Americans spend roughly $100 billion a year on cannabis. Washington taxes almost none of it. Not because it decided not to — but because, under federal law, most of that money changes hands in a market that officially isn’t supposed to exist. Alcohol pays an excise tax. Cigarettes pay an excise tax. Marijuana, the third member of the usual “sin tax” trio, pays nothing at the federal level, for the simple reason that the federal government still classifies it alongside heroin.
Economists have wondered for years what that gap is actually worth. On 17 August 2026, a team at The Budget Lab at Yale University published the most careful public attempt yet to answer the question. Their report, Federal tax implications of legalizing marijuana, models what would happen if Congress legalized cannabis for both medical and adult recreational use and then taxed it the way it taxes whiskey and tobacco. The headline figure is $57.9 billion over ten years. The more interesting part of the report is everything the authors admit they don’t know.
Key Takeaways
- Federal legalization plus a THC-based excise tax would raise an estimated $57.9 billion over ten years, according to The Budget Lab at Yale.
- If every remaining state legalized as well, the ten-year figure roughly doubles to $111.3 billion.
- The modelled tax is $0.00625 per milligram of THC — about $1.31 per gram, or a 15% bump to the tax-inclusive price of an average $8.59 gram.
- Roughly 75% of the total US cannabis market remains illicit, and how much of it would move into the formal economy is the single biggest unknown in the estimate.
- The authors deliberately excluded several revenue streams they consider real but unmeasurable, meaning the true figure is more likely to be higher than lower.
Why the Federal Government Can’t Tax America’s Biggest Grey Market
Marijuana has been a Schedule I controlled substance under the Controlled Substances Act since Congress wrote the statute in 1970. Schedule I is the most restrictive category the federal system has: it formally declares that a substance has no accepted medical use and a high potential for abuse. Heroin sits there too.
That classification is now visibly wobbling. In December 2025, an executive order directed the administration to loosen federal regulation of medical marijuana. In April 2026, the Justice Department and the DEA moved FDA-approved marijuana products and cannabis sold under state medical licences into Schedule III, while opening a hearing process to consider rescheduling marijuana more broadly. None of this makes cannabis legal without a prescription — that would almost certainly require an act of Congress — but it is the first meaningful federal retreat in half a century.
Meanwhile the states raced ahead. Medical cannabis has been legalized in some form in 47 states and the District of Columbia, starting with California in 1996. Adult recreational use is legal in 24 states plus DC. The result is a regulatory patchwork with strange consequences: cannabis businesses generally cannot use federally regulated banks, cannot deduct ordinary business expenses on their federal returns, and in some cases cannot draw water from federally managed sources. They are legal enough to be licensed and taxed by their state, and illegal enough to be locked out of the ordinary machinery of American commerce.
How Do You Tax a Plant? Yale Says: By the Milligram
The Budget Lab’s model doesn’t tax the price tag. It taxes the drug.
That choice mirrors how the federal government already handles alcohol and tobacco. Spirits are taxed by proof gallon — effectively by pure alcohol content. Beer is taxed by the barrel, wine by alcohol content, smokeless tobacco by weight. The logic is that the thing causing the social cost is the substance itself, not the marketing budget wrapped around it.
So the researchers modelled an excise tax of $0.00625 per milligram of THC, collected from producers when product moves to retail. In practice that works out to about $1.31 on a gram of marijuana. Against an average retail price of $8.59 per gram, the tax would raise the final tax-inclusive price by roughly 15%.
A potency-based tax has three practical advantages the authors highlight. It targets the psychoactive compound rather than the packaging. It stays stable when market prices swing — and cannabis prices have swung violently in mature markets. And the revenue scales with how much THC people actually consume, not with how much they happen to spend.
$57.9 Billion or $111.3 Billion? It Depends What the States Do
The report runs two scenarios, both assuming the tax starts in calendar year 2027 and both measured across the standard ten-year federal budget window.
In the first, Congress legalizes marijuana federally and the states change nothing. Prohibition states stay prohibition states. Ten-year federal revenue: $57.9 billion.
In the second, Congress legalizes and every remaining state follows. Sales grow in the newly legal states — partly genuinely new demand, partly demand redistributed from neighbouring states that used to soak up “cannabis tourism.” Ten-year federal revenue: $111.3 billion.
The underlying market is already substantial. Using state tax revenue data, the researchers estimate the US recreational cannabis market was around $25 billion in 2024, and project it reaching nearly $40 billion by 2035 even with no further legalization anywhere. The medical segment, older but smaller, accounts for about 20% of activity — between $5 and $8 billion in annual sales across the same period. These are not the only numbers on the market; independent forecasts of the global cannabis market vary enormously depending on methodology, a problem we’ve examined before in our look at why market-size forecasts range from $47 billion to $137 billion.
The Ghost Market: Three of Every Four Dollars
Here is the number that makes the whole exercise uncertain. Citing data from the industry research firm Whitney Economics, the report estimates that roughly 75% of the total US cannabis market — around $100 billion in combined legal and illegal activity — is still illicit. About two-thirds of that illegal supply is grown domestically.
Crucially, this isn’t just a prohibition-state phenomenon. Even in states that legalized a decade ago, the underground market has stayed stubbornly price-competitive, because illicit operators pay no licensing fees, no state excise taxes, no compliance costs, and no federal tax penalty. California seized more than $534 million worth of illegal cannabis in 2024 alone, and has confiscated and destroyed nearly 800 tons worth over $2.8 billion since 2019. The persistence of parallel illegal supply alongside legal shops is a pattern documented internationally as well — the UNODC has tracked how legal markets and illegal ones coexist and feed each other worldwide.
Federal legalization would strip away several of the structural advantages holding that shadow economy together: no banking ban, no interstate commerce ban, no barrier to formal employment. How much activity migrates as a result is anyone’s guess, and the report says so plainly. Converting even a quarter of the illicit market would produce a dramatically larger tax base than a scenario where a high federal excise rate accidentally hands the black market a fresh price advantage.
Why the Authors Think Their Own Number Is Too Low
This is the part that separates a serious fiscal estimate from a lobbying document. The Budget Lab identifies at least three revenue streams it believes are real, then declines to count them.
The first is income and payroll tax from workers coming out of the shadows. Cultivators, trimmers, distributors and sellers in the illicit economy currently earn income that is neither reported to the IRS nor subject to payroll withholding, because their employers have no lawful basis for formal employment at the federal level. Legalization would split that invisible workforce in two: employees absorbed into licensed firms receiving W-2 wages, and owner-operators filing as self-employed. The state-legal industry already supports over 440,000 full-time jobs. The Tax Foundation has estimated around $1.5 billion in annual federal revenue from this channel — a figure the Yale team suspects understates things, but which depends on assumptions about worker transition rates that nobody can currently observe. So it’s left out.
The second is Section 280E of the Internal Revenue Code. Because it applies to anyone trafficking Schedule I or II substances, cannabis retailers cannot deduct rent, payroll, utilities, marketing or insurance — only cost of goods sold. They are effectively taxed on gross income rather than profit, producing effective federal rates well above the 21% statutory corporate rate. Legalization would lift that burden and likely pull more businesses fully into the formal sector. Without data to model the shift, the report excludes it.
The third is straightforward non-compliance. Some cannabis businesses operate on a cash basis outside the tax system entirely. Legalization would bring some fraction into compliance. Again: unmeasurable, so uncounted.
The report is equally candid about uncertainty running the other way. Consumer response to a 15% price increase could be larger or smaller than modelled, and the size of the post-legalization tax base genuinely isn’t knowable in advance. What we do know from state-level experience is that legalization tends to reshape consumption patterns more than it inflates them outright — as seven years of California data showed, the shape of use changed considerably more than the headline numbers did.
Why Anyone Outside America Should Care
The US isn’t the only jurisdiction staring at this arithmetic. Germany, Canada, Uruguay, Portugal, Colombia and a growing list of others have each had to decide how hard to tax a newly legal product without accidentally subsidising the dealer down the street. The Yale report is, in effect, a rigorous illustration of the universal trade-off: tax too lightly and you leave revenue on the table; tax too heavily and the illicit market wins on price and keeps its market share.
It is also a reminder of how much of the cannabis debate runs on numbers that are far softer than they look. $57.9 billion is a specific, confident-sounding figure. The report that produced it spends most of its length explaining why you shouldn’t treat it as a promise.
FAQ
Not necessarily. The modelled excise tax alone would raise the tax-inclusive price of an average gram by about 15%. On the other hand, lifting Section 280E and opening up banking and interstate commerce would cut operating costs substantially for legal businesses. Which effect dominates would depend on how competitive the legal market becomes and how much of the illicit supply converts.
No. The April 2026 order moved FDA-approved marijuana products and state-licensed medical cannabis into Schedule III, and opened a hearing process on broader rescheduling. Marijuana itself remains a Schedule I controlled substance, and everyone handling it remains subject to the full range of federal controls and criminal penalties. Making recreational use legal without a prescription would require congressional action.
Because prices in cannabis markets have proven volatile, while the substance being regulated has not. A potency-based tax produces revenue that tracks actual consumption, stays stable when retail prices fall, and targets the compound responsible for the social cost — the same logic behind taxing spirits by alcohol content rather than by bottle price.
It’s a careful projection, not a forecast. The authors flag an unusual degree of uncertainty because much of the activity being taxed is currently illegal, making the future tax base genuinely unknown. They also exclude several revenue channels they believe are real, which suggests the estimate is conservative. Treat it as a well-reasoned order of magnitude rather than a precise figure.
Disclaimer
This article is an informational overview of a published economic analysis and does not constitute legal, tax, financial or medical advice. Cannabis laws differ substantially between countries, and within the United States between federal and state law; activity that is permitted in one jurisdiction may carry criminal penalties in another. Nothing here should be read as encouragement to purchase, cultivate, possess or use cannabis. Readers should consult qualified professionals in their own jurisdiction before making any legal, tax or business decisions.
