Fed staff: stablecoins could be counted in M1 or M2
Federal Reserve staff on Sept. 4 outlined steps to add regulated payment stablecoins to M1 or M2, warning reserve overlap and global circulation must be resolved first.
Federal Reserve staff published a Sept. 4 note describing how regulated payment stablecoins could be classified in official U.S. money measures M1 or M2. The paper says accounting adjustments are required before stablecoin balances are added to those aggregates.
The note treats M1 as transaction money-currency and balances readily used for payments-and M2 as a broader measure that includes savings-type assets. Whether a particular stablecoin would enter M1 or non‑M1 M2 depends on its predominant economic use: tokens used mainly for household and business payments could fit M1, while tokens used mainly as stores of value or trading liquidity would fit non‑M1 M2. The paper is staff research and does not change current definitions.
The authors identify three accounting tasks that must be completed before adding stablecoin balances: functional classification, reserve consolidation and geographic isolation. The reserve issue arises because permitted reserves under the GENIUS Act can include bank deposits, Treasury securities and government money‑market funds, some of which are already counted, at least in part, in M1 or M2. If an issuer receives dollars, places some into a bank deposit or money fund and issues stablecoins against that backing, including the tokens at face value without consolidation could count the same dollar twice.
The note says overlap must be measured at the issuer level and removed before any net addition is made to M1 or M2. A defensible net estimate would match reported reserve categories against the exact money‑stock components already counted, remove only genuine overlaps and preserve backing assets that sit outside the aggregates.
Geography is a separate measurement task. Dollar stablecoins issued by U.S. regulated firms can circulate globally on public blockchains, and transaction records often lack reliable indicators of holder residency. GENIUS applies to U.S. issuers without distinguishing domestic from international circulation, so additional reporting would be needed to separate tokens held or used inside the United States from global activity.
Measuring economic use requires transaction‑level classification. Raw blockchain transfer counts can overstate payment activity because single smart‑contract transactions often emit multiple transfer events that bundle trading, lending, arbitrage, liquidity provision and settlement. A working paper that examined 2025 Ethereum data involving USDT, USDC and PayPal USD analyzed 141 million transactions and about 593 million event logs; roughly one‑third of transactions generated multiple steps and nearly 60 percent of transfer events occurred inside complex transactions.
The note uses Circle’s USDC as an example of the data challenge. Circle’s July 31 assurance reported 71.826 billion USDC in circulation and reserve assets with a fair value of $71.904 billion. Much of Circle’s reserve is held in the Circle Reserve Fund, an SEC‑registered government money‑market fund that can hold cash, short‑dated Treasuries and overnight Treasury repurchase agreements. Those point‑in‑time figures do not provide the consolidation calculations needed to determine a net addition to M1 or M2.
Market figures show scale differences. Global stablecoins were about $292.1 billion across 73 assets, with USDC near $74.5 billion in market capitalization. By comparison, seasonally adjusted U.S. M2 stood at $23.218 trillion for July 2026. The staff note sets out the three accounting jobs-functional use, reserve consolidation and isolating U.S. circulation-that must be completed before stablecoin balances are folded into official money measures.
The paper states that failing to adjust for reserve overlap and geographic circulation could produce statistical increases that reflect reclassification rather than newly created dollar liquidity. The note is independent staff research and reflects only the authors’ views.








