Will Stablecoins Drain Bank Deposits? The $6 Trillion Debate

On 15 September 2026 the US Senate voted 49 to 50 on whether even to begin debating the crypto market structure bill. It fell well short of the 60 votes required. The bill did not fail on token classification or exchange rules. It failed, in large part, because banks and crypto firms could not agree on one narrow question: should a company be allowed to pay you something for holding a dollar stablecoin? Behind that question sits a bigger one that every finance ministry and central bank is now asking. If digital dollars become a normal place to park money, what happens to bank deposits, and to the loans those deposits fund?

The numbers thrown around in this debate range from about $2 billion to $6 trillion. That spread is not noise. It reflects completely different assumptions about who holds stablecoins, what the issuers do with the reserves, and whether holders get paid. This piece lines up the main estimates, explains why they disagree, and argues that the real risk sits less in Main Street America than in emerging-market banking systems where a dollar token is a genuine upgrade on the local currency.

Why deposits matter more than the headline number

A bank deposit is not just money sitting in a vault. For most banks it is the cheapest funding they have, and a checking account that pays close to nothing is the cheapest of all. Banks lend those balances out as mortgages, small-business credit and farm loans. If customers move money into a stablecoin, the issuer typically buys short-dated Treasury bills with the reserves. The dollars do not vanish, but they stop sitting on a bank balance sheet as low-cost funding.

That is the mechanism banks fear. It is also why the debate is less about total stablecoin supply, which DefiLlama put at about $306 billion as of October 2026, and more about which deposits move, and how fast.

The estimates, side by side

The most quoted figure comes from the Treasury Borrowing Advisory Committee, a group of bond dealers that advises the Treasury. Its April 2025 presentation on digital money put US transactional deposits at about $6.6 trillion and identified them as the deposits most exposed to substitution, because they already pay little or no interest and move easily. It also used a Standard Chartered projection of a roughly $2 trillion stablecoin market by 2028. The $6.6 trillion is a pool at risk, not a forecast of outflows, but it is often quoted as if it were.

Bank of America chief executive Brian Moynihan turned that pool into a warning on the bank's January 2026 earnings call. According to The Block, he said up to $6 trillion, roughly 30 to 35 percent of US commercial bank deposits, could move into stablecoins if they are allowed to pay interest, citing Treasury work.

Read the table from top to bottom and the estimates fall by three orders of magnitude. They are not really disagreeing about the same thing. The top rows describe how much money could theoretically move. The bottom rows try to measure the damage once you account for where the money goes next.

Where the money goes after it leaves

This is the part most coverage skips. When a household swaps $10,000 of deposits for a stablecoin, the issuer receives the $10,000. If it buys Treasury bills, the seller of those bills receives the cash and usually deposits it at a bank. The deposit has changed owner and possibly changed bank. It has not left the system.

The Council of Economic Advisers paper published in April 2026 builds its whole argument on this recycling. It estimates that banning yield entirely would raise bank lending by only about $2.1 billion, around 0.02 percent of loans, while costing users about $800 million a year in forgone returns. Its baseline assumes stablecoins equal about 1.7 percent of US deposits. Only the slice of reserves held as cash at banks, which the paper puts at about 12 percent for USDC, is fully siloed. To get a large lending hit, the authors had to stack several assumptions they themselves call implausible, including stablecoins growing to six times their current share and the Fed abandoning its ample-reserves framework.

The Kansas City Fed analysis by economist Stefan Jacewitz takes a less relaxed view. It estimates that each dollar moving from deposits into regulated stablecoins reduces bank lending by about 50 cents and raises net Treasury holdings by about 30 cents, because issuers hold roughly half their assets in government paper while banks hold about half in loans. Scale the stablecoin market from about $250 billion to $900 billion with deposit funding and you get roughly $325 billion less in bank loans.

The aggregate number is reassuring. The distribution is not.

Both papers can be right at once. The CEA is measuring the whole US system. Jacewitz is pointing at the composition of balance sheets. Even when deposits come back to the banking system, they tend to come back to large banks that serve Treasury dealers and custodians, not to a community lender in Kansas. Standard Chartered's January 2026 note, reported by The Block, made exactly this point: regional banks, which rely most on net interest income, are the most exposed, while investment banks are the least.

A worked example: one community bank

Here is an illustration using round numbers, not a forecast for any real bank. Take a community bank with $1 billion in deposits, of which 30 percent, or $300 million, sits in non-interest checking accounts.

1. Assume 5 percent of those checking balances, $15 million, moves into stablecoins over two years as local businesses start paying suppliers on-chain.
2. Apply the Kansas City Fed ratio of about 50 cents less lending per dollar moved: the bank's lending capacity falls by roughly $7.5 million.
3. If the bank replaces the lost $15 million with wholesale funding at about 3.65 percent, the reserve rate used in the CEA paper, its annual funding bill rises by about $547,500.
4. On a bank that might earn $10 million to $12 million a year, that is a visible but survivable hit to margins.

Now change one variable. If stablecoins could pay 3.5 percent, the competitive yield the CEA assumes, the share of checking balances that moves would plausibly be much larger than 5 percent, because the customer is choosing between roughly zero and 3.5 percent rather than between zero and zero. That is the entire reason banks have spent so much effort on the yield question.

The yield fight in Congress

The GENIUS Act, signed on 18 July 2025, already bars stablecoin issuers from paying holders interest or yield. Banks argued this left a gap: exchanges and wallet providers could still pass on rewards funded by the issuer's reserve income. The fight moved to the CLARITY Act, the broader market structure bill.

On 1 May 2026, Senators Thom Tillis and Angela Alsobrooks released compromise text that, as CoinDesk reported, bans returns on balances that are economically equivalent to deposit interest but permits rewards tied to genuine activity, much like card rewards. Treasury and the CFTC would have a year to write the detailed rules. Coinbase accepted it. The big bank trade groups said it did not go far enough and ran a letter campaign for tighter language.

The Senate Banking Committee advanced the bill 15 to 9 on 14 May. Then, according to DLA Piper's summary, the 15 September cloture vote failed with three Republicans joining Democrats, with objections spanning ethics rules, enforcement and state authority as well as yield. The practical result: the GENIUS Act's issuer ban stands, while the treatment of exchange and wallet rewards remains unsettled at the federal level.

The real exposure is in emerging markets

In the US, a stablecoin is mostly a dollar competing with another dollar. In Lagos, Cairo or Colombo it is a dollar competing with a currency that may have lost a large share of its value in a few years. That changes the calculus completely.

Standard Chartered's October 2025 report, covered by Reuters via BNN Bloomberg, estimated that stablecoin savings in developing economies were about $173 billion and could reach $1.22 trillion by the end of 2028, implying up to $1 trillion leaving emerging-market banks. It identified 16 high-risk countries, including Egypt, Bangladesh, Sri Lanka, Kenya, Turkey, India, Brazil and South Africa. Even the full $1.22 trillion would equal only about 2 percent of deposits across those 16 countries, but deposit flight rarely spreads evenly. It concentrates in dollar-hungry savers during a currency scare.

Moody's, in a September 2025 report, called this "cryptoization": dollar tokens weakening a central bank's grip on rates and the exchange rate, and banks losing deposits as savers shift into wallets. It noted that fewer than a third of countries have comprehensive digital asset rules. This is why central banks such as the RBI have been far more cautious than Washington, and why several emerging economies are studying tokenised deposits and CBDCs as a domestic answer.

What banks are doing about it

Banks are not only lobbying. Many are building. The defensive play is the tokenised deposit: a bank liability that moves on a blockchain but stays on the bank's balance sheet and inside deposit insurance and capital rules. Others are positioning as reserve custodians for issuers, which keeps at least part of the reserve money as deposits. A third option is to issue a stablecoin through a subsidiary under the GENIUS Act framework. Each of these keeps the funding inside banking, and that is the real goal.

Our view: the systemic US risk is modest as long as issuers cannot pay yield and reserves recycle through the banking system. The distributional risk for regional banks is real but slow. The acute risk is in emerging markets with weak currencies and open crypto access, where a dollar token offers something the local bank cannot.

What to watch next

• November 2026: Comptroller Jonathan Gould said the OCC would publish its final GENIUS Act rule by November. Watch how it treats payments to holders by affiliates and partners. See our GENIUS Act implementation tracker.
• Post-midterm session, November-December 2026: whether Senate leaders try another vote on CLARITY before the 119th Congress ends in January 2027. If they do not, the bill must restart in the new Congress.
• 18 January 2027: the latest date the GENIUS Act can take effect, 18 months after enactment, unless final regulations arrive earlier and trigger the 120-day clock.
• Quarterly bank results: regional bank non-interest deposit balances are the single best real-world test of whether outflows are showing up.
• EM central bank reports: any data from the RBI, Central Bank of Egypt or Central Bank of Kenya on dollar stablecoin holdings would sharpen the emerging-market estimate considerably.

This article explains research and policy debates. It is not investment advice. For how issuers earn from reserves, see how stablecoin issuers make money.

About this researchMuffinPay InsightsIndependent analysis of stablecoins, CBDCs, cross-border payments and digital public infrastructure. Every figure is linked to its source. Read our editorial standards →