The House Majority Leader Called Cannabis a ‘Gateway Drug.’ The Federal Data Out the Same Week Told a Different Story.

House Majority Leader Steve Scalise was asked this week about Louisiana’s marijuana-arrest rate, one of the highest in the country. His answer reached back fifty years: cannabis is a “gateway drug” that “causes other problems down the road.” It’s a line that has outlived nearly every study built to test it.

The talking point and the paperwork arrived in the same news cycle, and the paperwork didn’t cooperate.

The Government Accountability Office published a new report on cannabis-business banking access this week — the federal government’s own audit of an industry it still classifies as criminal. GAO found what dispensary owners have said for a decade: getting a bank account remains genuinely hard when your legal, state-licensed revenue is federally untouchable. But GAO also found something that cuts against the enforcement framing Scalise reached for: no indication that any bank has actually been penalized solely for serving the marijuana industry. The chilling effect, in other words, is mostly banks guessing at a risk that federal regulators aren’t actually enforcing.

Sitting next to that finding, in the same week’s policy roundup, was a NORML analysis of FBI arrest data with a harder edge: in several states, marijuana possession accounted for more than half of all drug arrests last year. Not a majority of cannabis enforcement — a majority of all drug enforcement, full stop. That’s not what a system built around fentanyl, meth, and the opioid death toll looks like when it’s using its resources well. It’s what a system looks like when the easiest arrest is still the one it reaches for.

The “gateway drug” theory itself has had a rough run in the actual research literature. NIDA’s own research summary is careful on this point: most people who use cannabis never go on to use other drugs, and the risk factors that predict cannabis use — trauma, peer environment, mental health, access — are largely the same risk factors that predict use of anything else. A foundational 2012 paper in Drug and Alcohol Dependence went further, arguing the “common liability to addiction” model — a shared, measurable vulnerability rooted in genetics and environment — explains the sequencing researchers observe far better than a pharmacological gateway effect does; drug-use initiation order, the paper found, is “variable and opportunistic,” not the fixed staircase the gateway theory assumes. That reframe matters here, not as a debate point, but as a resource-allocation one: if the gateway theory is doing the work of justifying where arrest dollars go, and the theory itself is contested science, the dollars are moving on momentum, not evidence.

Not every state is moving the same direction. Sen. Tim Sheehy of Montana broke from his own party’s push for a blanket federal hemp ban this week, arguing for “commonsense guardrails” instead — regulation, not prohibition, aimed at the synthetic and unregulated products actually causing harm. That’s the more honest fight: not whether cannabis is a gateway, but whether enforcement dollars and banking rules are pointed at what’s actually dangerous, or just at what’s easiest to count.