Why Europe's central bank is putting its money on-chain

On October 1, 2026, Isabel Schnabel stood up at the Bank of England to speak at a conference held in honour of Charles Goodhart, the economist who spent much of his career explaining where central banks came from. Goodhart's answer was unglamorous: central banks did not begin as inflation fighters. They grew out of the plumbing — the place where banks settled their debts with one another. Control over that settlement layer came first. Monetary policy came later, as a consequence.

Schnabel, a member of the Executive Board of the European Central Bank (ECB), used the occasion to argue that the plumbing is being rebuilt, and that central banks need to move with it. Her presentation, titled "Central banks on-chain," condensed a longer speech she gave at the Jackson Hole Economic Policy Symposium in August. The message in both is the same: as stocks, bonds and bank deposits migrate onto blockchain-style ledgers, central bank money has to follow them there — directly, not through a private middleman.

That sounds technical, and much of it is. But the question underneath it is old and political: who issues the money that everything else ultimately settles in? Schnabel's answer, backed by a new ECB system that went live on September 21, is that it must remain the central bank — and the ECB is building the infrastructure to make sure it does.

What tokenisation promises

Tokenisation means representing financial assets and money as digital tokens on a programmable ledger, a shared database often built on distributed ledger technology (DLT). Schnabel's presentation reduces its appeal to two properties.

The first is atomicity: both legs of a trade — the security and the cash — settle together or not at all. That removes the risk that one side delivers and the other does not. The second is programmability: settlement can be made conditional on rules that execute automatically, instead of moving through a chain of manual, sequenced steps.

Schnabel is candid that neither idea is new. Europe's existing securities settlement platform, TARGET2-Securities (T2S), already delivers securities against cash simultaneously and can automatically generate a short-term loan if a buyer runs short of cash mid-trade. What tokenisation adds is generality. Market participants could write their own conditions covering the whole life of an instrument.

Her example is the repo, a short-term loan secured by collateral and the workhorse of money markets. A repo involves the initial swap of cash for bonds, margin calls if prices move, substitution of one bond for another, and the return of everything at maturity. On a programmable ledger, most of that could run on smart contracts, cutting out layers of messaging and reconciliation between institutions. The gains are largest across borders and time zones, where collateral often has to be parked overnight in advance.

Europe's particular problem

For the euro area, Schnabel frames tokenisation as more than an efficiency upgrade. It is a chance to fix a structural weakness that European officials have complained about for two decades.

Europe's financial infrastructure is still split along national lines. One slide in her presentation shows the tangle of exchanges, clearing houses and central securities depositories that make up the European system. An investor who wants to buy government bonds across several euro countries may need access to several different depositories. Large institutions can absorb that cost. Smaller ones often cannot.

Tokenisation, in Schnabel's telling, offers a market that is integrated by design rather than stitched together after the fact. She ties it explicitly to the EU's savings and investments union, the long-running effort to channel European savings into European companies. She also points to smaller-scale benefits: France's "Lightning Stock Exchange" (Lise), a tokenised venue aimed at small companies, and fractional ownership that lets investors hold a slice of an asset — a gold bar, a building — that would otherwise require a large minimum stake.

There is a legal catch she flags in the longer speech. A token is only as uniform as the law behind it. If the same bond token means different things in 20 jurisdictions, the ledger does not deliver integration. That is why she connects the project to a proposed "28th regime," a single optional EU-wide legal framework sitting alongside national law.

Why stablecoins cannot be the foundation

The heart of Schnabel's argument is about which money tokenised markets should settle in. The leading private candidate is the stablecoin, a token issued by a private company and backed by reserves such as government bonds.

Modern monetary systems, she explains, run on two tiers. At the top is central bank money, which banks use to settle with each other. Below it are commercial bank deposits, which circulate at par — one euro at Bank A is worth exactly one euro at Bank B — because both banks can settle in central bank money. Her slides label this structure trusted, safe and scalable.

That arrangement was not designed on a whiteboard. It emerged after long experience with private money. During the U.S. free banking era in the 19th century, state-chartered banks issued their own notes, and notes from distant or doubtful banks often traded at a discount. A dollar was not always a dollar.

Schnabel sets out two tests any settlement asset must pass. It must be safe, meaning free of credit, liquidity and redemption risk. And its supply must be able to expand elastically when demand for liquidity spikes. She grants that a carefully built stablecoin — backed only by short-term, floating-rate government debt, as economist Darrell Duffie has proposed — could pass the first test. No private issuer can pass the second. A stablecoin company cannot conjure liquidity in a panic. It can only sell what it holds.

The lesson of 1907

To show why elasticity matters, the presentation reaches back more than a century. One chart tracks the call money rate on the New York Stock Exchange — the overnight rate for loans secured by stock — through the Panic of 1907. On October 24 of that year, it spiked to around 100% a year.

At the time, the U.S. money supply was tied to banks' holdings of eligible government bonds. When demand for cash surged, there was no mechanism to expand it quickly. Private bankers organised rescues, but they were improvised and lacked legal backing. The experience led directly to the Federal Reserve Act of 1913 and the creation of a central bank that could supply liquidity on demand.

The second chart on the same slide shows the modern contrast. In early 2020, as COVID-19 hit, the ECB's balance sheet expanded from roughly 40% of euro area gross domestic product (GDP) toward about two-thirds of it by mid-2021, while euro repo rates stayed below zero. That is elasticity in practice: the central bank created the liquidity the system demanded.

Schnabel's conclusion is that stablecoins are best understood as complements to central bank money, not replacements for it. They may widen the range of payment options for households and businesses. But in her words, financial markets can only scale safely if transactions settle in "a risk-free asset that can be supplied elastically" — and only a central bank can provide one.

Three ways to bring central bank money on-chain

Once stablecoins are ruled out as the foundation, the question becomes how central bank money reaches tokenised markets. Schnabel lays out three models.

In the first, the central bank issues tokenised reserves directly on a programmable ledger. Reserves become native digital tokens.

In the second, a bridge connects the existing settlement system to a DLT platform. When a trade happens on-chain, a message triggers the cash payment in the traditional system. Reserves never leave their current home.

In the third, a private intermediary holds reserves in an omnibus account at the central bank and issues tokens on the ledger backed one-for-one by those reserves. The token is a private claim, not a claim on the central bank.

Schnabel makes clear she favours the first. She compares the omnibus model to the pre-1913 U.S. system, in which interbank settlement ran through a pyramid of private correspondent claims. The token's value would depend on the intermediary's ongoing soundness, and if every platform built its own wrapper, settlement could fracture into competing private claims instead of one common asset. Under both the bridge and omnibus models, she adds, the central bank stays a passive balance sheet in the background and gains none of the technology's benefits.

Why the central bank wants programmable money

This is where the argument turns from defence to ambition. If reserves live on the ledger alongside collateral, Schnabel argues, the ECB could run monetary policy operations natively through smart contracts. A standard repo between the central bank and a commercial bank could settle atomically. Collateral rules — requests for more margin, substitution of securities, different interest rates for different uses — could be written into the settlement process itself.

She points to Project Pine, a study by the Federal Reserve Bank of New York and the Bank for International Settlements (BIS), which showed that smart contracts could make policy implementation faster and more flexible. And she notes a step the ECB has already taken: since January 2026, marketable assets issued through DLT-based depositories are eligible as collateral for Eurosystem credit operations.

The deeper reason is speed. Tokenised markets will move faster, and so will their crises. Automated margin calls triggered by price moves could force rapid asset sales and amplify a downturn. More frequent payments that cannot be netted against each other could raise banks' need for liquidity during the day. A central bank working on yesterday's rails may be unable to respond in time. On-chain, Schnabel argues, it could launch new facilities and adjust interest rates, collateral requirements and access conditions with immediate effect.

This is the most consequential claim in the presentation. The case for going on-chain rests partly on efficiency, but mainly on the central bank keeping its ability to stabilise the system once that system runs at machine speed.

Pontes and Appia: what is already being built

The ECB is not just theorising. Its presentation highlights two projects that put the argument into practice.

Pontes — Latin for "bridges" — launched on September 21, 2026. It connects the Eurosystem's real-time settlement system to market DLT platforms and adds a Eurosystem-operated ledger for settling DLT-based trades in central bank money. According to the presentation, it offers a dual settlement model, through either the existing TARGET2 system or the DLT platform. At launch, legal finality for the cash leg still sits in TARGET2. The planned next steps are 24/7 availability and programmability on the ECB's own ledger.

The name is somewhat misleading. Pontes starts as a bridge — option two in Schnabel's framework — but the roadmap points toward option one, with native tokenised central bank money and smart contracts running on infrastructure the Eurosystem controls.

Appia is the longer-term programme exploring what Europe's tokenised system should ultimately look like. The presentation sketches a spectrum: a single unified ledger holding central bank money, bank deposits and securities together; a Eurosystem ledger linked to separate private networks; or several shared ledgers that each host both reserves and assets.

The trade-off nobody has resolved

Schnabel does not pretend the architecture question is settled. A single unified ledger maximises the benefits: central bank money and assets sit in one place, settlement is strictly atomic, and every asset follows the same rules. South Korea's Project Hangang, led by the Bank of Korea, is already testing this model. Schnabel's own slide shows how a tokenised system can reproduce the familiar two tiers, with tokenised central bank money at the top and tokenised bank deposits below.

She draws a pointed conclusion from that. If tokenised bank deposits can be programmed, settle in central bank money and work seamlessly with tokenised assets, they deliver most of what stablecoins promise. Under a unified ledger, she says, it becomes hard to see a compelling role for new private money in domestic payments.

But one ledger also means one point of failure. It concentrates governance questions that have no easy answers: who is liable if a smart contract fails, who admits new participants, who decides on software upgrades and confidentiality rules. Today, the Eurosystem owns TARGET Services while private firms such as Euroclear and Clearstream own other parts of the infrastructure. A shared ledger blurs those boundaries. It also risks technology lock-in and a processing bottleneck.

Several interoperable ledgers avoid some of those problems but create others. Every connection adds complexity, and if reserves cannot move freely between networks, liquidity fragments and banks need to hold more of it. Schnabel's tentative conclusion is that strong interoperability may capture most of the benefits without the concentration risk — and that old and new systems will likely run side by side for a long time, perhaps permanently.

What it means for ordinary people

Strip away the terminology and the presentation describes a defensive move with offensive ambitions. Central banks derive their power from sitting at the centre of settlement. Tokenisation threatened to build a new centre somewhere else, possibly around privately issued stablecoins, which today are overwhelmingly denominated in U.S. dollars. The ECB's response is to make sure the new centre is still the central bank, now running on programmable infrastructure.

For readers trying to make sense of this, a few points stand out.

First, this is wholesale money, not the digital euro. The presentation keeps them separate. Pontes and Appia deal with money that banks and financial institutions use to settle with each other. The digital euro is a separate retail project for households. Nothing in this presentation changes the money in your bank account or your wallet in the near term.

Second, the stablecoin debate is really a debate about monetary sovereignty. When the ECB says stablecoins cannot be the settlement foundation, it is making a technical argument about elasticity. It is also drawing a line around who controls the euro's plumbing. If European markets tokenise and settle in dollar stablecoins, part of the euro area's financial system would rest on assets the ECB cannot supply in a crisis.

Third, programmability is a central bank tool as much as a market tool. Schnabel openly describes the ability to change interest rates, collateral terms and access conditions instantly through code. In wholesale markets, that is about crisis response. But it is worth noticing how naturally the language of programmable money now sits in central bank speeches, because the same idea will be debated on the retail side.

Fourth, the place of Bitcoin is clearly marked. On Schnabel's map of digital assets, Bitcoin and Ether sit with unbacked cryptoassets under "trading or investment assets," not settlement assets. The ECB does not see them as part of the monetary system's foundation, and this programme is not designed to accommodate them.

Fifth, the timeline matters more than the announcements. The things to watch are when Pontes adds 24/7 operation and smart contracts, which Appia architecture the Eurosystem chooses, whether the EU adopts a 28th legal regime that makes tokens uniform across borders, and how governance questions such as liability for failed smart contracts are answered.

A very old fight on new rails

Goodhart's insight was that central banks became powerful because they held the ledger everyone else settled on. In 1913, the United States decided after a near-collapse that the ledger should be public and that its money should be able to expand on demand. Schnabel's presentation, delivered at an event in his honour, makes the same decision again for a new technology.

The rails are changing. The argument about who sits at the centre of them has not changed at all.