Which Coherent Corner? An Options Appraisal for Europe's Stablecoin Choice

1. From diagnosis to choice

A companion note argued that MiCA’s regime for euro stablecoins is built on an incoherent corner: a privately issued token that promises par, holds credit-bearing reserves, and has no public backstop. That argument was deliberately one-sided — it diagnosed a problem and named the three coherent escapes without ranking them. This note does the ranking, or rather refuses to: it sets out the case for and against each of the three coherent designs and shows that the right choice is conditional on what the European Union is trying to achieve.

The three options, stated plainly:

• Option A — Narrow money. The token is fully backed by central-bank money (or, in a weaker variant, by short-dated euro-area sovereign debt in a bankruptcy-remote structure). The issuer takes no credit risk and does no maturity transformation. The instrument is, in effect, a privately distributed claim on the central bank — a “synthetic central-bank digital currency.”
• Option B — Bank money. The issuer is regulated as the bank it functionally resembles, with the full safety net: deposit-style protection, a resolution regime, and access to the lender of last resort. In practice this is the world of bank-issued stablecoins and tokenised deposits.
• Option C — A public anchor. Money itself stays public — carried by a digital euro and by central-bank money settled on-chain — while private tokens are reduced to settlement rails tethered to that anchor. This is the European Central Bank’s “separate functions from instruments” position.

The instinct to treat these as rival answers is natural but wrong, and correcting it is the most useful thing this note can do.

2. The reframe: three coherent designs, not one winner

Be precise about what the trilemma claims. It is a constraint on a single instrument: no one token can promise par, take credit risk, and do without a backstop all at once — reach for the third and one of the other two gives way. That is why MiCA’s non-bank token is unstable. But the constraint binds the token, not the system. A monetary order can run several kinds of money side by side, each obeying the rule in its own way.

That is the move that unlocks the whole debate. Instead of asking “which of the three should the euro stablecoin be?”, ask “which kinds of euro money should Europe operate, and in what proportion?” Put that way, the three options stop being rival answers and become three distinct designs — three lanes that a single system can run in parallel (Figure 1):

• In the narrow-money lane, a private firm issues the token but holds the backing as central-bank money. The firm takes no credit risk; the public balance sheet does the heavy lifting; there is nothing to run on. The issuer competes on distribution and technology, not on risk-taking.
• In the bank-money lane, the token is a bank’s liability — a tokenised deposit or a bank-issued coin — and the public safety net stands behind it, exactly as it stands behind ordinary deposits today. Credit creation continues; so does the moral hazard the safety net carries.
• In the public-anchor lane, the money itself is public — a digital euro, or central-bank money settled on-chain — and private tokens are demoted to rails that settle against it. The token is no longer money; it is plumbing.

Running in parallel does not mean peaceful partition. The lanes compete: narrow money (A) and a digital euro (C) chase the same demand for safe digital money, and both draw deposits away from bank money (B). “Coexistence” therefore means managed competition — and the instrument that manages it, the disintermediation dial, is the subject of §9.

What separates the three is not cosmetic. It comes down to three questions: who issues the claim, what backs it, and who absorbs the loss when something breaks. Each lane answers them differently, and coherently. The corner MiCA occupies today answers them incompatibly — a private issuer, credit-bearing backing, and no one clearly on the hook for the loss — which is precisely why it is the one design that does not hold.

Why can that corner not simply be fixed with tighter rules? Because the par promise is a coordination problem, not a capital problem. A holder redeems when she expects others to redeem first; the run can arrive even if the issuer is solvent (Diamond & Dybvig, 1983). Ex-ante prudence — more capital, safer reserves, stress tests — lowers the probability of that bad equilibrium but cannot remove it. Only two things do, and each gives up one leg of the corner: a credible backstop that pays out when everyone runs at once (a public safety net), or reserves carrying no credit or liquidity risk at all (giving up private credit creation). History is the witness: privately issued par money without a backstop has, every time, ended in one of three ways — notes trading at a discount (the nineteenth-century free-banking era), promises the issuer could not keep under stress (the offshore eurodollar market), or survival only on an emergency public rescue (the constant-value money-market funds bailed out in 2008 and again in 2020). The instability is structural, not a calibration error. That is why this note treats “fix the dials” as the wrong project and “choose a coherent corner” as the right one.

So the EU’s real decision is not “which one?” but “what mix, and in what order?” That is a political-economy question, because each lane advances some objectives at the expense of others. The appraisal therefore takes the objectives first, then weighs each lane against them.

3. The objectives the choice must serve

Six objectives recur in the European debate, and they pull in different directions. Naming them is half the work, because the disagreement between the Commission, the ECB and the EBA is at bottom a disagreement about their weighting.

1. Monetary sovereignty — resisting “digital dollarisation” and keeping the euro the anchor of euro-area payments and settlement.
2. Financial stability — run-resistance, and containment of contagion into banks and asset markets.
3. Monetary-policy transmission — preserving the bank-deposit base through which euro-area policy mostly works, i.e. avoiding disruptive disintermediation.
4. Innovation and competitiveness — a dynamic private sector that can build and ship, and entry for non-bank fintechs, not just incumbents.
5. The international role of the euro — a safe, programmable euro instrument usable beyond the bloc.
6. Feasibility and speed — what can actually be delivered, and how fast, while dollar tokens are entrenching.

Two cross-cutting considerations sit underneath all six: the public cost and central-bank balance-sheet burden each option implies, and who captures the rents — incumbents, fintechs, or the state. The second is where the politics is sharpest.

4. Option A — Narrow money (the synthetic CBDC route)

How it works. Issuers may offer a euro token only against fully safe, fully liquid backing — ideally reserves held at the central bank, accessed through a narrow licence; in the weaker form, 100% short-dated, high-quality liquid assets — which, for a euro token, means short-dated euro-area sovereign debt — held in a bankruptcy-remote vehicle. There is no credit creation and no maturity transformation, so there is nothing to run on.

The case for. This is the only private design that removes credit and liquidity risk from the backing: with reserves held as central-bank money, par is mechanical, not a promise (Goel, 2024; Adrian & Mancini-Griffoli, 2019). It is as close to run-proof as a private instrument gets — though operational, governance and legal risks remain, which is why even its advocates avoid the absolute term. It keeps private firms in the game — they compete on wallets, distribution, programmability and user experience — while safety comes from the public balance sheet, a clean public–private division of labour. It protects the singleness of money, needs no deposit insurance or lender of last resort, and puts no taxpayer credit risk on the table. It is also comparatively fast: it leverages private distribution rather than waiting for the state to build a retail rail. Bruegel’s recent work makes exactly this argument — give EU-regulated issuers access to the central-bank balance sheet, and you “embed” rather than “contain” the technology (Bruegel, 2026).

The case against. Three costs are serious. First, it requires the central bank to open reserve access to non-banks — a structural change in the monetary system that widens the set of policy counterparties, and one the ECB has historically resisted. Second, it is the sharp edge of the disintermediation problem: a safe, convenient, central-bank-backed token competes directly with bank deposits for the public’s safe money, and at scale it shrinks bank funding and therefore bank credit (the classic narrow-banking critique; Williamson, 2024, finds narrow banking “never welfare-improving” — though that is a model-dependent result, and other models find a safe public instrument can raise welfare, so the welfare question stays contested). Third, the business model is thin: with no spread, issuer viability depends on fees or on remuneration of reserves, which drags the central bank into politically loaded decisions about paying private firms to hold public money. There is also a tension internal to the EU’s own plans: a privately distributed synthetic CBDC is a near-substitute for the ECB’s own digital euro, and the two would compete for the same demand. In the weaker, sovereign-bond-backed variant the euro has a problem the dollar does not: there is no single euro “safe asset.” The backing has to be spread across national issuers — Bunds, OATs, BTPs and the rest — each with its own credit spread, its own issuance volumes, and, in stress, its own redenomination and sovereign-risk premium. Concentration limits (the EBA’s draft cap of roughly 35% per sovereign) force diversification into weaker credits; a stress event can gap those bonds and feed the bank–sovereign “doom loop.” Even short-dated euro paper, then, carries the liquidity and fire-sale risk that March 2020 exposed in Treasuries, compounded by fragmentation — so in the euro area only true central-bank-reserve backing is fully clean, and the sovereign-bond variant is shakier here than the US analogue would suggest. And there is a hard institutional edge: granting non-banks access to the central-bank balance sheet runs into the ECB’s mandate, its counterparty framework, and the politics of who may hold central-bank money. (TFEU Arts 127–128 are usually read as supporting a bank-based system; they contain no outright ban, so the barrier is one of mandate interpretation and framework change, not a single textual prohibition.) Either way, this is not a dial the ECB can simply turn. So the feasibility of the strong (central-bank-reserve) variant is low, even though the analytically cleanest version of A depends on it — the more buildable sovereign-bond variant is the one that keeps the liquidity-and-fragmentation risk.

Who it suits. An EU that weights innovation and competitiveness highly, and is willing to clear a high legal and political bar to expand the central-bank balance sheet to non-banks (or to accept the less-clean sovereign-bond variant, with its euro-specific fragmentation risk) and to manage the disintermediation dial actively.

5. Option B — Bank money (bank-issued stablecoins and tokenised deposits)

How it works. Par-promising tokens are issued by entities inside the banking perimeter — credit institutions, with capital, supervision, resolution and the backstop that makes a bank’s par promise credible. What places a token in this corner is the issuer — a bank inside the safety net — not the legal wrapper, and today there are two permitted wrappers. The first is a bank-issued e-money token, regulated under MiCA’s EMT regime. The second is a tokenised deposit — an on-chain claim on an ordinary bank deposit, already an allowed category of money. They run on different rule-books (the EMT regime versus deposit law) but share the decisive feature: a bank stands behind the claim. They are not, though, equally protected at the holder level. A tokenised deposit is a deposit, covered by the deposit-guarantee scheme up to the limit; a bank-issued EMT is e-money, which sits outside that scheme even when a bank issues it. So while the issuer is inside the safety net in both cases, the bank-issued EMT gives the holder the weaker claim — it sits at the edge of this corner rather than its centre, backstopped at the institution level but not guaranteed at the instrument level.

The case for. This is the most proven and least novel design: it is simply the settlement we already use for bank deposits, with runs handled by insurance, resolution and the lender of last resort (Gorton & Zhang, 2023; Awrey, 2022). Because the money stays bank money, credit intermediation and monetary-policy transmission are preserved — there is no disintermediation, because it is the same banks. It reuses the existing supervisory machinery rather than inventing a new one, and it is where the market is already heading: European banks are moving on both fronts — some on tokenised deposits, others on bank-issued coins. The prominent euro-stablecoin consortium now forming is, as it stands, structured as a bank-issued e-money token under MiCA rather than a tokenised deposit. Because its issuers are banks inside the safety net, it is designed to sit in this corner (at the edge just described, since the token itself is e-money and so outside deposit guarantee); and were the review to settle on B as the favoured form of euro “money token,” such a consortium would be very well placed. It is, in short, the pragmatic, incremental path.

The case against. The costs are about competition and reach. Confining issuance to banks entrenches incumbents and shuts out the fintech entrants who drove the technology — a real competitiveness and capture cost, and the reason a “competitiveness” agenda resists it. It also extends the public safety net, with its moral hazard and contingent taxpayer liability, to a fast-growing new instrument, widening the too-big-to-fail perimeter. And there is a structural limit, though it bites unevenly across the two wrappers. Tokenised deposits are not fungible across banks the way a single stablecoin is: a deposit-token at Bank A is not automatically worth one at Bank B without a common settlement layer — precisely the public anchor of Option C. A shared, bank-issued EMT (the consortium model) sidesteps that, because it is one fungible token backed by a common pool, so it delivers singleness within its own programme — but it does so by becoming a private money in its own right, governed by the EMT regime, with the multi-issuer and holder-protection questions that brings (an e-money claim is not a deposit, and so falls outside the deposit-guarantee scheme even when a bank issues it). So tokenised deposits need C underneath them for cross-bank singleness, while a shared consortium EMT supplies its own singleness but reopens the question of exactly what stands behind the token-holder. Finally, bank authorisation is slow and costly, which risks ceding the fast-moving global market to dollar incumbents in the meantime.

Who it suits. An EU that weights stability, transmission and feasibility highly and is relaxed about incumbency — and that pairs B with a public settlement anchor to solve singleness.

6. Option C — A public anchor (the digital euro and on-chain central-bank money)

How it works. The monetary function is lodged in public money: a digital euro for retail, and central-bank money settled natively on-chain for wholesale. Private tokens — whether bank-issued or narrow — are admitted only as instruments that settle against, and are tethered to, that public anchor.

The case for. This is the cleanest answer to monetary sovereignty: money stays public, singleness is guaranteed by the anchor, and private tokens become interoperable plumbing rather than competing private monies. It supplies the common, neutral settlement asset that the Eurosystem’s own surveys have flagged as the missing precondition for tokenised finance — the “tokenised cash” that lets tokenised securities settle delivery-versus-payment without a private substitute. The wholesale leg, in particular, is tractable in the near term and avoids the controversy of a retail digital euro, while still anchoring the market. If delivered in time, it is the strongest defence against the “infrastructure dollarisation” of European settlement (Cipollone, 2026).

The case against. The retail digital euro is politically contentious — privacy, the spectre of “the ECB in your wallet,” and disintermediation fears — and is therefore slow, and hemmed in by holding limits and non-remuneration designed to stop it draining bank deposits (Kumhof & Noone, 2021). Those very safeguards cap its usefulness as money. There is a real risk that a dominant public option chills private innovation, with the state effectively choosing the rail. And the binding constraint is execution time: public infrastructure is slow to build, and the Bruegel critique bites here — while Europe perfects the public anchor, dollar tokens entrench as the default settlement layer. Option C also does not, by itself, give the private sector a viable euro-stablecoin business; it relies on others building on top of the anchor.

Who it suits. An EU that weights monetary sovereignty and long-run control above speed, and is prepared to spend the political capital and time the digital euro requires.

7. The appraisal in one view

The table scores each option against the six objectives — strong, moderate, weak — as a way to make the trade-offs legible, not as a precise metric. The point is the pattern of strengths and weaknesses, which differs sharply across options.

No column is dominant. Narrow money wins on innovation and run-resistance but is worst on transmission and faces a high feasibility bar in its cleanest form (reserve access). Bank money wins on transmission, feasibility and reuse of existing tools but worst on competition and cross-bank singleness. The public anchor wins on sovereignty and singleness but worst on speed and private dynamism. That is exactly why the debate is unresolved: the institutions are not disagreeing about facts but about which row matters most. A caution on the table itself: these are directional judgements, not measured scores — the disintermediation magnitudes that would pin several cells down are unmodelled here (see §9a).

8. How the answer changes with objectives

Three stylised priority-orderings show how the “best” mix flips:

A sovereignty-and-stability-first EU. If the overriding aims are to keep money public and the system run-proof, the mix tilts to C as the foundation plus B for private credit money, with free-standing non-bank private money actively discouraged. This is, in fact, the revealed preference of the ECB and the European Systemic Risk Board today, and the recommendation of recent European Parliament work (van ’t Klooster, Martino & Monnet, 2025): build the public anchor, let banks tokenise, and do not weaken MiCA to nurture a riskier private class.

An innovation-and-competitiveness-first EU. If the overriding aim is a dynamic, entrant-friendly euro-stablecoin sector that can compete with dollar tokens now, the mix tilts to A — a narrow-money lane with calibrated central-bank-reserve access — plus B for scale. This is the “embed, don’t contain” prescription (Bruegel, 2026): give regulated issuers a safe public backing and let them compete on everything else. The catch is that this scenario’s defining move — reserve access for non-banks — is the very one that clears the highest legal and political bar (§4, §9a), so even an innovation-first EU may have to settle for the less-clean sovereign-bond variant of A in the near term — which, in the euro area, inherits the fragmentation of a market with no single safe asset.

A feasibility-first EU. If the binding constraint is what can be shipped quickly with existing institutions, the mix tilts to B — tokenised deposits and bank-issued coins — built on whatever wholesale central-bank settlement (the near-term slice of C) can be stood up fast. This is the lowest-risk path but the slowest to contest the dollar, and it leaves the competitiveness objective largely unmet.

The honest observation is that none of these is “the EU’s view,” because the EU does not yet have one — which is why the 2026 review reads as a contest rather than a decision.

9. A defensible synthesis (conditional, not categorical)

What follows is one conditional route, not the appraisal’s verdict — a sovereignty-first or stability-first Europe would sequence differently (§8). On today’s revealed preferences — the ECB and ESRB’s direction, and the market’s drift to bank issuance — the EU’s near-term trajectory leans toward lanes C-plus-B. That reading is interpretive, and the wider debate is messier: cross-border dollar tokens, multi-issuer schemes and big-tech entry all press on it. But within the euro-denominated, EU-regulated frame, the contested margin is how much room, if any, to leave for a non-bank narrow-money lane. With that framing, a sequenced portfolio (Figure 2):

1. Build the public anchor first, starting wholesale. On-chain central-bank money for settlement is the foundation that makes every other lane coherent: it gives bank tokens their cross-bank singleness and gives narrow tokens something to tether to. It is also the least politically fraught slice of Option C, and it directly addresses the sovereignty and dollarisation objectives. The contested retail digital euro can follow on its own timetable.
2. Let bank money and tokenised deposits carry private credit (Option B). This preserves transmission and reuses existing supervision, and it is where the market is already moving. The safety net it extends is a known quantity.
3. Optionally, add a narrow-money lane for non-bank issuers (Option A). The disclosure at the foot of this note bites hardest here: this is the author’s own corner, and a sceptical reader should discount the recommendation accordingly. On the merits, the case for the lane is competition and non-bank entry without taxpayer money standing behind private credit; the case against is the one this note has stressed — in its cleanest form it needs central-bank-reserve access the ECB resists on mandate grounds (TFEU), it is the sharp edge of disintermediation (a financial-stability risk, not merely a political price), and its business model is thin. The lane earns a place in the portfolio only if Europe actively wants non-bank entry and is willing to pay those costs. Absent that political choice, lanes C and B suffice.
4. Close the incoherent middle — concretely, and only the non-bank version. A bank-issued EMT already sits in coherent corner B, behind the safety net; the incoherence is specific to the non-bank EMT — par promise, credit-bearing reserves, no backstop. The concrete fix is to amend MiCA so that a non-bank issuer of a par-promising euro token must either hold 100% of its reserve as central-bank money or Level-1 high-quality liquid assets (moving it into lane A) or obtain a banking licence (lane B). Be clear about what that means: it would, in effect, abolish the free-standing non-bank e-money token as MiCA created it, pushing every par-promising euro issuer into either a fully reserved (A) or a bank (B) structure. That is a strong and contestable proposal — reasonable people will call it too radical — and it plainly benefits lanes A and B, the author’s own included; it is offered as the logically clean endpoint, not as a costless tidy-up. With that caveat stated, the underlying conclusion — that the no-backstop non-bank corner is the one design that does not hold — is robust to the objective-weighting; it is the one thing the appraisal settles unconditionally.

Underlying all three lanes is one variable: how freely safe, central-bank-anchored money is allowed to substitute for bank deposits. It is tempting to call this a single disintermediation dial, but it is really three distinct levers — reserve access for non-banks (lane A), holding limits on the digital euro (lane C), and the remuneration of safe balances — and they are not politically equivalent. A central bank will turn the dial for its own instrument (a digital euro hemmed by holding limits) far more readily than it will hand public-balance-sheet access to private, for-profit issuers. The levers must be set together but judged separately. Stated carefully, the EU’s choice is not among three monies but about how much of its safe-money demand to route through the public balance sheet, and through whom — and that is a question it should answer deliberately rather than by leaving each institution to turn its own lever.

9a. Hard edges this appraisal does not resolve

Three dimensions a policymaker will press on are flagged here rather than settled, because each deserves separate treatment:

• Legal basis. Both contested lanes run into the Treaty. Granting non-banks access to central-bank reserves (lane A) and fixing the digital euro’s design (lane C) are constrained by the ECB’s mandate and counterparty framework (TFEU Arts 127–128). These are questions of legal competence, not just policy preference, and they shape what is feasible — which is why lane A’s cleanest form scores low on feasibility above.
• The cross-border and FX dimension. The objectives table scores the international role of the euro, but the body keeps to the domestic system. Each lane affects the euro’s external use differently: a narrow-money euro token usable by non-residents is the most outward-facing instrument but also the most sovereignty-sensitive, while a digital euro’s cross-border use raises questions of capital flows and reciprocity this note does not model.
• Scale, AML and interoperability. The disintermediation magnitudes that would pin down several cells of the matrix are unmodelled here and must be estimated before any lane is sized. Anti-money-laundering obligations, and the technical mechanics of cross-bank settlement (atomic settlement, trigger solutions) that decide whether lanes B and C actually deliver singleness, bear directly on feasibility; they are out of scope for this appraisal but not out of mind.
• Who manages the competition. Calling this “managed” competition presumes a manager. Setting the three levers together casts the central bank as an active arbiter of competition between banks, non-banks and its own digital euro — a governance shift in its own right, which this note names rather than resolves.

10. What turns on this

The first note argued that calibration is the wrong task because Europe has not chosen its machine. This note adds that the choice is not binary and not even ternary — it is a weighting across three coherent lanes, governed by one dial, and the weighting is a statement of what the EU values. A sovereignty-first Europe, an innovation-first Europe and a stability-first Europe would each build a different, internally coherent system. What none of them would build is the thing the EU has now. The review’s real work is to make the weighting explicit and choose it on purpose — and then to build each lane in its coherent corner, rather than continue to run a single instrument that tries, and fails, to occupy all of them at once.

Disclosure: the author is the CEO and founder of Eurodollar, a regulated stablecoin issuer based in Denmark, and therefore has a direct commercial interest in the regulation discussed here. This note is written as an even-handed options appraisal rather than advocacy for any single design; readers should weigh the author’s interest accordingly. A longer, academic treatment of the underlying framework is in “What the Rule-Book Cannot See” (MPRA No. 129388).

References (verified June 2026)

Adrian, T., & Mancini-Griffoli, T. (2019). The rise of digital money. IMF FinTech Note No. 2019/001. (Journal version: Annual Review of Financial Economics, 13, 57–77, 2021.) Awrey, D. (2022). Unbundling banking, money, and payments. Georgetown Law Journal, 110, 715. Bruegel (2026). A new strategy to contain stablecoin risks in the European Union. Policy Brief 09/2026 (drawing on Bindseil, 2026). Cipollone, P. (2026). Stablecoins and the future of money: separating functions from instruments. Speech, European Central Bank, 8 May. Diamond, D. W., & Dybvig, P. H. (1983). Bank runs, deposit insurance, and liquidity. Journal of Political Economy, 91(3), 401–419. Goel, T. (2024). Reserve-backed tokens: a money for the future? Journal of Payments Strategy & Systems, 18. (Earlier BIS/SSRN version, 2023.) Gorton, G. B., & Zhang, J. Y. (2023). Taming wildcat stablecoins. University of Chicago Law Review, 90(3), 909. Kumhof, M., & Noone, C. (2021). Central bank digital currencies — design principles for financial stability. Economic Analysis and Policy, 71, 553–573. van ’t Klooster, J., Martino, E. D., & Monnet, E. (2025). Cryptomercantilism vs. monetary sovereignty: dealing with the challenge of US stablecoins for the EU. Study for the European Parliament (ECON), PE 760.274. Williamson, S. D. (2024). Deposit insurance, bank regulation, and narrow banking. Journal of Economic Theory.