A wallet can be useful without becoming a large savings account. That distinction sits at the centre of the European Central Bank’s argument for a digital euro: give people access to public money for everyday payments, while limiting the amount they can accumulate outside commercial-bank deposits.
In an October 6 speech, ECB Executive Board member Piero Cipollone defended holding limits and linked bank accounts as complementary safeguards. He did not announce a final numerical limit. His explanation matters because the same design choice affects payment convenience, banks’ funding and the economics of distributing a new payment instrument. A reassuring system-wide simulation answers only part of that question.
A balance ceiling can coexist with a larger payment
A digital euro balance would be central-bank money. A conventional bank deposit is a liability of the commercial bank. Moving funds from the latter into the former can therefore reduce a bank’s deposit funding, even if the customer continues using the bank’s app to initiate payments. Keeping the customer interface does not, by itself, preserve the balance-sheet funding behind it.
The proposed architecture separates the amount held from the amount paid. Cipollone described linking a digital euro wallet to a bank account so that payments can draw additional funds when needed, while excess incoming balances can return to the linked account. The wallet would not pay interest. Together, these features aim to make it useful for transactions without encouraging large, persistent balances.
A holding cap is consequently different from a cumulative spending allowance. Money can enter, be paid out and be replenished; payment turnover need not equal the balance sitting in the wallet at a particular moment. That distinction also explains why adoption cannot be judged solely by the maximum permitted balance. Reliability at checkout and the convenience of the link to a bank account would matter.
The Council’s December 2025 negotiating position envisages limits on total holdings and a framework for setting them. It is a negotiating position, not proof that the final legislation or wallet limit has been settled. Cipollone said negotiations were continuing and issuance would require a later ECB decision after legislation was adopted.
A sector average can conceal a funding mismatch
The ECB’s technical analysis published in 2025 tested hypothetical limits from €500 to €3,000 across 2,025 banks. Its supervisory inputs include first-quarter 2024 data. These are modelled effects on a historical data set, not observations of a digital euro operating in October 2026.
One result is particularly useful for reading the debate: with a €3,000 limit under normal conditions, and excluding the assumed deposit inflow from declining cash use, the aggregate liquidity coverage ratio falls from 166% to 163%. That is a model output measuring liquid-asset protection against short-term stressed outflows, not a prediction of deposit growth or a guarantee about every bank.
The distinction between aggregate and individual results is material. A bank serving many customers with relatively small balances could have a different exposure from one funded by larger corporate accounts. Access to replacement funding, available collateral and existing liquidity buffers would also change its response. A sector ratio cannot identify the cost of replacing deposits at a named lender.
The study additionally assumes that continuing payment digitalisation brings deposits into banks as cash use declines. Its estimated €127 billion inflow by 2034 is a scenario-dependent projection, rather than money already received. Whether those inflows reach the same institutions losing deposits is a separate distribution question.
The extreme case assumes simultaneous demand up to permitted balances and no new monetary-policy response. It still models access to ordinary central-bank operations where conditions permit. This is a deliberately severe exercise, not a forecast of a bank run or an assertion that central-bank liquidity would disappear. Conversely, its assumptions should not be mistaken for a complete description of behaviour during a future crisis.
Liquidity comfort does not settle the implementation bill
Banks would have another calculation beyond replacing deposits: the cost of integrating wallets, customer support and payment systems. An industry-commissioned study published by European banking associations in June 2025 used estimates from 19 retail banks. Its scope excluded offline functionality, multiple accounts and ongoing operating costs. The EBF’s April 2026 supplement says participants revisited their estimates and confirmed the original assessment after further design work.
Those are interested parties’ estimates, not an audit of actual expenditure. Their scope also differs from a balance-sheet stress model. Neither exercise can be used to declare the other irrelevant: avoiding a funding shock does not pay an integration invoice, while a large estimated project bill does not demonstrate a liquidity crisis.
The Council’s position would make specified basic services free to consumers and provide a compensation framework for intermediaries. That shifts attention towards the eventual rules for merchant charges, bank remuneration and added-value services. A distribution role could preserve customer relationships and support new services, but participation alone does not establish a positive return on implementation spending.
There is a credible case that capped, non-interest-bearing public money would create less scope for deposit migration than an unrestricted savings substitute. There is also a credible demand for bank-specific evidence before dismissing funding concerns. Both can be true. The evidence that would sharpen the analysis is practical: a final legal framework, calibrated limits, pilot results showing actual funding and replenishment behaviour, and bank disclosures separating implementation costs from recurring revenue. Until then, the cap is a design variable with competing objectives, not a completed verdict on bank profitability.