the main driver.

Surveys are good at tracking trends but vulnerable to reporting effects. Hospital and poison-centre records do not depend on anyone admitting to drug use, but they count only cases that reach a service.

The United States: legal by state, illegal by federal law

Colorado and Washington voted to legalise adult use in 2012, and many states have since followed. The result is a large, varied and largely for-profit market that remains illegal under federal law for non-medical use. Because states changed their laws at different times, and some not at all, they offer ready-made comparison groups for difference-in-differences studies.

Adolescents: the prediction that did not come true

The most frequently stated fear was that legalisation would drive up use among teenagers, by making cannabis more available, more normal and less risky in young people's eyes. The fear deserves to be taken seriously: starting cannabis use at a younger age is associated with a greater likelihood of later problem use (National Academies of Sciences, Engineering, and Medicine, 2017).

So far the prediction has not been borne out. Using Youth Risk Behavior Survey data for 1993 to 2017, Anderson et al. (2019) found that recreational marijuana laws were associated with an 8% decrease in the odds of high-school students using cannabis (odds ratio 0.92). An updated analysis to 2019, covering about 1.6 million students, found no association in either direction (Anderson et al., 2021). The authors suggest that licensed retailers who check identification may be harder for teenagers to buy from than illicit dealers.

The safest summary is therefore not that legalisation reduced teenage use, but that it has not increased it on average. Two qualifications apply. Follow-up is still short. And Cerdá et al. (2020) found a small increase in cannabis use disorder among adolescents who used cannabis after recreational legalisation. The number of teenagers using may not have risen, but there is a signal of more problematic use among those who do.

Adults: the shift towards daily use

Among adults the picture is different. Caulkins (2024) analysed national survey data from 1979 to 2022. He found that in 2022 an estimated 17.7 million Americans used cannabis daily or near-daily, more than the 14.7 million who drank alcohol that often. Per capita daily or near-daily use rose fifteen-fold between 1992 and 2022, and about 40% of current users now use daily or near-daily, a pattern closer to tobacco than to alcohol. Many drinkers drink only occasionally, whereas tobacco smokers typically smoke every day. American cannabis use appears to be moving from the first pattern towards the second.

Caulkins is careful about what his data can show. He notes that greater willingness to report use explains part of this rise, but not all of it. As cannabis became legal in more states, people who once would have denied using it in a government survey may have stopped denying it. Nobody knows exactly how much of the rise this accounts for.

A second line of evidence points the same way. Cerdá et al. (2020), covering 2008 to 2016, found that recreational legalisation was associated with increases in frequent use and in cannabis use disorder among adults aged 26 and over.

Frequency matters because it is the strongest predictor of dependence, the most common serious harm of cannabis. It also matters for psychosis: in a large European case-control study, daily users of high-potency cannabis were about five times as likely as never-users to have a psychotic disorder (Di Forti et al., 2019). A market in which a growing share of customers use every day is one in which the main harms are likely to grow, even if the number of occasional users stays stable.

Young children and edibles

Edible products created a risk that scarcely existed under prohibition: accidental poisoning of young children who find and eat them. Exposures to edible cannabis among US children under six reported to poison centres rose from 207 in 2017 to 3,054 in 2021, an increase of 1,375% (Tweet, Nemanich and Wahl, 2023). Poison-centre figures count calls rather than every case, but the scale of the rise, and its link to one type of product, make this one of the clearest findings in the field. It is a harm caused by a specific design choice, allowing edibles that look like sweets, rather than by legalisation as such.

Arrests and racial disparity

The criminal justice record is where reform's supporters stand on firmest ground. The American Civil Liberties Union analysed arrests from 2010 to 2018 and found that Black people were 3.64 times more likely than white people to be arrested for marijuana possession, despite similar rates of use, and that the disparity existed in every state (ACLU, 2020). Legalisation sharply reduced arrests for possession, and states that had legalised possession had, on average, lower racial disparities in the arrests that remained than states that had only decriminalised or kept prohibition. Since a possession conviction can block employment, travel and housing for years, this matters well beyond the arrest itself, as the second article explains.

Two questions, not one

The US experience separates two questions that debate tends to run together. Has legalisation harmed adolescents? On average, so far, not visibly. Has the commercial market changed how adults use cannabis? Here the answer appears to be yes: commercial markets seem to have normalised and intensified heavy use among adults, which is where most cannabis-related harm is concentrated (Hall and Lynskey, 2020).

A debate framed entirely around teenagers can miss the larger effect. An argument against legalisation built mainly on the risk to young people has so far found little support in the evidence; an argument for it that assumes nothing else would change has found even less.

Canada: a national market, measured closely

Canada legalised adult use nationally on 17 October 2018. Edibles and extracts were permitted from late 2019 and on sale from early 2020. Because the whole country changed at once, there is no domestic comparison group. Against that, Canada has unusually good monitoring data: the annual Canadian Cannabis Survey and population-wide health data from Ontario.

What the national survey shows

Source: Health Canada (2024).

Overall use rose by four percentage points over six years. Without a comparison group, and with use already rising in many countries, the survey cannot say how much of that legalisation caused, and reporting effects apply here too.

The most striking change is where people buy. The share of users buying mainly from legal sources roughly doubled, a large shift from illicit to licensed sellers and one of legalisation's central aims. It is also what the non-commercial models in Uruguay and Germany have struggled to achieve, as the previous article describes. Use among 16- to 19-year-olds was not rising, consistent with the US finding of no adolescent surge. Daily or near-daily use was stable rather than climbing, though at about a quarter of users it is high. Reported driving after use fell, though that too is self-reported.

What hospital records add

Population-wide health data from Ontario add a more cautious picture. These records do not rely on anyone admitting to drug use, and they can be analysed as interrupted time series.

Emergency department visits for cannabis poisoning in children under ten rose about nine-fold after legalisation. The largest increase followed the legal sale of commercial edibles, after which nearly 10% of all paediatric poisoning visits in Ontario involved cannabis (Myran et al., 2022). The biggest rise came not with legalisation itself in October 2018 but after edibles reached the shelves, linking the harm to a product decision, the same pattern as in the US.

Schizophrenia: a rising share, and a careful reading

In a cohort of more than 13.5 million people, the share of new schizophrenia cases associated with cannabis use disorder rose from 3.7% before legalisation to 10.3% after it (Myran et al., 2025).

At first sight this looks like strong evidence that legalisation caused more schizophrenia. The authors warn against that reading. They stress that the rise was steady over the whole period, with no clear acceleration at the point of policy change; in time-series terms, the trend did not bend when the law did. It may therefore reflect long-run increases in potency and heavy use more than legalisation as such.

That caveat cuts both ways. The Ontario data do not show that legalisation caused a jump in schizophrenia. But the underlying problem is real and growing, and it fits evidence linking psychosis to daily use of high-potency cannabis in cities under quite different legal regimes (Di Forti et al., 2019).

A mixed record

Canada has achieved its market-displacement goal to a large degree and has not seen an adolescent surge. It has not reduced adult use, and product liberalisation, especially edibles, carried predictable harms to children. Its record is mixed rather than clearly positive or negative.

The commercialisation problem

The US and Canadian records share a pattern. Many of the harms that followed legalisation come less from cannabis being legal than from the incentives of a commercial industry.

Who the best customers are

In a commercial market, the most profitable customers are the heaviest users. Someone who uses cannabis once a month contributes little to a seller's revenue; someone who uses it several times a day contributes a great deal. A company that wants to grow has an incentive to make heavy use easier and cheaper, whatever its intentions.

Kilmer (2017) argues that this gives a for-profit cannabis industry incentives similar to those of the alcohol and tobacco industries: to increase frequent use, lower prices and lobby against restrictions. Hall and Lynskey (2020) reach a similar conclusion from the US evidence, noting that commercial legalisation has been followed by falling prices, rising potency and more daily use among adults. The review concludes that these adult harms are concentrated among a minority of heavy users, who are also the industry's most valuable customers.

The tobacco experience suggests that once an industry becomes economically and politically powerful, it is very difficult to regulate back. Rules that are easy to write before an industry exists can become very hard to pass once it does. None of this shows that legal supply is wrong in itself. It shows that the form of legal supply matters, and that early decisions may be hard to reverse.

The opposite risk

Non-commercial models avoid much of this marketing pressure but meet a different problem. Uruguay's state model and Germany's non-profit associations limited promotion but under-supplied the market, leaving space for grey and illicit sellers (Queirolo et al., 2024; EKOCAN, 2026). Policy design therefore faces a genuine trade-off between displacing criminal markets and avoiding the promotion of heavy use. Canada sits towards one end, Uruguay towards the other.

Between prohibition and an open market lies a set of regulatory levers: limits on THC content and on edibles that resemble sweets, child-resistant packaging, taxes linked to THC content, marketing bans, outlet licensing and age checks, and a choice of ownership model, from state monopoly to non-profit associations to private firms (Room et al., 2010; Rolles and Murkin, 2016). Several North American harms map directly onto these levers: child poisonings rose with edibles, and heavy use rose alongside falling prices and rising potency. A legal regime with weak controls and one with strong controls are different policies with different outcomes.

The United Kingdom: medicine on paper, a market for those who can pay

The United Kingdom has remained a prohibition state for non-medical use. Since 1 November 2018, specialist doctors have been able to prescribe cannabis-based products for medicinal use (Home Office, 2018).

What happened in practice

The evidence for medical use is real but narrow. The US National Academies found conclusive or substantial evidence that cannabis or cannabinoids are effective for chronic pain in adults, for chemotherapy-induced nausea and vomiting (oral cannabinoids) and for patient-reported spasticity in multiple sclerosis; for most other conditions marketed by private clinics, the evidence remains limited or absent (National Academies of Sciences, Engineering, and Medicine, 2017).

NHS access has been minimal. NHS Business Services Authority data indicate that fewer than five patients have received NHS prescriptions for unlicensed cannabis-based products in the community (NHSBSA, n.d.). Almost all prescribing takes place in private clinics, where prescriptions more than doubled between 2021/22 and 2022/23 according to Care Quality Commission data, and where an estimated 45,000 patients were being treated by 2024 (Pharmaceutical Journal, 2024).

A two-tier outcome

The result is a two-tier system in which lawful access depends on the ability to pay. A patient who can afford a private clinic can obtain a lawful prescription. A patient with the same condition who cannot remains exposed to the criminal law if they obtain cannabis elsewhere, and growing the plant for personal medical use remains a serious offence. This is an outcome of partial reform, not of the evidence on efficacy: if the evidence supported prescribing for a condition, it would support it on the NHS too.

There is a fair case on the other side. Caution about prescribing unlicensed products with limited trial evidence is a defensible clinical position, and the thinness of that evidence is partly a legacy of prohibition, which made cannabis hard to study (National Academies of Sciences, Engineering, and Medicine, 2017). But caution has not prevented a large private market; it has only decided who pays.

To be precise about the law: in the UK, cultivation and possession of cannabis remain criminal offences, and growing it needs a Home Office licence even for patients with a prescription. This article is analytical and gives no guidance on obtaining, growing or using cannabis.

What the review suggests for the UK

The review is cautious about transferring North American results to Britain, which has a different illicit market, policing culture and health system. It identifies three priorities that would not require full legalisation: ending the criminalisation of personal possession, reviewing the use of the smell of cannabis as grounds for stop-and-search, and reforming the medical scheme so that access does not depend on ability to pay. Any of these would remain consistent with the international conventions. These are the review's policy conclusions, not a description of current law, which is unchanged. And because Britain has no commercial adult-use industry, it has none to regulate back; if wider reform were ever considered, the design of supply would deserve the most attention.

Above the nation state: the UN vote and Schedule III

Two changes at the top of the control system are symbolically important but practically limited.

The 1961 Single Convention sorts drugs into schedules. Schedule I covers substances under full control; Schedule IV is a stricter list within it for substances regarded as particularly dangerous and of little therapeutic value. In December 2020 the UN Commission on Narcotic Drugs removed cannabis from Schedule IV while keeping it in Schedule I (UNODC, 2020). The vote did not authorise non-medical markets, and countries such as Canada remain in tension with a treaty that restricts cannabis to medical and scientific purposes.

US federal law has its own schedules. Schedule I is for substances with a high potential for abuse and no currently accepted medical use; Schedule III covers drugs with accepted medical uses and a lower potential for abuse. In April 2026 the US Drug Enforcement Administration moved FDA-approved cannabis products and state-licensed medical marijuana into Schedule III. Non-medical marijuana remains in Schedule I, and proceedings on wider rescheduling were stayed at the time the review was written (Drug Enforcement Administration, 2026; Congressional Research Service, 2026). The adult-use markets in many states therefore still operate in conflict with federal law.

Both changes acknowledge medical value without endorsing non-medical legalisation. In that respect they resemble the UK's 2018 reform more than Canada's.

Weighing the two camps

The strongest case for commercial legalisation. Prohibition did not stop cannabis becoming widely used, and its enforcement fell hardest on Black and minority communities. Legalisation sharply reduced arrests and narrowed racial disparities. The feared rise in teenage use did not happen; licensed shops may even be harder for teenagers to buy from than dealers. Canada moved most of its market to legal sellers within a few years. Some of the rise in reported use may be more honest survey answers.

The strongest case against. In the United States, more people now use cannabis daily or near-daily than drink alcohol that often, and frequency is the strongest predictor of dependence. Adult cannabis use disorder has increased. Reported edible exposures among children under six reached 3,054 in a single year. A growing share of new schizophrenia cases in Ontario is associated with cannabis use disorder. An industry that profits most from its heaviest users has every incentive to deepen that pattern, and the tobacco experience shows how hard such an industry is to rein in.

Both cases rest on real evidence. The table sets the main claims against it.

The two camps are often answering different questions. Reform's supporters are largely right about teenagers and about the unequal burden of enforcement. Its critics are largely right about heavy adult use and edibles.

What we know, and what we don't

Reasonably well established:

• US adult-use legalisation has not, on average, increased cannabis use among high-school students on the evidence so far (Anderson et al., 2019; Anderson et al., 2021).
• Daily and near-daily use among US adults has risen sharply, though part of the rise reflects greater willingness to report use (Caulkins, 2024).
• Legal edibles have been followed by large increases in accidental poisonings of young children in the US and Canada (Myran et al., 2022; Tweet, Nemanich and Wahl, 2023).
• Legalisation reduced possession arrests and was associated with lower racial disparity in those that remained (ACLU, 2020).
• Canada moved most buyers from illicit to legal sources (Health Canada, 2024).
• In the UK, lawful access to cannabis-based medicines runs almost entirely through private clinics (NHSBSA, n.d.; Pharmaceutical Journal, 2024).

Still uncertain:

• How much of the US rise in daily use is real and how much is a reporting effect.
• Whether the rising share of schizophrenia associated with cannabis use disorder in Ontario owes anything to legalisation itself rather than long-run increases in potency and heavy use (Myran et al., 2025).
• What harms will appear as commercial markets mature, and whether the small rise in cannabis use disorder among adolescent users (Cerdá et al., 2020) will grow.
• How well any of this transfers to the UK.

The broader lesson is that commercial legalisation is not one policy but many. The North American outcomes followed largely from specific choices: allowing edibles, letting prices fall and potency rise, permitting marketing, and handing supply to firms whose best customers are their heaviest users. The UK's medical-only route avoided those choices but produced a different problem, lawful access that depends on ability to pay. The question is less whether cannabis should be legal than how its supply, where it exists, should be governed. That is the subject of the final article in this series.

This article is based on the critical review it expands; figures are reported as that review gives them and were not re-checked against every original study.

Sources

ACLU (American Civil Liberties Union) (2020) A Tale of Two Countries: Racially Targeted Arrests in the Era of Marijuana Reform. New York: ACLU. Available at: https://www.aclu.org/publications/tale-two-countries-racially-targeted-arrests-era-marijuana-reform.

Anderson, D.M., Hansen, B., Rees, D.I. and Sabia, J.J. (2019) 'Association of marijuana laws with teen marijuana use: new estimates from the Youth Risk Behavior Surveys', JAMA Pediatrics, 173(9), pp. 879–881. doi:10.1001/jamapediatrics.2019.1720.

Anderson, D.M., Rees, D.I., Sabia, J.J. and Safford, S. (2021) 'Association of marijuana legalization with marijuana use among US high school students, 1993–2019', JAMA Network Open, 4(9), e2124638. Available at: https://pmc.ncbi.nlm.nih.gov/articles/PMC8424481.

Caulkins, J.P. (2024) 'Changes in self-reported cannabis use in the United States from 1979 to 2022', Addiction. doi:10.1111/add.16519.

Cerdá, M., Mauro, C., Hamilton, A., Levy, N.S., Santaella-Tenorio, J., Hasin, D., Wall, M.M., Keyes, K.M. and Martins, S.S. (2020) 'Association between recreational marijuana legalization in the United States and changes in marijuana use and cannabis use disorder from 2008 to 2016', JAMA Psychiatry, 77(2), pp. 165–171. doi:10.1001/jamapsychiatry.2019.3254.

Congressional Research Service (2026) Department of Justice Eases Control of Medical Marijuana (Legal Sidebar LSB11424). Washington, DC: Library of Congress. Available at: https://www.congress.gov/crs-product/LSB11424.

Di Forti, M., Quattrone, D., Freeman, T.P., Tripoli, G., Gayer-Anderson, C., Quigley, H. et al. (2019) 'The contribution of cannabis use to variation in the incidence of psychotic disorder across Europe (EU-GEI): a multicentre case-control study', The Lancet Psychiatry, 6(5), pp. 427–436. doi:10.1016/S2215-0366(19)30048-3.

Drug Enforcement Administration (2026) 'Schedules of Controlled Substances: Rescheduling of Food and Drug Administration approved products containing marijuana from Schedule I to Schedule III', Federal Register, 91 FR 22714, 28 April 2026.

EKOCAN (2026) Evaluation des Konsumcannabisgesetzes (EKOCAN): 2. Zwischenbericht, 1 April 2026. Hamburg, Düsseldorf and Tübingen: EKOCAN consortium. Available at: https://www.uke.de/landingpage/ekocan/.

Hall, W. and Lynskey, M. (2020) 'Assessing the public health impacts of legalizing recreational cannabis use: the US experience', World Psychiatry, 19(2), pp. 179–186. doi:10.1002/wps.20735.

Health Canada (2024) Canadian Cannabis Survey 2024: Summary. Ottawa: Government of Canada. Available at: https://health-infobase.canada.ca/cannabis/.

Home Office (2018) Home Office Circular 2018: Rescheduling of cannabis-based products for medicinal use in humans. London: Home Office.

Kilmer, B. (2017) 'Recreational cannabis — minimizing the health risks from legalization', New England Journal of Medicine, 376(8), pp. 705–707. doi:10.1056/NEJMp1614783.

Myran, D.T., Cantor, N., Finkelstein, Y., Pugliese, M., Guttmann, A., Jesseman, R. and Tanuseputro, P. (2022) 'Unintentional pediatric cannabis exposures after legalization of recreational cannabis in Canada', JAMA Network Open, 5(1), e2142521. doi:10.1001/jamanetworkopen.2021.42521.

Myran, D.T. et al. (2025) 'Changes in incident schizophrenia diagnoses associated with cannabis use disorder after cannabis legalization', JAMA Network Open, February 2025. Available at: https://pubmed.ncbi.nlm.nih.gov/39903464/.

National Academies of Sciences, Engineering, and Medicine (2017) The Health Effects of Cannabis and Cannabinoids: The Current State of Evidence and Recommendations for Research. Washington, DC: The National Academies Press. Available at: https://www.ncbi.nlm.nih.gov/books/NBK425762/.

NHSBSA (NHS Business Services Authority) (n.d.) FOI-01304: NHS and private prescriptions for unlicensed cannabis medicines. Available at: https://opendata.nhsbsa.net/dataset/foi-01304.

Pharmaceutical Journal (2024) 'Prescriptions for cannabis-based medicines double in a year, CQC report says'. Available at: https://pharmaceutical-journal.com/article/news/prescriptions-for-cannabis-based-medicines-double-in-a-year-cqc-report-says.

Queirolo, R., Repetto, L., Alonso, J., Álvarez, E., Sotto, B., Pardal, M. and Kilmer, B. (2024) 'Intended and unintended effects of cannabis regulation in Uruguay'. Available at: https://repositorio.ausjal.org/handle/20.500.12032/179430.

Rolles, S. and Murkin, G. (2016) How to Regulate Cannabis: A Practical Guide. 2nd edn. Bristol: Transform Drug Policy Foundation.

Room, R., Fischer, B., Hall, W., Lenton, S. and Reuter, P. (2010) Cannabis Policy: Moving Beyond Stalemate. Oxford: Oxford University Press.

Tweet, M.S., Nemanich, A. and Wahl, M. (2023) 'Pediatric edible cannabis exposures and acute toxicity: 2017–2021', Pediatrics, 151(2), e2022057761. doi:10.1542/peds.2022-057761.

UNODC (2020) CND votes on recommendations for cannabis and cannabis-related substances. Vienna, 3 December 2020. Available at: https://www.unodc.org/unodc/en/frontpage/2020/December/cnd-votes-on-recommendations-for-cannabis-and-cannabis-related-substances.html.