ECB's Schnabel maps three models to put central-bank money on-chain

Isabel Schnabel, member of the ECB executive board, laid out three concrete architectures for settling in central-bank money on distributed-ledger platforms in a speech delivered at the Bank of England's "Future of Money" conference on Oct 1. All three preserve the Eurosystem's two-tier monetary structure — reserves at the central bank, commercial money for the public — and all three slot into the ECB's existing Pontes and Appia DLT settlement work.

What happened

Schnabel's brief was narrow: wholesale central-bank money must remain the risk-free settlement asset even as collateral, repo, and the DvP leg of securities trades move to DLT venues. She framed the choice as three models, each a trade-off between ambition and interoperability with the current real-time gross settlement system (TARGET).

Model 1 — direct issuance on DLT. The central bank issues reserves directly on a programmable DLT platform. The reserve itself is tokenized and sits natively on-chain. Highest change-cost, highest expressivity — smart contracts can hold and move reserves, but TARGET and the DLT platform are now two homes for the same claim.

Model 2 — interoperability layer. The existing RTGS (TARGET) stays intact; reserves are not tokenized. A hash-linked bridge coordinates the DLT platform's cash leg with a central-bank-money leg on TARGET. Lowest change-cost; delivery-versus-payment across the two is coordinated, not atomic.

Model 3 — tokenized settlement claims. Reserves at the central bank are tokenized, but what circulates on the DLT platform is a private claim on those reserves — a fully backed token, not a direct central-bank liability. Middle ground on change-cost; the on-chain asset is legally not the reserve itself but a redemption right.

Mechanism

The distinction is important because it decides who bears the credit risk during a settlement leg:

• Model 1 keeps the claim on the Eurosystem throughout.
• Model 2 keeps the claim on the Eurosystem but splits the on- and off-DLT legs — the bridge is the point of trust.
• Model 3 shifts the on-chain leg's credit risk to the intermediate issuer, even if that issuer is itself fully covered at the central bank.

Schnabel did not pick a model. The ECB has committed to parallel exploration via Pontes — interlinking existing TARGET services with DLT platforms, closer to Model 2 — and Appia, a longer-horizon architecture with Model 1 features on the table.

Impact

• Wholesale DLT venues: operators running tokenized collateral and repo pilots — the Fnality, HQLA-X, Clearstream D7, Partior cohort — now have an ECB-approved shortlist of integration patterns. Any production tokenized-settlement workflow will be sized against these three.
• Stablecoins: Model 3 maps directly onto the fully-reserved stablecoin shape — a point of continuity with MiCA's reserve and redemption requirements for e-money tokens.
• Timeline: Pontes was first scoped in 2024 and is targeted for first-wave go-live in 2027; Appia is longer-horizon. Schnabel's framing suggests the three models are not sequential but parallel options being tested in both tracks.

What to watch

1. ECB Pontes progress reports. The model choice will become visible in each DLT platform's integration interface — the data format, the message exchange, the finality rule.
2. Bank of England alignment. The BoE, which hosted the speech, has synchronization-bridge tooling under test in its RTGS Renewal programme. Whether London lands on the same three-model taxonomy decides how cross-currency DvP works when both sides sit on DLT.
3. BIS Agorá. The cross-border variant of Model 1 remains the comparator the ECB will be benchmarked against — the differences between Agorá's architecture and the Eurosystem's eventual pick are the shape of future interoperability.
4. MiCA review tie-in. The ESMA/EBA consultation on MiCA 2 is open; whether tokenized settlement claims (Schnabel's Model 3) land in or out of MiCA's scope — alongside the EMT regime — is the question that affects every private issuer building against EUR rails.

Context

The three-model framing closes a long open question in the Eurosystem's wholesale DLT work: whether the central bank would commit to a single architecture or pluralize. Pluralizing lets commercial DLT venues stay interoperable without rebuilding, but it also means the ECB will not be issuing a single wholesale CBDC the way the digital euro project targets retail. For developers building on-chain settlement rails against EUR, the choice of model does not dictate what contracts look like — it dictates who stands on the other side when settlement finality hits.