GENIUS Act may leave stablecoins exposed to blockchain runs

Congress’s Guiding and Establishing National Innovation for U.S. Stablecoins Act, known as GENIUS, requires permitted payment stablecoin issuers to hold one-to-one reserves and submit to new supervisory rules. A Federal Reserve staff paper warns that congestion and higher transaction fees on public blockchains can trigger coordinated redemptions even when reserves are fully backed.

The Federal Reserve staff paper, first published June 2, 2026 and updated Aug. 31, 2026, was written by Federal Reserve economists and carries the standard Board disclaimer. The authors model a stablecoin that is fully and safely backed to isolate a separate source of fragility: rising on-chain fees can reduce usefulness as a payment medium and, under weak network effects, push individual exits into a coordinated run. The paper measures “redemption” primarily as a decline in a stablecoin’s circulation on Ethereum; that decline can reflect cashing out to fiat or migration to another blockchain.

The empirical analysis uses an unbalanced weekly panel of five stablecoins from November 2017 through December 2025. The authors report that a one-standard-deviation increase in gas costs, $10.83 in their specification, corresponded to about a 0.9 percentage-point rise in weekly redemptions when network effects were low. The low-network-effects state covered roughly 7% of observations. The paper uses empty Ethereum blocks as an exogenous capacity shock and finds that such shocks raise gas prices. A matched-transfer test of USDT flows on Ethereum and Tron from May 2020 through December 2025 found that higher Ethereum fees coincided with reallocation toward Tron; in that sample, a $1 increase in lagged gas was associated with 3% to 4% more net matched value moving off Ethereum.

The authors note limits to the measures. The circulation metric captures pressure on Ethereum-based supply rather than a direct count of customers redeeming at an issuer. Matched cross-chain transfers cannot identify every owner or motive. The empty-block design establishes a link from capacity to fees but does not prove that a single congestion event caused every later migration or redemption.

The paper presents distributional evidence on transfer economics. Between 2021 and 2025, the fee-to-value ratio for below-median USDC transfers at the 75th percentile frequently exceeded 100%, while above-median transfers almost never had fees above 5% of transfer value. The pattern indicates that small self-custody users may face fees larger than the value of intended transfers during congested periods.

GENIUS requires reserves in specified liquid assets, public redemption procedures, disclosure of issuer purchase and redemption fees, monthly reporting, examinations and certification, and regulatory standards covering capital, liquidity, diversification, operations and information technology. Treasury published an implementation proposal on Aug. 17, 2026 in the Federal Register on Aug. 18; comments are due Oct. 19. The proposal sets an expected effective date for the issuer licensing framework of Jan. 18, 2027 and an expected date for broader digital-asset service provider restrictions of July 18, 2028. The proposed rule distinguishes uncompensated peer-to-peer transfers, including self-custody, from compensated services such as exchanges and custodians, a distinction that affects which entities carry compliance duties.

Stablecoins are distributed across chains with different fee systems. A recent snapshot placed stablecoin supply at about $147.3 billion on Ethereum, $93.2 billion on Tron and $15.7 billion on Solana. At a calm fee moment in early September, Ethereum gas near 0.127–0.128 gwei and an illustrative ERC‑20 transfer using 65,000 gas units implied a network cost near two cents. Tron and Solana use different fee mechanisms and can be cheaper or more expensive depending on transfer type and account state.

The Fed paper frames its result as a conditional warning rather than a forecast of an imminent run. The paper highlights a potential gap between issuer-level safety, ensured by reserves and supervision, and the price and capacity of the public rails that carry token transfers.