Interview with Central Banking

Interview with Sharon Donnery, Member of the Supervisory Board of the ECB, conducted by Thomas Chow on 11 June 2026

23 June 2026

What do you see as some of the most prominent risks to financial stability and banking system resilience facing Europe right now?

Broadly speaking, over the last number of years, European banks have remained resilient through a number of shocks. This resilience is very important. But, of course, we cannot be complacent about the risks and shocks that are potentially on the horizon. I would say our main priorities at the moment include strengthening banks resilience to geopolitical risk, making sure that banks remediate outstanding issues in a timely manner, and making sure that banks have sound digital transformation and AI-related strategies in place.

Geopolitical risk is obviously very topical in the world now, and a number of things of significant concern have happened in the last few months. Given the way things have been changing over the last number of years and the shocks that we have experienced – including the Russian invasion of Ukraine – geopolitical risk has been much more on the radar. We do not see it as a new risk, because banks have been exposed to geopolitical risk for a long time. We want to ensure that banks really understand their exposures and how they feed through: whether through the real economy, through financial markets, or through security like operational disruption or defence issues. Banks also need to know what these exposures mean for their credit risk and market risk, etc.

The prominence we have given to geopolitical risk is illustrated by our reverse stress test, for which we ask banks to look at their own business model and exposures. The test is not about passing or failing, and it is also not about choosing the right scenario. It focuses on what really matters to banks’ business models and exposures, and how the banks would deal with a geopolitical shock. We will have the results of that test later in the summer.

I would next highlight the issue of timely remediation. In supervision, we give banks feedback and ask them to address certain issues. There are always things that banks must follow up on. And we have some long-standing issues – including for example risk data aggregation – on which we want to see banks make more timely progress, where we are also thinking about how to escalate that. So, there is considerable work to be done on that, and we are changing our approach to make it simpler.

The last key area I would emphasise is digital transformation, which encompasses many different aspects. There is what it means for banks’ business models, IT systems, cyber and operational resilience and third parties, as well as topical issues like AI, which have come more to the fore in recent months. It is up to banks to ensure they understand what digitalisation means for their business and whether they are adequately addressing their vulnerabilities. Digitalisation has picked up speed in recent months, particularly in terms of AI, and banks must keep on top of that. Of course, in Europe we have the Digital Operational Resilience Act (DORA), which applies since last year.

What was the thinking behind choosing a reverse stress-testing methodology rather than a regular one?

Every second year we participate in the European Banking Authority (EBA) stress test, which could be considered a more traditional stress test, i.e. banks are asked to respond to a given scenario. We have their capital depletion and other findings to follow-up.

In years when there is no EU-wide EBA stress test, the ECB conducts thematic stress tests. We conducted the climate risk stress test in 2022 and the cyber resilience stress test in 2024. And the 2026 thematic stress test to capture geopolitical risk is ongoing.

The geopolitical reverse stress test was driven by our serious concern about geopolitical risk. In the EBA stress test we had a scenario that looked at trade shocks and what it meant for the banks, but it was a standard shock across all the banks at the same time. We know the banks have very different business models, very different exposures to countries, currencies, etc. We wanted to focus more on individual bank exposures. We also wanted to foster the banks’ own capabilities for scenario planning, particularly for looking at geopolitical risk and carrying out their own stress testing. The reverse stress test was designed accordingly, and as I said, it is not pass-fail exercise.

We asked the banks to look at scenarios that would lead to a capital depletion of 300 basis points. They have come back to us with a lot of data and many different scenarios. We are collating all of that material, while following up on some issues with individual banks. In the summer we will publish a report that identifies some key trends and findings from the test, and each bank will receive feedback and follow-up from the Joint Supervisory Teams (JSTs). We expect that the feedback will be mostly qualitative. Let me stress that this is not about whether the banks identify the right scenario, because there are many different scenarios. It is more about the banks’ capability to consider scenarios and how they would recover. The world is extremely uncertain so our focus is on understanding how a bank would respond if something happened: what their board would do, what management measures would they take, etc.

Is there a concern that the banks would window-dress their stress-testing exercise? How do you ensure they come up with a scenario that is genuine and informs the supervisor of their most vulnerable weaknesses?

In all stress tests, including this one, we have checks and balances. There are basic things like data quality review and assessment. Another consideration is plausibility. The stress-testing team and teams composed of experts on each individual bank, check whether what they are seeing is plausible and reasonable. And of course, the ECB has the great advantage of directly supervising over 100 banks. We can look across all the banks. We can look at banks by individual business model, or we can look at a group of banks that have a particular type of exposure and make a peer comparison. This all helps to inform our assessment of whether the whole thing is reasonable and plausible. We have not got all the results yet – as I said we are in a dialogue with the banks – but no doubt we will need to follow up on a range of issues. We will communicate the overall results in the summer.

In December the ECB published a report outlining how it is streamlining supervision, and among them are major reforms to the Supervisory Review and Evaluation Process (SREP) and methodologies to calculate the Pillar 2 requirement (P2R). However, some industry stakeholders still complain that the final P2R decision – which relies on “constrained supervisory judgement” – could feel arbitrary at times. Does the ECB share this concern? And if yes, is it looking to improve transparency and communications regarding the P2R?

Let me first put the P2R in context. There is a big global debate at the moment about simplification, modernisation, competitiveness or whatever you want to call it. But back in 2022, even before this debate started, the ECB had appointed experts to review our SREP process. These experts took stock after almost ten years of European banking supervision and gave us a number of recommendations on reforming the SREP process, all of which we accepted.

Our “Streamlining supervision, safeguarding resilience” report, published in December 2025, includes our work to make our SREP process simpler, clearer and more transparent, as well as enable quicker feedback to banks to encourage more timely remediation. Almost all of the reforms have been implemented. In the wider debate about simplification, I think this report is an important element.

On P2R, we have reformed the methodology, and our approach is consistent with those themes: we wish to make our methodology clearer and more transparent, to make it more connected to the risks. Constrained judgement remains an important element. Having the flexibility to make judgements lies at the heart of banking supervision. But we do have constraints in the sense that we have internal quality assurance, internal benchmarking and second-line reviews during which various experts look at the judgements that have been made. And because we are also making our SREP communications with the banks much clearer, crisper and more focused, I would hope that the banks see a more direct connection between their capital requirements and our risk assessment of them. We are also working to make sure that overlaps between Pillar 1 and Pillar 2 requirements do not happen and that there is a clear process around that issue. Banks should feel assured that if something is unclear, they can discuss it with their JST or the senior management of the ECB.

How has the ECB’s model of horizontal expert teams supporting vertical supervision been working in practice, including during stress events? What further improvements could be made?

The ECB’s horizontal and vertical teams are absolutely crucial to how we work. We have the JSTs, which are responsible for the day-to-day supervision of individual banks and are composed of staff from the ECB in Frankfurt and the relevant national supervisors. We have our on-site inspectors, who are dedicated experts who carry out targeted, in-depth supervisory analyses. However, we are also performing more short and quick reviews in some cases. While we will retain the big deep-dives, we are trying to become somewhat more agile and have some of our inspectors working on shorter, quicker and more focused inspections. In addition, we have horizontal expert teams, such as our stress-testing team, which are a fundamental part of our set-up. Overall, the structure has proved effective, as it combines a bank-specific perspective with specialised expertise. We have experts who can step back and carry out in-depth inspections as well as experts performing thematic work across the banks.

The feedback from the banks – particularly over the last couple of years – is that this may have become a little cumbersome. We touched on the SREP, but we also have a wider reform programme called “next-level supervision”. An essential part of that is making sure we are more coordinated across all teams. Banks need to have a clearer understanding of the work we do, no matter what team is involved. As part of our reforms, we are streamlining and improving the coordination of data requests, and we make it simpler and easier for banks to understand how we work.

What efforts has the Supervisory Board made to upgrade supervisory methods to match the velocity of modern banking – such as instant payments, digital asset tokenisation, T+1 settlements, instant balance transfers, etc.?

There are two aspects to it. One is understanding the banks and what is happening in the banks. And the second is the ECB’s own technology. First, on the banks, the topic of digitalisation – of the financial system, of the economy, AI, cyber – we have been working on these issues for a number of years. Some of our teams have been working on really understanding the banks, also in close collaboration with the central banking side of the ECB, which is working on payments, the digital euro and blockchain, among other things.

Then there is technology used for supervision, known as suptech, which we have been focusing on for four or five years now, if not longer. It is about how we use suptech tools, for example in dealing with the vast quantities of data we have. Another element of the next-level supervision project is streamlining our processes: making fit and proper procedures quicker and simpler and fast-tracking decisions for lower-risk capital transactions. Using technology helps our supervision in two important ways. It helps banks because it is easier for them to engage with us and thus get quicker decisions. And it frees up our teams from the clunky data processing work so they can spend more time on core supervision matters, and on the challenges related to emerging risks and things that we need to keep up with.

What efforts has the Supervisory Board taken to ensure banks and their IT systems remain resilient to new cyber security threats – Mythos, attacks from state-actors, etc – as well as resilience related to outsourcing IT to third parties?

If we go back to just after the global financial crisis, there was a huge focus on capital and liquidity – on rebuilding the financial resilience of the banking system. Over the last couple of years, although financial resilience has continued to be a core, important issue, non-financial resilience – and particularly operational resilience – has come much more to the fore in all of the elements you mentioned, whether cyber, reliance on third parties, digitalisation of banks’ business models or investing in new IT systems. It has become much more prominent in our supervision as well. In Europe we have also been implementing DORA together with colleagues at the EBA, meaning banks have had to get much more on top of all these issues.

I think what has happened with AI in recent months is a real change of pace. We knew AI was getting better and moving quickly, but there has been a significant change over the past few months in terms of what we expect from some of the models. But in the end, dealing with this still comes down to good risk management – to banks being on top of their systems and change management issues as well as their cyber vulnerabilities. We are engaging with the banks. We had a seminar with a number of banks a couple of weeks ago and we are following up with banks to make sure they are looking at these issues and dealing with them. In the end, these issues need to be taken forward in the context of banks’ governance and risk management systems. A big consideration, especially in recent years, is recovery capacity: if something goes wrong, how will a bank recover?

Moving on to regulation – the ECB has suggested introducing more proportionality in the framework and raising the current €5 billion threshold at which Basel rules would begin applying in full to banks. Is there a concern that this might sow the seeds for a future crisis – not unlike how the 2018 Dodd-Frank amendment in the United States contributed to the failure of Silicon Valley Bank in 2023?

In this simplification and modernisation debate, the topic of proportionality has been raised quite a bit. The existing framework – the “small and non-complex institution” (SNCI) regime you mentioned – already allows for proportionality. In many cases, the national supervisors supervise the smaller banks and the ECB supervises the bigger ones. Over the past few years there have been efforts to make the system more proportionate. Our assessment is that even within the existing framework there is room for more proportionality in areas like data and other requirements. We have suggested raising the current €5 billion threshold, and it is now up to the European Commission to decide on proposals.

We are very conscious that, even if the threshold were raised, requirements would still apply to the banks below that threshold. You must always take into account the nature, scale, complexity and risk profile of an individual bank. If a national supervisor or the ECB is concerned about a smaller bank, it is still part of our work to make sure this concern is followed up on. It is a very important issue for the ECB as well because we have an oversight role over the smaller banks – the less significant institutions. We work with our colleagues at the national supervisors to make sure that we are looking at the highest-impact issues in the smaller banks. Let’s see what the European Commission recommends, but I think raising the threshold will be subject to adequate safeguards to make sure that we do not have problems in the future.

The ECB has also proposed reforming the Additional Tier 1 (AT1) regime. Specifically, it suggests that the loss-absorbing capacity of AT1 instruments in going concern could be “enhanced” and “clarified” or that AT1 could even be fully replaced by Common Equity Tier 1 instruments. What is your view on that? And for the former option, how might such a reform look like in practice?

This recommendation was part of a wider set of recommendations that we made to the European Commission about potential legislative reforms. A number of these were related to capital. One was on the capital stack, which has become quite complicated at European level, and another proposed taking a more holistic view of the capital stack. We raised all three issues – including AT1 – in light of the prominent debates surrounding them. People have been talking about AT1 and whether it can be used effectively in the system, so the recommendations came from that.

AT1 is the one area in the ECB’s Governing Council proposals on simplification where we had different options. We did acknowledge in the report that the option of phasing out AT1 could be problematic and may not be Basel-compliant, so it would require further work and discussions. The issue was raised because people were talking about it, and we think it is an issue that particularly relevant among investors, for example. Something may need to be done to address this, either a clarification of the role of AT1 or something else. We will wait and see what the Commission recommends in its forthcoming report.

The United States has abandoned the Basel III output floor (for internal models used to calculate market risk) in its most recent package of the post-crisis capital reforms. What do you think the EU’s response should be, given the capital relief that large US banks might get from this rule change?

At the European level, it is very important that we have our own view on our rules, our supervision, what we want to do, and how it applies. We must make our own decisions. The US proposal is subject to consultation, so I should not comment on it. But as we have consistently emphasised at the ECB, including in our Governing Council proposals on simplification, the implementation of international standards is essential. From the ECB’s point of view, we remain very committed to Basel and its implementation. Anything we do at the European level should be in the context of our assessment of our framework and our banks and their exposures.

I hear a lot of commentary about whether European banks are at a disadvantage. There may be slightly different issues or nuances that mean you cannot really think about competitiveness in relation to individual banks per se, but I do know that for the biggest banks, a level playing field is very important. That is why we remain very committed to the international standards overall. And, as discussed, the Commission is actively looking at all of these issues.

More broadly speaking, are you worried that the deregulation drive in the United States is putting EU banks and financial institutions on the back foot? How do you ensure the European banking sector remains competitive on the international landscape?

As I said, in Europe our focus is not necessarily to get into a big debate with other jurisdictions. We have been very clear at the ECB that, at European level, the debate about simplification and modernisation should not be about deregulation. It should not weaken resilience. We have spent a large part of this interview talking about the risk landscape, which remains complex and uncertain. So the resilience we have built up has been very hard fought, and we need to preserve it. We have been clear that, for us, it should not be about deregulation and it should not be about lowering resilience.

We accept that there are undue complexities that can be reduced in both the regulatory framework and our supervisory approach. The framework can be made simpler and clearer, and our supervision can be more risk-based, efficient and effective. These are the things that we should focus on, because we believe that in the current risk environment, resilience will be really important in the future.