England and Wales already spend an estimated £2.3 billion a year on cannabis. Almost none of that is taxed. Almost none of the jobs are on the books. And the police still open around 90,000 possession files a year for a drug most officers already treat as low-level.
That is the starting point for any serious conversation about a regulated recreational market in the UK. Not a fantasy industry waiting to be born. A consumer market that already exists — currently run by unlicensed growers, county lines and whoever answers a WhatsApp message — and two North American experiments that show what happens when a government tries to put the same trade on a licence.
Ontario built shops, a Crown wholesaler and a dividend back to the province. New York built a fight with unlicensed storefronts, then watched tax money appear only once the legal shops finally showed up. Britain would have to choose which lesson it wants.
The honest ceiling, once a legal market is mature, is not a new North Sea. It is roughly £0.5–1.5 billion a year to the Treasury, 7,000–15,500 jobs, and police processing time back. The illegal market does not vanish on Royal Assent. It vanishes if the legal product is good, local, and not taxed into stupidity.
The market is already here
In its 2023–24 proceeds of crime analysis, the Home Office estimated cannabis consumption in England and Wales at £2.3 billion, using a street price of £10 a gram from the National Crime Agency. A second method in the same paper produces £3.6 billion. Ministers still cite Dame Carol Black’s earlier figure of £2.4 billion. The range moves. The order of magnitude does not.
That money is already being spent. Border Force and police recorded 152,660 cannabis seizures in the year ending March 2024, including a record 85 tonnes of herbal cannabis. Crime Survey data put last-year illicit drug use among 16- to 59-year-olds at 8.8 per cent — about 2.9 million people — and cannabis is the drug most of them use.
A legal market would not create that demand. It would decide the terms on which it is met: tested products, age checks, PAYE, VAT, and a chance to pull spending away from organised crime. Or it would fail, and the WhatsApp market would keep the lot.
Canada is the caution against lazy scaling. Past-year cannabis use there is around 22 per cent. Copy Ontario’s shop takings onto the UK population and the numbers look heroic. Copy them onto UK prevalence and they look like the models we already have.
Five hours, two officers, a possession wrap
In the year ending March 2024, police in England and Wales recorded 90,405 cannabis possession offences. That is 71 per cent of all drug possession. Only 16 per cent ended in a charge or summons. 49 per cent were closed with a cannabis warning or a community resolution. Drugs remain the reason for 59.6 per cent of stop-and-searches.
The state has already decided, in practice, that most of this is not serious crime. It has not decided to stop doing the paperwork.
Asked in March 2025 for the annual cost of policing cannabis, the Home Office said it could not isolate the figure. That is not a dodge so much as a description of how the work actually happens.
The best observational account is still Tiggey May, Hamish Warburton, Paul Turnbull and Mike Hough’s 2002 Joseph Rowntree study, “Times they are a-changing”. They sat with officers in four basic command units, read custody records, and costed the 1999 caseload.
That year there were 76,769 recorded possession offences. 69,377 people were cautioned or convicted — one in seven of all known indictable offenders in England and Wales. A typical case took about five hours, usually with two officers: 770,000 officer hours, the equivalent of 500 constables. Police time alone came to £38 million, or about £500 a case. Add the Crown Prosecution Service and the magistrates’ courts and they put the total at around £50 million. A cruder method, sharing all drug-offence spending across the cannabis caseload, produced £350 million. They rejected it as an overestimate.
Uprated by consumer prices, that £38–50 million is about £72–95 million in 2025 money. The caseload did not wither in the meantime. By 2015 it was still 87,247 files — a figure the Liberal Democrats used in a 2017 briefing that also claimed a million police hours. The hour-count sits in the same family as May’s. The costing in that press release does not add up, and should not be used.
May’s other findings matter more than the cash. About three-quarters of the arrests were simple possession, with no concurrent offence. Most came to light as a by-product of other stops, or because someone was smoking in the street. Arrests very rarely opened the door to serious crime. Three per cent of the officers who made any possession arrests accounted for 20 per cent of them. Two-thirds of officers had already dealt with possession informally. Front-line practice had quietly decriminalised cannabis. The forms had not.
That is the honest shape of the saving. Legalisation does not abolish stop-and-search. It abolishes the file that starts when the only thing in the pocket is a bag of weed — the custody clock, the warning, the community resolution, the record. It is processing time, not a promise that 500 extra officers appear on burglary the next morning. Forces would still have to choose to spend the time on violence and county lines.
Transform’s 2025 “High Returns” model, which is advocacy-adjacent but built on later North American data, puts policing savings at £50–88 million and the whole criminal justice system at £174–284 million, depending on how commercial the legal market is. An older University of Essex study for the Institute for Social and Economic Research put police savings higher, around £291 million. Different package, same direction.
The live number is the 90,000 files. Everything else is an estimate of what those files cost.
The jobs are in the rooms, the vans and the tills
A legal cannabis industry is not one job. It is cultivation and processing, then wholesale and logistics, then retail — plus testing, security, compliance and the dull work of moving a controlled plant from a licensed room to a licensed counter.
Transform’s three UK scenarios put employment at 7,000 full-time jobs if the country copies a clubs-and-home-grow model with no commercial shops; about 15,500 if it licenses retail; and about 15,000 if the shops are run as a state monopoly. Some of those roles would be genuinely new. Some would be people already growing, trimming or delivering, coming onto PAYE. Both count as legal employment. Only one of them is a new shift.
Ontario is what the commercial version looks like after seven years. The Ontario Cannabis Store, a Crown wholesaler, supplies privately owned shops. In 2025 those shops sold 441 million grams of tested cannabis for C$2.28 billion — about £1.2 billion at September 2026 rates — from 1,780 authorised stores. A Deloitte study for the OCS estimates the provincial industry supported an average 31,900 jobs between 2018 and 2024, of which 8,600 were direct. Canada-wide, the same work puts supported jobs at 98,200 and direct jobs at 22,900. The distinction matters. A multiplier is not a shop payroll.
It worked because the shops were allowed to exist. When Mississauga opted into retail, a resident survey found legal purchasing rising from 62 per cent to 80 per cent. Past-year use barely moved, from 40 per cent to 39 per cent. That is the cleanest available story of shops displacing dealers without creating a new generation of consumers.
Health Canada’s 2024 cannabis survey now finds 72 per cent of past-year consumers usually buy from a legal storefront or website, up from 4 per cent in 2018. Statistics Canada put national legal recreational sales at C$5.5 billion in 2024/25 — about £2.9 billion.
The unglamorous half of the Ontario story is also the relevant half for Britain. Prices fell. Producers consolidated. Some towns have more cannabis shops than bookshops. A legal market that captures demand is not a cottage industry. It is a distribution system.
New York shows the other path. The Office of Cannabis Management talks about “thousands” of jobs. Industry research from Vangst put the 2024 headcount at 12,500, up 209 per cent in a year — a useful figure, not a state census. Adult-use stores went from 34 at the end of 2023 to 261 in 2024 and 556 by late 2025. Sales followed the store count, not the statute. Combined adult-use and medical takings were $317 million in 2023, $1.0 billion in 2024 and about $1.6 billion through late 2025. The legal market has now passed $2.5 billion in adult-use sales since the Marihuana Regulation and Taxation Act — roughly £1.9 billion.
The jobs arrived when the shops did. Until then, the high street belonged to whoever opened without a licence.
Germany is the footnote. The 2024 CanG law allowed home-grow and non-profit clubs, not commercial dispensaries. Licensing has been slow. There is no serious excise take and no shop payroll to count. It is what a Britain without retail would look like: some relief on possession, little of the economic story.
One figure that should not enter this piece except as a warning: an APPG-linked plan that talks about 594,000 jobs and £5.5 billion in tax. That folds in hemp, industrial cannabis and a Colorado-scaled fantasy. It is lobbying. It is not a forecast.
The tax only appears if people use the legal door
The Treasury does not tax a market it refuses to recognise. Once it does, the take depends on three things: how much of the £2.3 billion moves into licensed shops, what duty is stacked on the ticket, and whether the state also wants the retail profit.
Transform’s 2025 model is the most detailed current UK attempt. It assumes a five-year mature domestic market, no export, and — in the commercial scenarios — 20 per cent VAT plus 30 per cent excise, about half the retail price, with a £10/g average held up by minimum unit pricing. It is a reform organisation’s model. It should be read as a range, not a budget line.
Clubs and home-grow, no shops
Tax and other state revenue of £345 million — £87 million excise, £175 million VAT, £83 million in income tax and National Insurance. Criminal justice savings of £174 million. After a thin regulator, a net Treasury benefit of about £513 million. Legal share by year five: 45 per cent.
Licensed commercial retail
Revenue of £1.09 billion — £462 million excise, £376 million VAT, £184 million in income tax and NI, £68 million corporation tax. Criminal justice savings of £284 million. Net Treasury benefit of about £1.33 billion. Legal share by year five: 80 per cent.
State monopoly on retail
The same tax stack, plus about £180 million in shop profit to the state, for £1.23 billion in revenue. Same £284 million justice saving. Net Treasury benefit of about £1.49 billion. Legal share still 80 per cent — the extra money is the till, not a bigger market.
Two older independent checks sit in the same band. Bryan, Del Bono and Pudney at ISER, in 2013, put licensed-cannabis tax in England and Wales at £0.4–0.9 billion and the total improvement in the public finances at £0.5–1.25 billion. The Treasury later looked at that work. The Institute of Economic Affairs’ 2018 “Joint Venture” paper put product tax at £495–690 million on a market of about £2 billion, and warned that duty set too high — the California, Colorado and Washington error — keeps the illegal market alive. A 30 per cent duty on top of VAT is already about 36 per cent of the ticket: less than spirits, petrol or cigarettes.
Ontario shows what a working till looks like. In 2024–25 the OCS made C$246 million net income and paid a C$223 million dividend to the province — about £132 million and £119 million. Deloitte’s wider six-year number, C$5.3 billion in Ontario taxes, includes indirect and induced effects. It is not a single excise line, and should not be sold as one.
New York shows the ramp, and the politics. Cannabis taxes, fees and fines were $80 million in 2023–24 and $165 million the following year. From April 2023 to November 2025 the state collected $341 million — about £254 million — from adult-use and medical programmes combined. The statute spends it before the argument starts: 40 per cent to schools, 40 per cent to community reinvestment, 20 per cent to treatment and public education. The first Community Grants Reinvestment Fund round sent $100,000 each to 50 youth-focused nonprofits; more than 80 per cent of them serve ZIP codes hit hardest by the old drug laws.
That hypothecation is the most politically saleable design detail in the file. It is also a reminder that tax only arrives when there is a legal shop to ring it up. New York’s Office of Cannabis Management logged 2,017 enforcement actions in 2025 and seized $20 million of illicit product. The legal market is growing. The parallel high-street market has not left the building.
Ontario built shops. New York built a fight
The two jurisdictions are useful because they are not metaphors. They are similar in population to a large English region, both legalised adult use, and they chose opposite operating systems.
Ontario’s system is dull in the way that works. A Crown agency buys from licensed producers — the OCS now partners with around 267 — and sells on to private stores. Municipalities could opt in. The ones that did saw the illegal share fall. Public polling for the OCS now finds most adult Ontarians accept that the industry pays taxes and employs people. The price of that success is a dense, competitive retail map and a producer sector that has already been through a shakeout.
New York’s system was built to repair prohibition first and stock shelves second. Licensing was slow. Unlicensed shops filled the gap. Prices in the legal channel stayed above the national median. Only when the store count doubled did sales, and then tax, start to look like a market. Equity was the point of the statute. Delay was the cost. Britain can want both. It cannot pretend they arrive on the same morning.
The design lesson is blunt. Access plus a competitive legal price is what kills the illegal market. Jobs and tax follow the shops, not the other way around. A Crown wholesale layer — the OCS model — is the closest real-world version of a hybrid in which the state keeps a hand on the middle of the chain without having to staff every till.
What Britain would actually have to decide
A UK law would not be a mood. It would be a stack of choices, each of which moves the three ledgers.
Who can buy. Eighteen, in line with alcohol, or 21, in line with some US states. The Crime Survey already includes teenagers. A legal market would not sell to them. An illegal one already does.
Who can grow. Home-grow sits in every Transform scenario and in Germany’s law. It is a pressure valve and a hole in the tax base. Clubs without shops raise less money and capture less of the market — 45 per cent in Transform’s year-five estimate, against 80 per cent once retail exists.
Who can sell. Licensed private shops are the plausible political landing zone: the thing voters can see, the thing that employs people on the high street, the thing Ontario used to move consumers. A state monopoly on retail, or an OCS-style Crown wholesaler with private shops, keeps more of the profit in public hands and reduces the risk of a few chains owning the map. Germany’s clubs-only model is what happens if commercial retail is too difficult to pass. It saves some police time. It does not build an industry.
How hard to tax. Tobacco-style duty would look tough and feed the illegal market. Beer-style duty would look soft and empty the WhatsApp groups. The working range in the UK papers is VAT plus a duty that leaves legal cannabis cheaper than, or close to, the current £10/g street price. Minimum unit pricing is how you stop a race to the bottom without handing the price war back to unlicensed growers.
Where the money goes. General Treasury, or a New York-style split into schools, community reinvestment and treatment. The second is easier to defend in a doorstep argument. The first is how most UK taxes actually work.
What happens to the medical market, and to records. Britain already has a licensed medical cannabis sector and a generation of people with possession convictions that still appear on enhanced checks. Any adult-use law that ignores both will look like it legalised the plant and left the damage in place.
What the product is allowed to be. THC caps, required CBD, labelling, shop density, opening hours, advertising. These are public-health choices, not decorations. They are also why some consumers will stay on the illegal market if the legal shelf is weak, expensive or boring.
None of this settles the arguments that opponents will bring, and should bring: youth access, high-THC flower and psychosis risk, workplace safety, driving. Canada’s higher use rate existed before legalisation; it is not a UK baseline. A serious piece has to hold those points without treating them as a veto on the arithmetic.
The arithmetic is this. A clubs-only Britain is a smaller fiscal story — hundreds of millions, not a billion, and about 7,000 jobs. A Britain with licensed shops, moderate tax and enough stores to be convenient is a £1 billion-plus Treasury story and a 15,000-job industry, if the models are in the right neighbourhood. A Crown wholesale or state-retail variant takes more of the profit and gives ministers more control over the shop floor.
The choice is who keeps the £2.3 billion
There is a version of this debate that treats legalisation as a jackpot and a version that treats it as a moral crusade. The evidence from Ontario and New York supports neither.
It supports something drier. Britain is already policing a mass-market drug it does not control, at a volume that has barely moved in 25 years, at a cost ministers say they cannot calculate and researchers keep putting in the same tens-of-millions band. Consumers are already spending low billions. The jobs already exist in the informal economy. The tax does not.
Ontario shows that if you open enough legal doors, most people walk through them, and use does not have to explode for that to happen. New York shows that if you open them late, you spend the first years raiding the shops that opened without you, and the Treasury waits.
A regulated UK market would be a decision about design: shops or clubs, duty or a black market, a dividend or a fight. The £2.3 billion is not the prize. It is the status quo.
Sources
Home Office proceeds of crime 2023–24, crime outcomes, drug seizures and stop and search; May, Warburton, Turnbull and Hough, “Times they are a-changing” (JRF, 2002); Transform, “High Returns” (2025); ISER / Bryan, Del Bono and Pudney (2013); IEA, “Joint Venture” (2018); Ontario Cannabis Store and Deloitte six-year study; Statistics Canada; Canadian Cannabis Survey 2024; New York Office of Cannabis Management, 2025 annual report. Sterling conversions use mid-September 2026 rates (about £1 = C$1.87 and $1.34).
