Higher deposit rates reliably slow stablecoin growth, and banks reprice deposits far faster once policy rates climb above 3%, according to research published by the Bank for International Settlements (BIS). The study also finds little evidence that decentralized finance (DeFi) yields shielded stablecoins from deposit competition, while the question of who may pay a return to stablecoin holders remains unsettled.

BIS Working Paper 1382, published on October 13, is by Rashad Ahmed of the Andersen Institute for Finance and Economics and Iñaki Aldasoro of the BIS, and it states that its views are not necessarily those of the BIS. The authors use weekly US data from January 2019 to June 2025, a stretch in which stablecoin market capitalization grew from under $5 billion to more than $220 billion.

What the paper found

Against average ten-week stablecoin growth of 15%, a one-basis-point rise in the national average of six non-maturity deposit rates in a single week predicts about 3 percentage points less growth, taking it to roughly 12%. The authors argue that figure is biased toward zero, because banks raise rates when stablecoins pull funds away. Correcting for that with an instrumental variable, the effect is about three times larger: roughly 9 points, or about 6% growth.

A basis point is 0.01 percentage points, but it is a large move for a single week: about 1.2 standard deviations of the paper's weekly series. On the smaller estimate, a one-standard-deviation rise cuts ten-week growth by about 3.75 points. A full growth cycle measures 20% over ten weeks, according to the paper.

Growth slows without reversing. Canceling average growth on the smaller estimate would take a five-basis-point rise in a single week, a six-standard-deviation event. In the 2022–23 tightening cycle, stablecoin market capitalization did fall about 30%, which the authors say is consistent with deposit repricing accumulating alongside other pressures, including Fed hikes and the Terra and FTX collapses.

The paper flags its own limits. It uses aggregate data, so it "cannot identify which holders substitute" or which banks lose funds. It covers one high-rate stretch. Its test statistic of 9.78 falls slightly below the conventional threshold of 10, which suggests some weakness. And the stablecoins in its sample paid holders no interest.

Why 3% matters, and why now

The authors' key step is a kink in how banks price deposits. Below about 3% on the fed funds rate, deposit rates barely respond to policy changes. Above it, banks reprice more aggressively as depositors' alternatives become salient. Treasury bill rates, by contrast, track the fed funds rate in a straight line, so the kink is specific to deposits. The threshold the authors use is 3.125%, the midpoint of a 3.00–3.25% target range.

The Fed raised rates in September, its first hike in three years, to a target range of 3.75% to 4.00%, back above that line. Pass-through is typically thin: the paper cites estimates that only 6% to 8% of fed funds changes reach demand deposits. In its sample, the average deposit rate was 0.26%, while three-month Treasury bills averaged 2.64%.

What a saver can earn as of October 2026 is framed by FDIC average and top rates reported by Fortune and high-yield savings and CD averages from Curinos via Experian.

Banks have done this before

The Federal Reserve's May note on banks and stablecoins points to a precedent. Banks "knew how to compete with MMFs – raise deposit rates," it says, but regulation capped what they could pay until Congress changed the rules. The note adds that banks have lobbied for limits on stablecoin interest and similar yield substitutes.

• 1982 — money market funds. Competitor: money market funds, about $220 billion by year-end, roughly 15% of bank deposits. Ob: savings rates capped at 5.25%. Fix: money market deposit accounts, authorized in December. Result: more than $300 billion in three months.
• 2026 — stablecoins. Competitor: payment-ready stablecoin balances, with exchanges paying the yield. Obstacle: matching rates costs margin. Fix: still open, including whether platforms may pay stablecoin rewards. Result so far: slower growth when deposit rates rise.

What a higher rate costs a bank

Cheap deposits fund a bank's loans and securities, so paying more keeps them but narrows the margin. On $1 billion of deposits, the gap between the 0.37% national average and the 4.50% top online rate is about $41 million a year: $45 million against $3.7 million, before any change in balances. On $10,000 it is $450 against $37. That is an illustration, not an estimate of what any bank would pay or keep.

Repricing every account can also pay customers who would have stayed anyway. The paper's tier test gets at that: when rates on preferred accounts of $25,000 or more rise relative to small savings accounts, stablecoin market capitalization falls by about 0.7% over ten weeks. The effect is much smaller because the test captures only part of repricing, but it points the same way. It also appears in certificate of deposit rates (−2.2% on the smaller estimate, −15.5% on the corrected one), which the authors flag as relevant to community banks that rely on time deposits. The Fed note says banks are already responding with more targeted rate strategies.

Who switches, and what yield doesn't change

Stablecoins do not respond equally. Over ten weeks, on the smaller estimate, a one-basis-point deposit rise is followed by about 5% less growth in USDC and about 2% less in USDT. The authors link USDC to US institutional and DeFi users, who are likelier to move between deposits and tokens, and USDT to offshore activity and developing economies.

A common objection is that exchange and DeFi yields make bank rates the wrong comparison. The paper controls for DeFi yields and finds little evidence that they shield stablecoins. Balances lent through DeFi protocols also fell when deposit rates rose, though imprecisely, and only a small share was lent at all: a median of 12.8% of USDC and 4.1% of USDT.

User-facing yields tend to come from exchanges, because issuers' direct interest payments are prohibited in most jurisdictions, a June BIS bulletin notes. It describes two models. Coinbase's USDC yield follows the policy rate because it passes on reserve income. Binance's USDT yield exceeded 20% in 2024, when borrowing demand pushed the borrowing rate to 40–50%. That return may also pay for counterparty risk: customer balances can fund an exchange's own activity, as the FTX collapse showed.

Where the money goes

Money that leaves one bank can come back to another. The Fed note explains that an issuer may hold reserves as bank deposits, or buy Treasury bills from dealers who deposit the proceeds, so stablecoins may change who holds deposits more than how many exist. The paper stresses that it does not show stablecoin growth shrinking aggregate deposits or lending.

The bank that loses the customer may not be the one that gets the money back, and the mix changes: retail and commercial deposits can be replaced by larger wholesale balances from stablecoin issuers, which the note says has implications for funding stability and liquidity management.

Banks are also building products. In the Fed's September 2025 survey, about half of large-bank respondents planned to prioritize issuing tokenized deposits, and about 40% reported plans to hold reserve assets for stablecoin issuers. One example is Citi's push into tokenized deposits for corporate payments.

Who is allowed to pay

The paper ends on regulation. If stablecoins come to be seen as digital cash, the authors write, adoption is sensitive to what banks pay depositors, so banks keep some influence over its pace. They name two levers for regulators: requiring issuers to hold some reserves in bank deposits, which supports liquidity but concentrates uninsured deposits in a few banks, and deciding whether issuers may pay holders.

The fight now is over platform rewards. Community banks objected to stablecoin rewards in the dispute that helped sink the CLARITY Act in the Senate on September 15, as AMINA reports, and the Senate has not settled it.

Before moving cash, savers may want to ask:

• Where does the yield come from: reserve income or exchange activity?
• Can the provider lend your balance, as exchanges can?
• Is it a bank deposit, insured up to $250,000 per institution?
• Do you need the tokens for payments or trading, or are they idle?

This article is provided for informational purposes only and does not constitute financial or investment advice. Research estimates describe historical relationships and do not guarantee future outcomes.