
Twenty-One Million Arrests Later: The Plant Was Never the Problem the Law Said It Was
I like round numbers about as much as I like being pulled over, so let me start with one that isn’t round: 21,366,700.
That’s how many marijuana arrests American police reported between 1990 and 2025, according to FBI-derived figures compiled by NORML and presented by stupidDOPE’s THE RECORD project. The decade totals tell the story before I do: about 5.1 million in the 1990s, 7.9 million in the 2000s, 6.9 million in the 2010s, and 1.4 million from 2020 through 2025. The single worst year was 2007, with more than 870,000 arrests, nearly half of all drug arrests in the country.
Before anyone tattoos that number on their forearm, the honest caveats, because I’d rather be accurate than outraged. Those are arrest events, not 21 million unique people. They aren’t 21 million convictions or 21 million prison sentences. The FBI also changed its data collection in 2021, which muddies the seam between old and new numbers. The people who compiled the figure say all of this themselves.
Even with the caveats, the scale is obscene. And it isn’t history. In 2025, police still logged 200,960 marijuana possession arrests. Divide that into a year and you get one every 2.6 minutes. I did the arithmetic twice because I didn’t believe it the first time.
This is the question I keep coming back to, because it’s the one the whole enterprise had to answer. If you spend three decades and tens of millions of arrests trying to remove a plant from society, the plant should get scarcer.
It didn’t. Monitoring the Future has surveyed high school seniors since 1975, and in every year through its 2001 report, between 83 and 90 percent said they could get marijuana fairly or very easily. Kids in the middle of the drug war could find weed about as easily as they could find a vending machine.
Now the twist that makes advocates uncomfortable. Once legal, age-gated markets appeared, teen access got harder. By 2024, only 65 percent of seniors said marijuana was easy to get, well below that old 83-to-90 range (with a caveat: the survey moved online in 2021, which can shift how kids answer). Teen use fell too: past-year use among 12-to-17-year-olds dropped from 14.2 percent in 2011 to 8.7 percent in 2025, per the federal NSDUH survey. A licensed store checks ID. A guy named Dre does not.
Here’s the part I won’t skip, and you shouldn’t either. Adult use went up, a lot. Researcher Jonathan Caulkins found that use hit its low point around 1992 and has climbed since, with per-capita past-year use rising 120 percent between 2008 and 2022. For the first time, more Americans reported near-daily marijuana use than near-daily drinking (17.7 million versus 14.7 million). Past-month use among everyone 12 and older reached 15.4 percent in 2024. NIH researchers, using survey data through 2013, estimated that about 3 in 10 users meet criteria for a use disorder, and the product on shelves today is considerably more potent than what your uncle smuggled through Tijuana.
So no, I’m not going to tell you cannabis is harmless. What I’m telling you is narrower and, I think, harder to argue with: prohibition failed by its own metric. It never made the plant scarce. It just made the consequences fall on people instead of the plant.
You’d like to think a policy that ruined this many lives started with sober, disinterested science. Let’s look at the receipts, with appropriate suspicion in both directions.
In 1994, journalist Dan Baum interviewed John Ehrlichman, Nixon’s domestic policy chief, who had done time for Watergate. Baum reported in a 2016 Harper’s piece that Ehrlichman told him the Nixon White House saw two enemies, the antiwar left and Black people, and that tying marijuana to hippies and heroin to Black communities, then criminalizing both, was a way to disrupt them. His reported punchline: did they know they were lying about the drugs? “Of course we did.”
I’ll be straight about what that is and isn’t. It’s one uncorroborated interview, published 22 years after it happened and 16 years after Ehrlichman died. His children publicly rejected it as inconsistent with the man they knew. Baum says he had no reason to doubt it, and that he left it out of his 1996 book because it didn’t fit the narrative style. So treat it as a striking piece of testimony, not a confession under oath.
But you don’t need the Ehrlichman quote to see the pattern, because the enforcement data does the talking. The ACLU found that Black Americans were 3.64 times more likely than white Americans to be arrested for marijuana possession, despite similar usage rates, in a period covering 6.1 million possession arrests from 2010 to 2018. In every state. In over 96 percent of the counties they examined. Disparity survived legalization in many places; arrests dropped, but the ratio stayed put. Political motive may be arguable. Selective enforcement is not.
The ACLU estimated that enforcing marijuana possession laws cost roughly $3.6 billion a year in its 2013 analysis, and 88 percent of the 8.2 million marijuana arrests between 2001 and 2010 were for simply having it. That’s the visible cost.
The invisible one is the record. I want to be careful here, because “millions carry felony convictions” is a line I see thrown around, and the data doesn’t support it as stated. Arrests aren’t convictions, and plenty of possession cases were misdemeanors, dismissals, or diversions. I can’t give you a clean national count of marijuana felonies, and neither can the 21-million figure.
What I can source is that a record of any kind travels. The Congressional Research Service lists ineligibility for certain federal jobs, federal housing and benefits, professional licenses, and immigration consequences as standard collateral damage of marijuana-related convictions, and the Justice Department has estimated that 70 to 100 million American adults face collateral consequences from arrest or conviction of any kind. Arrests alone, with no conviction, can follow you into a background check.
Notice how legalization itself admits the debt. New York’s 2021 law automatically expunged qualifying marijuana convictions, and Illinois said its law would clear close to 740,000 records (a figure from the Marijuana Policy Project, an advocacy group, so weigh accordingly). You don’t need to expunge a crime you believe was a crime. Expungement is the state whispering “we may have overdone it.”
Here’s the current situation, and it’s genuinely weird. In April 2026, the Justice Department moved FDA-approved marijuana products and marijuana under a state medical license to Schedule III. Everything else, including adult-use cannabis, remains Schedule I, alongside heroin. As of mid-September, no decision has been made on the broader question after the DEA hearing wrapped in July.
So the identical plant is now medicine in one ledger and a Schedule I narcotic in another, depending on who sold it and what paperwork it wore. If that doesn’t tell you the categories were built on politics rather than pharmacology, I don’t know what would.
Let me answer plainly, since you came here for plain.
Did three decades of arrests make America healthier? The data I can find says no. Availability stayed high, adult use rose, and the harm that clearly did spread was legal, not pharmacological: records, lost jobs, lost aid, lost time, distributed unevenly by race.
Is “cannabis was never the problem” true? I’d trim that sentence. Cannabis has real risks, particularly for adolescents, heavy daily users, and anyone driving. But it was never the problem the law said it was, and it was never a problem that justified 21 million arrest events. A society can take a plant’s risks seriously and still conclude that handcuffs were the wrong tool. Alcohol carries risks we regulate rather than criminalize, and we managed to remember that after Prohibition. It took us 13 years to fix that one. This one has lasted more than 50.
The uncomfortable follow-up question, the one I’d put to legislators if they ever returned my calls: if the record is this bad and the fix is this obvious, why is anyone still being arrested for possession in 2026? And why does the plant get to be legal for the licensed while the people who paid for that legality still wait for their names to be cleared?
Time for a rethink? Past time. I’d just like it done with the receipts in hand.
- 21 Million Marijuana Arrests: What Three Decades of Cannabis Enforcement Actually Look Like, stupidDOPE
- Monitoring the Future, key findings (Focus/APA) and 12th-grade availability, 1975–2024 (Statista/NIDA)
- Teen use decline, 2025 NSDUH (Marijuana Moment)
- Caulkins, self-reported cannabis use 1979–2022, Addiction
- SAMHSA 2024 NSDUH
- NIH: marijuana use and use disorder prevalence
- Baum, “Legalize It All,” Harper’s; CNN coverage; Ehrlichman family response
- ACLU, A Tale of Two Countries (2020) and The War on Marijuana in Black and White
- Congressional Research Service, collateral consequences
- Gersten, Rocha, Stiefel, Cannabis record clearance (Moritz College of Law)
- Marijuana Policy Project, criminal justice
- DOJ, Schedule III order (April 2026) and Cannabis Business Times, rescheduling status
