On September 4, Federal Reserve staff released a research note that contained no enforcement action, no legislative demand, and no dramatic language. It simply asked a question the industry has spent a decade avoiding: if a dollar sits inside a Treasury bill owned by Circle, and a user holds USDC representing a claim to that dollar, how many dollars now exist?
The correct answer, the note observes, may be two. And that is the problem.
The paper is part of the Fed’s Finance and Economics Discussion Series. It is not policy. It is not even peer-reviewed. It is a staff-level statistical document, the kind normally read by six people and then forgotten. Yet buried inside its discussion of M1 and M2 classification is the clearest official acknowledgment yet that the stablecoin boom has created something the U.S. money supply was never designed to measure: a single reserve dollar doing double duty inside the national ledger.
The crypto market will call this a regulatory catalyst. It is not. It is an accounting warning. Code does not lie; people do. The Fed has just discovered that the code was never the problem.
The context matters because the runway has already been built. The GENIUS Act, the U.S. stablecoin framework now moving through implementation, requires every major issuer to maintain a 1:1 reserve of identifiable assets and to publish monthly attestations. Circle, the issuer of USDC, already does this. Its public reports show roughly $71.8 billion in circulation, backed by cash, U.S. Treasuries, and money-market instruments, verified by a top accounting firm every month.
The industry reads this as maximum transparency. And it is—for a private balance sheet. But the GENIUS Act also assigns the Federal Reserve the statistical authority to decide whether stablecoins belong inside the official money supply. That is where the transparency ends and the complexity begins.
M1 measures the money people can spend today: currency in circulation, demand deposits, and other liquid balances. M2 expands the definition to include savings deposits, small time deposits, and retail money-market funds. Every dollar a bank holds as a reserve, and every dollar an investor holds in a money-market fund, is already sitting somewhere inside those aggregates.
Now ask the question Circle’s monthly attestation never answers: when a stablecoin issuer takes a customer’s bank deposit and converts it into Treasuries and money-market fund shares, which column of the money supply should still count that dollar in? And when the customer receives USDC in exchange, should the new token also be counted?
The Fed staff note identifies the double-count risk in careful bureaucratic language. The underlying logic is more brutal. Consider a simple example. A user transfers $1,000 from a commercial bank account to Circle. That bank deposit was already counted in M1. Circle then moves the funds into a government money-market fund. Money-market fund shares are already captured inside M2. The user now holds $1,000 in USDC. If a statistical compiler simply adds USDC balances to M1 or M2, that original dollar is counted at least twice: once as the reserve asset and once as the circulating token.
The result is not newly created purchasing power. It is repackaged purchasing power wearing a second label. The Fed’s note is essentially saying that the current measurement framework cannot tell the difference between a new dollar and a new wrapper for an existing dollar.
This is not a smart-contract bug. No re-audit of Circle’s code will fix it. The problem is structural: the reserve asset and the stablecoin liability are two different claims on the same underlying dollar. A stablecoin holder is not holding the dollar. The holder is holding a promise that the dollar exists somewhere in a qualified custodial account. As long as that promise is backed by an asset already present in the money supply, adding the token to the money supply is statistical double-spending.
From my own work auditing collateral structures, I have watched this pattern repeat across market cycles. In 2022, when I reconstructed the Terra collapse, the liabilities were fully visible on-chain: every mint and burn was recorded. Transparency did not save Terra because the backing was fiction. The stablecoin problem today is the mirror image. The backing is real, but the classification layer does not exist. I can verify that a reserve exists. No on-chain tool can verify that the same reserve is not being counted twice inside a national balance sheet.
The GENIUS Act’s 1:1 reserve requirement actually deepens this issue. It forces issuers to hold assets that are overwhelmingly the same instruments the Federal Reserve already tracks in its money-supply data. Bank deposits, Treasuries, money-market funds: these are not exotic collateral. They are the core components of M1 and M2. A stablecoin regime built on those assets is, from the Fed’s perspective, a parallel accounting system that copies entries from the official ledger without deleting the originals.
The note also makes classification conditional on what it calls functional, economic use. That phrase sounds like a routine legal test. In practice, it is an oracle problem.
M1 is defined around transaction media. M2 is defined around liquid stores of value. A stablecoin like USDC is used for both. It settles trades in under a second in some venues and sits idle for months in self-custody wallets in others. The same token can be a payments vehicle at 10:00 AM and a savings vehicle by noon. The Federal Reserve cannot observe those distinctions from public blockchain data. On-chain records are event logs, not survey responses. They show that a transfer occurred. They do not show whether that transfer was a purchase of coffee or a movement between two trading accounts.
The deeper problem is geography. U.S. money-supply measures are constructed for the U.S. economy. They estimate how much dollar liquidity is held and used by U.S. residents. Stablecoins, by design, do not respect that boundary. A USDC token held in a self-custody wallet in Tokyo is indistinguishable on-chain from a USDC token held by a corporate treasury in Chicago. Public keys do not carry ZIP codes.
This is the quiet killer inside the Fed note. To include stablecoins in M1 or M2 without double-counting, the Fed needs to know not just how many tokens exist, but who holds them, where they are located, and why they are being held. That data is not on the ledger. It lives inside the know-your-customer files of exchanges and issuers—and it does not exist at all for the growing universe of unhosted wallets.
What the Fed would need is a reporting chain, not a blockchain. Issuers would have to compile standardized breakdowns of circulating supply by geography and by economic use. Exchanges would have to certify holder locations. The entire system would require the kind of survey infrastructure that took the Federal Reserve decades to build for the conventional banking sector. No stablecoin issuer has built that infrastructure yet. No legislative timeline in the GENIUS Act requires them to.
Now the contrarian angle, because the bulls deserve credit for what they have gotten right.
The Federal Reserve does not publish staff papers on technologies it intends to strangle. The very existence of this note is a recognition that stablecoins have moved from crypto-native speculation into the institutional perimeter. That is a legitimacy milestone no amount of double-counting rhetoric can erase.
The bulls are also right that a formal classification would be transformative. If the Fed ever determines that a specific stablecoin is functionally equivalent to a transaction deposit, the token’s legal status shifts. A Howey analysis becomes harder to sustain when the asset is officially categorized as a component of the U.S. money supply rather than as an investment contract. Institutional treasurers who currently treat USDC as a risky crypto asset would be forced to reconsider it as a cash equivalent. The winners would be the issuers with the cleanest reserves and the most disciplined reporting. That points directly to Circle, not to the offshore issuers whose attestations are weaker and whose reserve compositions are less transparent.
But this is where I stop being charitable. The same classification that grants legitimacy also imposes a cost most bulls refuse to model.
Inclusion in M1 does not make a stablecoin issuer a hero of decentralized finance. It makes that issuer a statistical component of the U.S. monetary system, with all the supervision that implies. The Fed cannot count a private company’s token as money without treating that company as a quasi-bank. Quasi-banks receive stress tests. They receive capital standards. They receive examination teams. The “we are not a bank” defense dies the moment the Federal Reserve starts netting your reserves against your tokens inside M1.
The deeper illusion is that stablecoin issuers want to be counted as money. They have spent years describing themselves as neutral infrastructure. Once neutral infrastructure becomes a line item in the national money supply, it also becomes a policy tool. The Fed does not adjust interest rates based on the health of private payment networks. It adjusts them based on the health of the economy. If stablecoins matter enough to measure, they matter enough to regulate as instruments of monetary policy.
Audit the promise, not the poster. The promise from the stablecoin lobby is that M1 inclusion is the final step toward legitimacy. The poster is beautiful. The fine print is that being counted as money means being managed as money.
None of this will be resolved in the next quarter. The Fed staff note will be read, cited, and then shelved while the statistical infrastructure is debated. That delay is not neutral. Every month of ambiguity means the market keeps pricing stablecoins as something between a payment token and a digital bond. The double-count risk does not disappear. It just compounds.
High yield is a warning, not a welcome. In this market, though, the warning is not coming from a protocol or a yield farm. It is coming from the quiet statistical desk at the Federal Reserve, which has just demonstrated that the $71.8 billion in USDC everyone treats as digital cash may already be counted inside the U.S. money supply under a different name.
When the next crisis arrives, the blockchain will settle a stablecoin transfer in seconds. The dollar behind that transfer will take three days to move through the banking system. The Federal Reserve will be watching both clocks. The question is not whether stablecoins should enter M1. The question is which half of that settlement fails first—and whether the ledger can tell us in time.
