GAO traces the cost of a cannabis bank account to dedicated staff and monitoring systems

Back out the denominator and the number gets more useful: a thousand filers at 11% of insured depositories implies about 9,100 institutions in total, leaving roughly 8,100 that filed nothing on a cannabis transaction in all of 2024 ^[4][6]. GAO's own caveat pushes the same way, since some of the thousand were logging occasional transactions rather than running accounts ^[5], putting 11% at the outer bound of engagement rather than a census of cannabis deposit franchises.

Then there is who GAO talked to: seventeen focus groups, nine with financial institutions and eight with cannabis businesses, drawing on 74 financial-institution participants and 51 owners and managers ^[7], with five of the nine institution groups including firms that already serve the sector ^[8]. So the finding that existing cannabis bankers were less likely to say a safe harbor would materially change their operations ^[14] comes substantially from firms that have already paid for dedicated staff, special transaction monitoring and customer due diligence ^[10], and that hold a scarce permission the same report shows is priced accordingly ^[2]. "It would not change what we do" is a compliance answer and a competitive one at once.

My read is that legal cover moves the count of banks while compliance cost sets the price, and those are separate levers. Institutions that stay out named regulatory risk, including adverse supervisory action and the conflict between federal and state law ^[9][11], which is precisely what a statute can speak to, and many of them told GAO a safe harbor might make them more willing to serve some cannabis businesses ^[14]; GAO sets reducing BSA burden alongside that as potentially equally important ^[15]. Here is the more interesting version of the counter-thesis: if program cost per account is small next to the spread and fee income a cannabis deposit earns, then supervisory exposure was the whole barrier and a safe harbor clears it. What the report supplies is the inputs, staff and systems and due diligence ^[10], not dollars per relationship, so that comparison cannot be run on this evidence.

The falsifier is pricing. If a safe harbor passes and account fees and business-loan rates for plant-touching operators drift toward ordinary small-business terms ^[2], legal risk was the binding constraint; if the filer count climbs while pricing holds, the fixed cost of the program was. Twenty-four states and the District of Columbia allow recreational use while federal law does not ^[16], and regulators still leave each bank to size its own appetite ^[17]. Appetite gets priced case by case, not set by statute.