The exit test: why the single market in banking depends on how banks fail

Keynote speech by Pedro Machado, Member of the Supervisory Board of the ECB, at the A&M International Banking – crisis management conference

Madrid, 1 October 2026

This forum provides me with the opportunity to connect two topics I have discussed at other events^[1] in recent days; namely, crisis management and competitiveness. And this is the right place to ask difficult questions.^[2] Two in particular: first, how can we tell whether a crisis management framework works before a real crisis puts it to the test? And second, what does a sound crisis management regime contribute to the resilience – and, I will argue, the competitiveness – of the banking system?

The title of this speech comes from the economist Albert Hirschman.^[3] He described two ways people react when an organisation declines. They can exit, meaning leave, or they can use voice, meaning stay and push for change. Standard thinking about competition relies almost entirely on exit: unhappy customers leave, and that pressure forces firms to improve. But Hirschman added a third term, “loyalty”, and saw it as a virtue: it holds the customer back and gives voice a chance to work. In banking, the loyalty of depositors and counterparties rests on trust. When that trust is eroded, a crisis management regime must enable an orderly exit through credible use of the resolution toolkit.

In normal times, depositors and clients can and do change banks when a better offer appears. But in a crisis the exit of customers can become a run, and a run destroys the competition an exit is supposed to drive. That is what a resolution regime supplies instead: an orderly, administered exit of the firm, in which shareholders and creditors absorb the losses and contribute to recapitalisation, while critical functions continue. I will thus focus on how effectively Europe has organised the crisis management of banks. Everything I will say hinges on a single variable: trust in how failure would be handled.

I now want to examine one document with you through that lens: the European Commission’s Communication of July this year on the competitiveness of the banking sector and the single market in banking^[4], a key deliverable of the savings and investments union strategy^[5] and the blueprint for the legislative banking package announced for the first quarter of 2027. Overall, the Communication is largely a going-concern document: it is about scale, efficiency, integration and simplification. No less importantly, it also devotes a chapter to crisis management, and it is that chapter I want to focus on. My argument is simple: a competitive banking sector needs sound prudential rules, but it also needs credible crisis management. Crisis management is not a cost Europe pays in exchange for integration. It is a driver of resilience and competitiveness in the banking union.^[6]

The Commission’s diagnosis and its prescription

Let’s begin with the diagnosis, which is candid. European banks are sound, well capitalised and liquid. Through the pandemic, the energy shock and the market turmoil of March 2023, they acted as shock absorbers rather than amplifiers. And yet the sector is underperforming its potential. The Commission pulls no punches when stating that the principal reason for this is fragmentation along national lines. Banking markets are still organised at the national level, consolidation is overwhelmingly domestic, and cross-border corporate lending in the euro area still only accounts for around one-sixth of total lending. In an economy where banks provide the bulk of corporate debt financing, a sector that cannot achieve scale across the Single Market constrains Europe’s capacity to fund its own priorities and limits banks’ ability to offer households and businesses better products at lower prices.

To foster integration, the Commission has announced a series of measures that go to the heart of how cross-border groups are regulated.

First, the Commission wants to create the conditions for groups to allocate capital and liquidity more efficiently across the EU by revisiting a framework in which requirements apply not only at consolidated level but also, entity by entity, at individual level. Today, a group operating in several countries must meet these requirements twice: once for the group as a whole, known as the consolidated level, and again for each subsidiary on its own, known as the individual level. In practice, this means that a subsidiary in one country must keep its own buffer of capital and liquid assets, even if the parent company or another subsidiary could step in to support it. The ECB estimates that the resulting constraints lock up €230 billion of high-quality liquid assets in cross-border subsidiaries, hence resources trapped, in effect, by geography.^[7]

The waivers that would lift individual requirements exist in the legislation^[8], but they have so far very rarely applied across borders. The Commission will also propose measures to ensure that parent banks are credibly committed to moving resources to subsidiaries in a timely manner in case of need. It also plans to review the treatment of intragroup exposures so that a claim on a parent across an internal border is treated no less favourably than one within a Member State.

Second, the Commission has announced measures to further strengthen coordination between microprudential, macroprudential and resolution requirements. This stack of requirements has grown in layers – while each layer is rational in isolation, they do not always form a coherent whole.

And third, the Commission envisages a new proposal to review and simplify the structure of the deposit insurance framework, better aligning the responsibilities and the financing of crisis management and depositor protection. This would replace the existing legislative proposal on the European deposit insurance scheme (EDIS), which has been on the table for a decade.^[9] The Commission is also signalling that it will have a firmer hand when Member States interfere with cross-border consolidation without justification.

Alongside these integration measures, but separate from them, is a broad simplification of the regulatory framework – of the capital stack, of reporting and of the layers of requirements accumulated since the crisis. I will come back to this, because the logic behind this simplification is different from that of the integration measures.

This is, by any standard, an ambitious agenda. The question I want to pose is a simple one: what would make it work? Under what conditions can these measures be legislated, granted and used, rather than remaining, like the existing cross-border waivers, available in law and avoided in practice?

The three tests of a single market

To answer, let me transpose Hirschman’s own framing from the customer to the bank. A single market in banking has to pass three tests. First, a test of entry: can a bank established in one Member State do business in all the others? Europe answered that long ago, with the single licence and the passport. Second, a test of operation: are banks subject to the same rules, applied by the same standards, wherever they operate? The Single Rulebook and the Single Supervisory Mechanism have largely answered that too, although more can – and should – be done to harmonise prudential and non-prudential rules across Member States. Third, a test of exit: when a bank fails, is its failure handled in the same way, with the same instruments, at the same level, regardless of where it happens? Here the honest answer is: only partly. Resolution powers have been transferred to the European level and are exercised within the Single Resolution Mechanism. Yet the creditor hierarchy remains national.^[10] The going-concern regime is European, but much of the gone-concern regime is still national. A single market is only as single as its treatment of failure.

And that is why the market has organised itself the way it has. Local capital, local liquidity and pre-positioning are pricing the framework as it stands. Fragmentation in going concern is, to a large extent, the shadow cast by fragmentation in gone concern.^[11]

Every waiver will be assessed during an exit

Now read the Commission’s prescription against that background, measure by measure, because each of them draws on the same account.

Start with capital and liquidity waivers. What does a national authority do when it waives individual requirements for a subsidiary? It gives up resources under its jurisdiction today in exchange for a promise that resources will flow from the group tomorrow – precisely when tomorrow is bad. That is why the existing waivers have remained unused across borders: not because the legal text is unclear, but because the promise is discounted.^[12] If reliance on the group fails, the Member State concerned is left with a national insolvency, a national deposit guarantee scheme and a national political crisis. The waiver debate is not a prudential debate that happens to touch crisis management. It is a crisis management debate conducted in prudential vocabulary.

This starting point brings us to the first measure: how to create the conditions for national authorities to accept a legal claim on the parent in place of resources held on their own territory. Whether that claim is worth accepting depends on two exit questions. First, speed: can the obligation be enforced when it matters, against a parent that is itself under stress? And second, rank: if enforcement fails and the group’s entities enter resolution or insolvency, where does the subsidiary’s claim stand, and under whose law? An intragroup support commitment is only as good as the failure regime behind it. It may be legislated in going-concern language, but it will be tested in gone concern.^[13]

In resolution planning, the question of how much loss-absorbing capacity must be pre-positioned in subsidiaries through internal minimum requirements for own funds and eligible liabilities (MREL), rather than held at the resolution entity, is the gone-concern counterpart of the waiver debate. Where national authorities trust a single-point-of-entry strategy – losses up-streamed, recapitalisation down-streamed – they accept less pre-positioning; where they do not, they demand more. Going-concern waivers and gone-concern pre-positioning are two dials calibrated by one variable: trust in how the group would be handled if it fails. It would be incoherent to turn one dial while the other stays in place.

This has a practical consequence. A cross-border waiver is formally a supervisory decision, but it has a bearing on the resolvability of the group. The granting of a waiver will, in practice, be relevant for both the supervisor and the resolution authority – each waiver has to be assessed against the group’s resolution strategy, and each strategy has to be updated for the waivers that are granted.

The Commission’s remaining measures confirm the pattern. The second – treating a cross-border intragroup exposure like a domestic one – asserts that a claim on the parent is equally good on either side of the border, which is, fundamentally, an assertion about the law governing the parent’s failure. The third – coordinating microprudential, macroprudential and resolution requirements – cannot be done without the resolution authority at the table, because MREL is one of the layers being coordinated. And the fourth, deposit insurance, is not adjacent to crisis management – it is crisis management.

So when the Communication turns to crisis management – implementation of the reformed crisis framework, liquidity in resolution, a way forward on deposit insurance – it is not adding a chapter for completeness. The going-concern proposals of the 2027 package rest on the soundness and credibility of the gone-concern framework. The question that follows is obvious: is this actually the case?

Before answering this, I should deal with an objection linked to the very nature of the decision-making process, most cogently put forward in recent work at the Financial Stability Institute of the Bank for International Settlements, and which may impact its credibility.^[14] The critique is that the European resolution machinery is uniquely complex: the resolution procedure involves the SRB, the ECB, the Commission, potentially the Council and national authorities, whereas the United Kingdom and the United States rely on a single empowered authority. This criticism deserves to be taken seriously, and my response is twofold. First, the record shows that the machinery, complex as it is, can come to a decision swiftly: neither the Banco Popular nor Sberbank cases were delayed by the number of decision-making steps required.^[15] Second, where complexity cannot be removed without changing the Treaty, it can be managed – by preparation, by cooperation arrangements and by the legal clarity the courts have brought to the matter of who decides what, and when. While complexity can result in inefficiency, it need not mean a lack of credibility. So the right question to ask is not whether the machinery is too complex to be trusted. It is whether the rules and resources behind it are complete. And that is where the gaps are.

Completing the banking union to support competitiveness

Four gaps remain, and each represents a border at which the Single Market currently stops.

The first of these gaps is already being narrowed. This spring, the reform of the crisis management and deposit insurance framework – the CMDI package – was published in the Official Journal.^[16] It further clarifies the path from supervisory action to resolution, anchors the ECB’s early intervention powers in directly applicable European law and makes the escalation path, and the cooperation duties between the ECB and the SRB, more predictable. In a word, it strengthens the continuum from supervision to resolution.^[17] It also opens the resolution toolkit, with its European governance and funding, to a wider set of significant banks, notably medium-sized, deposit-funded institutions that were previously left to divergent national liquidation procedures. That is itself a de-fragmentation measure, and its effectiveness now depends on implementation. This being said, the reform stops short of a full solution.^[18] Access to funding for transfer strategies remains hedged by complex conditions and size thresholds, and MREL cannot be calibrated on the basis of prospective external support.

The second gap is funding – liquidity above all, but not liquidity alone. Alongside robust capital and liquidity ratios, the build-up of loss-absorbing capacity in the banking union is one of the underappreciated achievements of the decade. The SRB’s latest MREL dashboard, covering the end of 2025, shows banks meeting an average final target of close to 28% of risk-weighted exposures, with an aggregate shortfall of €0.2 billion confined to a handful of banks that are still within transitional periods.^[19] But a bank recapitalised over a weekend still needs to fund itself when markets open. It may have lost access to market funding for weeks; it may no longer hold eligible collateral for central bank facilities; and the amounts involved in a systemic case can exceed anything an industry-financed fund can bear.

What is needed is a European mechanism that is able to bridge that period, backed by a guarantee that is ultimately fiscal. That mechanism does not exist. The issue is back on the European agenda, and rightly so, but the difficulties that stalled it before have not gone away.^[20] Unlike the United States and the United Kingdom, Europe has no dedicated resolution liquidity facility. The agreed European Stability Mechanism (ESM) backstop to the Single Resolution Fund, significant as it will be^[21], is not the full answer under extreme stress.^[22] Beyond liquidity, external funding in resolution is deliberately constrained: the Fund can be tapped only after shareholders and creditors have absorbed losses of at least 8% of total liabilities and own funds, and its contribution is capped at 5%.

The corollary is that Europe leans heavily on internal loss-absorbing capacity – on MREL – because external funding is scarce.^[23] Other jurisdictions have recently made the opposite choice, pairing dedicated funding for transfer strategies with simpler loss-absorbency requirements. But what matters most for today’s argument is the feedback loop into the Commission’s own agenda: as long as the ultimate backstop remains national, national authorities will tend to resist precisely the liquidity waivers the package will propose.^[24]

The third gap is deposit insurance. The absence of a common scheme perpetuates asymmetries in depositor protection and prevents banks and investors from fully pricing risk across the Single Market; less than 2% of euro area household deposits are held across borders.^[25] National schemes also differ in what their funds may be used for, notably for preventive and alternative measures outside resolution – a further layer of divergence at the moment of failure, only partly addressed by the CMDI reform. To its credit, it approaches deposit insurance from the crisis management side, as part of the financing of failure rather than a consumer protection add-on. But depositor protection is where the Single Market’s promise is of the essence, and a safety net that stops at the border will not persuade savings to cross it. This issue belongs at the centre of the competitiveness agenda.

The fourth gap is the divergence of national insolvency regimes. The protection of creditors in resolution is benchmarked against the treatment they would have received in insolvency, and that counterfactual varies with each national regime, hence complicating the no-creditor-worse-off assessment for cross-border banks and letting the public interest assessment itself yield different outcomes. The outcomes thus depend, in part, on where a bank happens to fail: fragmentation in its purest legal form. And because national insolvency proceedings face fewer funding constraints than resolution, the incentives at the moment of failure may still, in some cases, point away from the common framework rather than towards it.^[26] Convergence is demanding but no longer uncharted: the UNIDROIT Legislative Guide on Bank Liquidation, adopted last year with the Bank for International Settlements’ Financial Stability Institute, may offer legislators a sound blueprint.

Let me compress the argument into the one sentence I would like you to take away from my speech today: trust in how banks fail unlocks freedom in how banks live. And let me be precise about whose trust I am talking about. It is the trust of national authorities and, behind them, of national parliaments and finance ministers – those who will answer for a failure whose cost falls on the host rather than the home. Behind this trust stands the EU’s fiscal architecture: where large resources are at stake, Europe still relies on national budgets, which limits European discretion and raises questions of legitimacy that no technical framework can answer alone. Arguably, further integration measures will gain broad acceptance to the extent that the framework for managing bank failures becomes sufficiently common, credible, and comprehensive that the location of a bank’s distress no longer determines who bears the cost. This is how effective crisis management contributes to competitiveness.

Conclusion

Let me conclude with two observations.

The first is about how the 2027 package should be read: as one package, not two. It is sometimes suggested that Europe could take the efficiency measures first and leave the safety net measures for a better political season. I know that for the industry the efficiency that matters most lies elsewhere – in the simplification of the capital stack and of reporting – and I do not underestimate it. Simplification can and should proceed on its own merits. Within their mandates, the ECB and the SRB are already heading in this direction, with common planning where requirements overlap, streamlined approvals and faster joint handling of own funds reductions and MREL decisions.^[27] But the integration measures are different. Their value is conditional on the safety net: they are secured on its collateral. Legislating the former without the latter would produce rights that exist on paper but are avoided in practice.^[28] That is the package that must not be split.

My second and final observation is the following: as much as the debate on European banking competitiveness can be approached from a global perspective, failure management remains anchored in jurisdictions, and every cross-border resolution ultimately rests on the credibility of regimes and on tested cooperation. Europe’s attempt to convert crisis management credibility into market integration is, in that sense, not a peculiarity but a precondition for any meaningful integration. And the dynamics can run in both directions. Each step that makes failure management more predictable increases the trust that national authorities can rationally extend, whether this is the CMDI reform being implemented in full, the ESM backstop being ratified, a framework for liquidity in resolution being agreed, waivers being assessed jointly by supervisory and resolution authorities, insolvency regimes converging or a common deposit insurance scheme with a timetable being implemented. Trust enables waivers; waivers let groups operate as genuine groups; and groups that operate as genuine groups make a common safety net cheaper and more legitimate because the risks it covers are already shared. That is a virtuous circle that is within reach. It requires no leap of faith, only a sequence of steps, each of which makes the next one easier.

Let me end where I began. Hirschman taught us that exits instil discipline. In banking, Europe has learned that the exit process must be properly organised, and that a single market is only as single as the exit it can guarantee. A bank whose failure is managed anywhere in the EU on the same terms is a bank that can be trusted to operate anywhere in the EU. That is the exit test, and Europe has not yet passed it. But it has never been closer to doing so, and the 2027 package is the key. It can make the banking union a union that is trusted in even the most challenging times and, therefore, a union that represents an appealing choice in good times.