Your Stablecoin Earns Interest. Who's Pocketing It?

A stablecoin yield ban is being sold as a way to protect bank lending. On the evidence, it protects very little. What it does decide is who keeps the interest on your dollars, and the person holding them comes last.

Bank of America paid an average 0.51% on its consumer deposits in the first quarter of 2026. At that rate, a hundred dollars earns fifty-one cents over a year (a blended average across accounts).

Put a hundred dollars behind a stablecoin, assume the reserves earn 3.9%, and the backing earns $3.90. Under the incoming US GENIUS Act, the issuer cannot pay you interest simply for holding or using the token.

The interest still gets paid. The question is who receives it.

Banking groups say restricting stablecoin rewards protects deposits and the loans they support. My reading is that the clearer benefit is protecting cheap bank funding. If distributors are also prevented from sharing the reserve income, issuers and exchanges could benefit too.

That second benefit gets much less attention.

Regulation Q, introduced in 1933, banned interest on demand deposits and capped what banks could pay on savings and time deposits. For much of its early life, market rates sat below the ceilings. As rates rose in the 1960s and 1970s, those ceilings began to matter.

Money market funds offered savers access to market returns outside the capped accounts. By autumn 1982, they held about $230 billion. Banks gained access to uncapped money market deposit accounts that December, and the remaining savings and time-deposit ceilings were removed by 1986.

The usual lesson is that the cap helped build the competitor. But the target of the restriction matters.

Then, the savings ceiling held back banks. Today, banks can pay market rates on deposits. High-yield savings accounts and money market funds already give savers alternatives to a low-paying account.

GENIUS puts the restriction on the payment stablecoin instead. That includes a stablecoin issued by a bank. The bank can pay interest on its deposit product, but the stablecoin issuer cannot pass yield directly to its holders.

The closer historical comparison is Regulation Q's ban on demand-deposit interest. Retail savers had ways around that ban, including interest-paying NOW accounts with cheque-writing access. The remaining prohibition was repealed in July 2011.

Suppose your hundred dollars leaves Bank A to buy a stablecoin. The issuer places the backing in a government money market fund, which buys existing Treasury bills from a dealer. The dealer receives the hundred dollars in its account at Bank B.

Your deposit has become the dealer's deposit. It has moved between owners and perhaps between banks.

The Treasury bills now back the stablecoin and generate its reserve income. The dealer's bank deposit is a separate asset, with its own terms.

Money placed at the Federal Reserve can leave the commercial banking system. Money retained as an issuer's bank deposit can face different liquidity treatment.

The White House Council of Economic Advisers modelled these differences in April. Its baseline estimated that eliminating stablecoin yield would add $2.1 billion to bank lending, roughly 0.02%, with a net welfare cost to households of $800 million.

The model uses Circle's roughly 12% cash share as a calibration and conservatively assumes that this cash backing supports no lending. That is an assumption about regulatory treatment, rather than a finding that 12% of all stablecoin reserves have disappeared from banks.

So, deposits can remain in the system while becoming less useful to the banks that receive them.

Bank of America's 0.51% consumer-deposit rate applied to about $951 billion of average deposits in Q1. Short-term market rates were several percentage points higher.

That gap is a funding advantage, rather than a reported profit margin. Banks also have operating costs, liquidity requirements and credit losses. But paying little for deposits is still valuable, especially when customers leave money where their salary arrives.

A stablecoin that shares reserve income with its holder adds another competitor for that money. Restricting the payment makes the competitor less attractive.

It also affects what happens to the income on dollars that do move.

Circle shares USDC reserve economics with Coinbase. Coinbase, in turn, pays rewards to eligible customers. Eligibility and rates vary by country and account type; the programme is not universally limited to Coinbase One members.

Coinbase reported about $305 million of stablecoin revenue in Q1 2026 and $292 million in Q2. Its earnings deck also shows broader USDC income totals of $324 million and $320 million, including income on Coinbase's own balances. Customer rewards sit in sales and marketing expense.

If rewards were restricted, Coinbase could retain more income per dollar held, before other costs. That does not guarantee higher total profit: fewer rewards could also mean fewer customers or smaller USDC balances.

The CEA's September follow-up treats a yield ban as an effective tax on holders, with the proceeds going to issuers and intermediaries. At its interactive model's baseline settings, holders who keep their stablecoins forgo about $8.5 billion a year in yield. That is an income transfer, separate from the model's net welfare cost.

The same question runs through Follow the Float: how much of the income reaches the holder, and how much stays with the companies distributing the dollar?

A wider rewards ban could benefit banks by weakening a competitor and benefit stablecoin companies by reducing what they pay holders.

There is a serious argument behind the banks' lobbying. The form of funding matters.

Millions of small retail deposits can be relatively stable. A large reserve deposit controlled by one issuer can leave much faster. A bank losing retail customers may have to replace their deposits with more expensive funding, even if another bank receives the money. The Federal Reserve's own analysis describes these trade-offs.

The CEA does consider this problem. In the same follow-up, applying a 5% to 15% reduction in lending capacity to deposits returning in a less useful form raised its estimate to $2.9 billion to $4.4 billion (up to 0.04% of bank lending).

That remains a small aggregate effect. It does not mean every bank is unaffected, and a static model is not a stress test of a sudden run.

The credit concern is real. GENIUS permits bank deposits as well as short-term government assets in reserves. A BIS working paper published in June models Treasury-backed stablecoin growth raising banks' funding costs and reducing loan supply, alongside cheaper government borrowing. In its baseline US scenario, with demand entirely domestic, $2 trillion of adoption lowers long-run output by about 0.03%. Part of that effect comes through banks paying depositors more. The model also finds a welfare gain once stablecoins' liquidity, payment and safety benefits are included.

Those are effects of stablecoin growth. Whether banning rewards meaningfully protects lending is a separate question. The CEA's estimates suggest only a small lending benefit from a ban.

There is also a defensible policy choice in keeping payment money separate from investment products. Europe's MiCA rules bar both issuers and crypto-asset service providers from granting interest on e-money tokens. Paying yield does not automatically make a stablecoin an unregistered money market fund, but the distinction between a payment product and an investment product deserves attention.

On 15 September, the Senate failed to advance the CLARITY Act, with 49 votes in favour and 50 against (the procedural vote required 60). Its proposed restrictions on interest-like distributor rewards have therefore not become law.

The issuer ban in GENIUS is already enacted. The live question is how far regulators will treat payments through other companies as an attempt to evade it.

The OCC's February proposal would presume that specified affiliate or related-third-party arrangements breach the issuer ban, unless the issuer can demonstrate otherwise. The definition covers issuer-directed reward services and white-label relationships. It is narrower than a blanket ban on every independent distributor's rewards; other arrangements can still be assessed case by case.

Comptroller Jonathan Gould has targeted November for the final rule. GENIUS takes effect on 18 January 2027, its statutory backstop; a final rule issued now can no longer bring that date forward.

The wording will matter more than another round of arguments about deposit flight. It will help determine which rewards can reach holders and which payments must stay with issuers or distributors.

The reserves can keep earning interest. The issuer's interest payment to you is the share the law fixes at zero.

See you next week.

James Smith