Why
A structured way to read every "bank issues a stablecoin" headline: the report separates three genuinely different balance-sheet moves that get talked about as one thing, and shows why a bank ends up needing all three rather than picking a winner. Useful background for anything touching bank-adjacent stablecoin rails or institutional settlement design.
Updated 2026-08-27 — the report's own headline example is now testing its central claim. The Wall Street Journal reports that JP Morgan has held preliminary internal discussions about issuing its own stablecoin, with the bank saying it has no current plans but will weigh options as customer demand and the regulatory environment change. That is family (1) considering family (2). JPM Coin and JPMD are tokenized deposits, and the bank does not issue a stablecoin today — a distinction most coverage flattens and the entire report depends on. If it happens, the complementarity thesis stops being an argument and becomes a disclosure, because the expensive move is the one being contemplated: converting a claim from deposit to redemption contract takes LCR runoff from 25% to 100% and NSFR ASF from 50% to 0%. A bank does not pay that to duplicate something it already has. It pays it to reach the venues a deposit token cannot.
How it works
Three families, each a different trade on the balance sheet. (1) Tokenized deposits — the claim stays a deposit, just wrapped in a token; cheapest Basel treatment (25% LCR runoff, 50% NSFR ASF for the permissioned variant) but narrow reach — JPMD on Base is the live example. (2) First-party stablecoins — the bank issues against a segregated reserve pool; the claim converts from deposit to redemption contract, which is the expensive move: LCR runoff jumps from 25% to 100%, NSFR ASF drops from 50% to 0%, and it can be issued on the parent's own balance sheet, through a licensed subsidiary, or through a bank consortium. (3) Third-party stablecoins — the bank doesn't issue at all, just facilitates access (prefunded inventory or a deposit-secured loan) to reach venues that don't accept bank-claim tokens, at the cost of the claim leaving the bank's balance sheet entirely. The report's core claim: these aren't competing options, they're complementary — a single institutional client might use a tokenized deposit for intra-network settlement, a first-party stablecoin for treasury rebalancing, and a third-party stablecoin to hedge weekend exposure on Hyperliquid-style venues, so a bank offering only one can't fully serve them. Source: https://research.4pillars.io/en/research/tokenized-money-for-banks
What to watch, 2026-08-27
The WSJ report is a preliminary discussion, not a plan, so the useful move is to write down in advance what would separate the readings — and the answer is a single structural choice the report already enumerates:
| If JP Morgan issues through | What it means | Basel consequence |
|---|---|---|
| The parent balance sheet | Maximum control, maximum cost | The full 100% LCR runoff sits on the bank |
| A licensed subsidiary | Ring-fenced — the most likely shape | Cost isolated from the parent's ratios |
| A bank consortium | Reach matters more than the brand | Cost shared, control shared |
The bank's own framing names the two triggers as customer demand and the regulatory environment, which is the report's thesis restated by the institution itself: nobody pays the deposit-to-redemption-contract cost for a customer they can already serve. And the consortium row is where this card meets build-rent-or-own-the-rail and korea-digital-asset-act — the same three structures, argued about in a different jurisdiction and a different currency, with Korea's central bank pushing for exactly the row JP Morgan is least likely to pick.
Why a bank needs its own stablecoin when it already has tokenized deposits
The first idea to throw away is that these are all just dollars turned into tokens. On the surface each is a token representing $1. Financially they are entirely different products. What matters is not that it was issued as an ERC-20 on some chain — it is who the holder has a legal claim against, and what that does to the bank’s balance sheet and liquidity regulation.
1. Tokenized deposit — an existing deposit, expressed as a token
JP Morgan’s JPMD is the clearest example. Holding JPMD is less owning a new kind of money than holding an existing deposit claim on JP Morgan in token form: bank deposit → tokenization → JPMD, and the underlying claim is still a bank deposit.
That matters, because to a bank a deposit is not merely a liability — it is the funding source that makes lending and investing possible. Under the permissioned structure the report assumes, LCR applies a 25% outflow and NSFR recognises 50% ASF.
- LCR (Liquidity Coverage Ratio) asks whether a bank holds enough high-quality liquid assets to survive 30 days of expected outflows in a crisis.
- NSFR (Net Stable Funding Ratio) takes the longer, roughly one-year view: how much of the asset side is funded by stable funding.
So of 100 in deposits, only part is treated as fleeing quickly, and a substantial part counts as stable funding. Very useful for a bank — it keeps the economics of a deposit while adding programmability and fast settlement.
But the weakness is reach. JPMD moving on Base does not mean every Base application, every other chain, exchange or DeFi protocol will accept it. It is powerful only inside the network the bank permits and the counterparty accepts.
2. First-party stablecoin — more expensive money that travels further
When a bank issues a stablecoin the character of the claim changes. It is no longer a deposit wearing a token; it is a separate redemption claim — a promise to redeem at a set price against reserve assets.
Technically small; under bank regulation, large. As the deposit becomes a stablecoin redemption liability, LCR outflow goes 25% → 100% and NSFR ASF goes 50% → 0%. The stablecoin is a far more expensive liability.
Which raises the obvious question: with JPMD already in hand, why would JP Morgan issue something this expensive?
The answer is reach — how far the money can go. An institutional client paying another company inside JP Morgan’s network is fine with JPMD. But suppose that client has to hedge a position on a crypto-native venue over the weekend, and the venue takes USDC and not JPMD. Then JPMD is money that cannot be used there. The client converts JPMD to cash, buys USDC, and leaves the bank’s rail — and the bank has handed part of its client’s money flow to Circle and the crypto-native system.
A first-party stablecoin creates another option: bank deposit → bank stablecoin → public blockchain / external venue. The bank accepts a higher regulatory cost and buys access to external rails with it. The cost is not waste — it is the price of reach.
3. And there is a second economic motive: reserve income
When a client uses USDC, the interest earned on USDC’s reserves accrues to Circle. Issue your own and the structure becomes client funds → bank stablecoin issued → reserve assets → short-dated Treasuries → interest income.
On a simple assumption — $10B outstanding at an average reserve yield of 4% — that is roughly $400M of gross reserve income a year.
> If our clients need a stablecoin in external on-chain markets anyway, why should that only ever be USDC? Issue it ourselves and the client gets external reach while the reserve income stays with us.
Not all of that $400M is profit. A first-party stablecoin carries LCR/NSFR liquidity cost, regulatory cost, compliance, operations and technology. The economics a bank actually sees are roughly:
Reserve yield + client-retention value + payments-business value − liquidity/regulatory cost − operating and compliance cost.
Even so, capturing reserve economics that would otherwise go to a third party is one of the clear attractions. That single line is what risk-free-rate-is-the-floor prices, and it is why the same product denominated in a low-rate currency is a much thinner business.
4. But a first-party stablecoin cannot go everywhere either
This is where the third product becomes necessary. JP Morgan issuing a stablecoin does not guarantee it will be accepted the way USDC is across exchanges, chains, DeFi protocols and payment services. USDC’s strength is not the token technology — it is distribution and network effect, already connected to a great many venues.
So in some markets it is more economically rational to give the client access to USDC than to force one’s own token. There the bank does not issue — it intermediates access.
5. The bank is not choosing one token
This is the report’s most important claim. The future bank is unlikely to be choosing between JPMD or a JPM stablecoin or USDC. A single institutional client may need all three: JPMD for large weekday payments inside JP Morgan’s network, the bank stablecoin to reach financial services on a public chain, USDC to hedge on a specific crypto-native venue at the weekend.
| Tokenized deposit | Bank stablecoin | Third-party stablecoin | |
|---|---|---|---|
| Where it goes | Inside the bank and permissioned networks | Wider external / on-chain rails | Crypto-native venues even the bank’s own token cannot enter |
| Regulatory cost | Relatively low | High | None to issue |
| What is kept | The deposit character | Reach and reserve economics | — |
| What is lost | Reach | LCR 25→100%, ASF 50→0% | The funds and the reserve income leave the bank’s economy |
So the bank’s role becomes routing rather than issuing — connecting the client’s money to the most appropriate form of money and rail for wherever it has to go. The future bank as a money router: deposit → tokenized deposit → bank stablecoin → third-party stablecoin, moving between them as needed while keeping the client’s funds inside its own financial relationship as far as possible.
Read this way, a bank entering stablecoins is not a story about banks doing crypto. The real competition is not who builds the better token — it is who is connected to more rails and venues.
And that is why JP Morgan, already holding a tokenized deposit, would still consider its own stablecoin. Not to duplicate the deposit — to send the client’s money where the tokenized deposit cannot go, to stop the client being handed wholesale to a third party like Circle, and, where possible, to keep the Treasury interest on the reserves inside its own economy.
In the end the value of a stablecoin to a bank is not the token. It is the reach.
> The tokenized deposit is cheap but does not travel far. The first-party stablecoin is expensive but travels further. The third-party stablecoin travels furthest, but the money leaves the bank’s balance sheet and its economy.
Which is why a bank goes toward offering all three as needed rather than picking one — and that is the most important frame for reading “tokenized money for banks” at all.
← All Dev Notes · Workspace Index · Top ↑ · Open on jaylabs.xyz →