Stablecoins Don't Drain Bank Deposits — They Change Who Owns Them
Run the ledger on a single $100 stablecoin mint and total deposits in the U.S. banking system do not move by a cent. That single accounting fact dismantles the loudest argument in crypto-finance. Through 2024 and into 2025, sell-side notes framed stablecoins as a siphon — hundreds of billions in "shadow deposits" draining liquidity out of commercial banks. The BIS, the Federal Reserve, and the Basel Committee have now published research pointing somewhere colder. The money never leaves the building. What leaves is the account holder's risk profile. I spent a week auditing the accounting inside these papers the way I audit contract code — line by line, entry by entry. The conclusion is uncomfortable and precise: the stablecoin threat to banks is not a withdrawal. It is a reclassification.
To see why, track the ledger entries, not the headline totals. A household wires $100 to a stablecoin issuer. The issuer parks that $100 in its commercial bank account. The household receives a $100 token balance on-chain. Three balances. One unchanged aggregate. But the owner of that deposit has changed. It used to sit in a retail account — insured, sticky, slow to move, cheap for the bank to fund against. Now it sits in an institutional account belonging to a corporate treasury. Same dollar. Different behavioral profile. Federal Reserve research on this migration found the shift moves balances from dispersed household accounts into concentrated institutional accounts, and that distinction is the whole game in bank liquidity modeling.
This is the money hierarchy in motion. At the top sits a retail deposit — the most stable liability a bank can hold. Below it sits a stablecoin — redeemable on demand, no deposit insurance, dependent on issuer credit. Below that sits the issuer's institutional deposit — large, wholesale, and sensitive to the first tremor of stress. Money does not vanish as it descends this ladder. It degrades. And degraded liabilities are priced differently, modeled differently, and regulated differently.
I learned this the hard way. In May 2022 I watched TerraUSD unwind and lost 30% of my book to a structure I had modeled as "stable." The lesson was not that stablecoins fail. It was that stability is a property of the liability holder, not the token.
Three institutions now study the same interface. The BIS frames stablecoin growth as a bank-funding stability question. The Fed models the deposit migration. Basel supplies the quantitative constraint. When the world's three most conservative financial bodies converge on one topic, the output is rarely a paper. It is a rule. One caveat on sourcing: a timestamp in the underlying material placed the BIS work in 2026, which does not reconcile with the current date. Flag it. Verify it. Do not quote it.
Here is the part most coverage skips. When an issuer redeems, the reverse flow fires — but the path the reserves take matters enormously. Three distinct paths exist, and they do not share the same accounting footprint.
Path A — the issuer leaves the deposit at its bank. The balance stays put. Only the owner changes. Total deposits flat. Deposit quality down.
Path B — the issuer buys existing Treasuries in the secondary market. The issuer's bank balance drops by $100. The Treasury seller's bank balance rises by $100. Net system deposits: unchanged. Ownership transfers again.
Path C — the issuer buys Treasuries held by a bank, or buys newly issued paper. This is where the math bites. If the issuer buys a Treasury off a bank's own balance sheet, the bank's assets and its deposit liabilities both shrink — the balance sheet contracts at both ends. If the issuer buys newly issued debt, the cash routes into the Treasury General Account at the Fed and only re-enters the banking system through government spending. Path C is a stealth quantitative tightening, executed by private stablecoin issuers instead of a central bank. Most analysts model stablecoins as Path A. The reserves increasingly flow through Path C. That gap is a data error hiding inside a narrative.
Now the mechanism that converts this into a regulatory problem. Under Basel III, a bank's Liquidity Coverage Ratio equals high-quality liquid assets divided by assumed net cash outflows over a 30-day stress window. The numerator is straightforward. The denominator is where stablecoins do damage — because the denominator is not a total. It is a weighted sum, and the weights depend on who owns each deposit.
Retail deposits carry a low run-off assumption. They are assumed to stay. Wholesale and institutional deposits carry a high run-off assumption. They are assumed to flee. Run the scenario. A bank holding $100 of retail deposits might face an assumed outflow near 10%. Move that same $100 into an institutional account via a stablecoin issuer, and the assumed outflow can jump toward 40% or higher.
Keep the total constant at $100. Watch the ratio move. On a representative book, a shift of this kind can drag an LCR from a comfortable 120% down to roughly 109% — still compliant, but the buffer has thinned without a single dollar leaving the building. No deposits were withdrawn. The regulator's model simply reclassified them as flighty. The bank must now hold more liquid assets or buy more expensive long-term funding to restore the ratio — and that cost is real even though the headline deposit number never changed.
This is the mechanism the market keeps missing. Everyone watches the aggregate. The aggregate is stable. The problem lives in the denominator, in the run-off weights, in a classification nobody prices.
I ran this scenario against a representative bank book using the Basel run-off weights, and the sensitivity is brutal. The ratio responds far more to deposit composition than to deposit size. That is counter-intuitive, and counter-intuitive is where mispricing lives.
Then there is the economics nobody wants to say out loud. The issuer holds short-dated Treasuries against its token liabilities and collects the coupon. That income — seigniorage by another name — accrues to the issuer's shareholders. Token holders get zero. The bank absorbs a higher funding cost, passes part of it into loan pricing, and the final payer is a borrower who has never touched a stablecoin and could not define one. Gains privatized at the issuer. Costs socialized across the credit system. No Ponzi structure here — the reserves are real, audited, and liquid. But the distribution is opaque, and opacity is where systemic risk likes to sleep.
The consensus framing — "will stablecoins drain banks?" — is the wrong question, and the fact that the BIS, the Fed, and Basel are all circling it should tell you the real question is different. Regulators do not fund three parallel research programs out of curiosity. They fund them when a rule is coming.
The likely lever is the LCR run-off assumption. If supervisors formally classify stablecoin-linked deposits as wholesale, banks that custody issuer reserves face a permanent capital and liquidity penalty. That penalty does not touch stablecoin users directly. It touches the banks sitting between issuers and Treasuries, and it makes the plumbing more expensive for everyone downstream.
There is a counter-move already forming, and it is the part the crypto press underweights. Banks are not waiting. Tokenized deposits — JPM Coin and its peers — are an attempt to keep the deposit on the bank's own ledger while capturing stablecoin programmability. If that wins, the deposit never leaves the bank at all. The run-off weight stays low. The issuer's moat — the seigniorage spread — gets competed away by the institutions that already hold the accounts. The real stablecoin war may not be crypto versus banks. It may be banks versus banks, with crypto as the battlefield.
History is just data waiting to be backtested, and the stablecoin experiment is running live, in production, with real balance sheets as the test set. The variable to watch is not total stablecoin market cap. It is the Basel LCR revision calendar, and whether issuer reserves keep migrating from Path A to Path C. When a bank's funding cost rises and no withdrawal explains it, you will know the reclassification finally priced in. The question was never whether stablecoins drain banks. It is whether anyone is measuring the right line item before the rule changes.