The Bank for International Settlements' 2026 analysis, "Anchoring Trust: Money, Innovation Beyond Stablecoins," demonstrates with a $100 example that stablecoin purchases can leave total bank deposits unchanged while making bank funding structurally more expensive. The finding reframes the industry's "trillions leaving banks" debate: the dollars return, but they return with a different, more demanding owner.

The mechanics are simple. A customer spends $100 of a bank balance on newly issued stablecoins. The issuer deposits those dollars at its own banking partner. The banking system's aggregate balance sheet looks identical, yet the original bank has lost a sticky retail relationship and gained nothing, while an issuer's treasury function — a single counterparty that can move the entire balance with one decision — now holds the funds.

Why does the deposit's owner matter more than its size?

Federal Reserve research, published as a FEDS note titled "Banks in the Age of Stablecoins: Implications for Deposits, Credit and Financial Intermediation" on Dec. 17, describes exactly this conversion of scattered household balances into large institutional accounts. Bankers call the first kind retail funding and the second wholesale funding. Wholesale balances can leave a bank even when the bank is perfectly healthy, because stablecoin issuers must honor redemptions and may need to withdraw reserves to pay token holders elsewhere.

Regulation makes this costly in a precise, measurable way. Under the Basel framework's Liquidity Coverage Ratio, banks compare liquid assets against estimated net cash outflows over a 30-day stress window, with different deposit categories carrying different runoff assumptions. A bank holding $120 million of qualifying liquid assets against $100 million of estimated outflows posts a 120% ratio. A customer mix shift that pushes estimated outflows to $110 million — with the same assets and no actual withdrawal — drops the ratio to roughly 109%.

That arithmetic, while illustrative, explains why funding desks cannot shrug at stable aggregate numbers. The bank may need more liquid assets or longer-term funding, and both cost money that can ultimately reach loan pricing — including borrowers who have never heard of a stablecoin.

Do Treasury purchases actually remove dollars from banking?

Not automatically. When an issuer buys an existing Treasury bill from a nonbank investor, the issuer's bank balance falls by $100 and the seller's rises by $100. The deposit simply changes hands; its future stickiness is unknown.

The accounting differs by counterparty:

• Buying a bill the bank itself owns extinguishes a deposit liability and shrinks both sides of the banking system's balance sheet, while removing a security the bank might have held for its own liquidity needs.
• Buying newly issued government debt routes payment to the Treasury's account, with government spending later returning money to the private sector.
• Secondary-market stablecoin purchases don't create new reserves at all — the buyer's payment goes to the selling token holder.

BIS General Manager Pablo Hernández de Cos put reserve composition at the center of the banking question in an August speech, "Pushing the Monetary Frontier: Stablecoins and Tokenised Deposits." The route the backing takes determines the effect on banks, which is why token-supply forecasts alone cannot predict lending impact.

There is also a distributional problem. A smaller lender can lose a customer's deposit while the issuer's larger banking partner captures the replacement account. The national total is unchanged, but the receiving bank is under no obligation to extend the same loans to the same local borrowers.

How are banks responding?

A separate Fed study published May 1, "Banks in the Age of Stablecoins: Lessons from Their Historical Responses to Financial Innovations," examined how banks adapted to money-market funds and earlier payment platforms. The options include paying higher deposit rates, improving payment services, replacing deposits with longer-term borrowing, tokenizing deposits on a blockchain while preserving the customer's claim on the bank, or issuing stablecoins directly.

The last route now has a regulatory track. The Fed's Sept. 24 proposals would set reserve and risk-management standards for payment stablecoin issuers under its supervision, alongside an approval process for supervised banks seeking to issue through a subsidiary. A key constraint: money committed to redeeming tokens cannot be treated as ordinary funding available on identical terms for long-term lending. Owning the issuer does not erase the promise to token holders.

Neither the BIS nor the Fed claims stablecoins have already forced lending cuts. Demonstrating that would require evidence from the banks involved — how they replaced deposits and what happened to their loan books — plus data on where buyers sourced the money, since new dollar demand from abroad need not mirror domestic deposit migration.

The competitive pressure itself is not a harm to correct. Faster payments carry real value, and banks competing harder for customers can benefit users even if dependable funding gets more expensive. The operative question for the coming rulemaking cycle is how much of that cost banks absorb themselves and how much reaches the next borrower walking into a branch.