Speech by Andrea Enria, Chair of the Supervisory Board of the ECB, at a dinner of The Institute of International Finance, in Washington, D.C., USA, 12 April 2019

For a long time, cross-border banks had been the showpieces of a global financial system. Then, in 2008, they turned into the poster children for a global financial crisis. And with this, the benefits of cross-border finance were called into question. It became clear that a global financial system can produce crises that wreak havoc on a global scale.

So what should be done? How can we benefit from global finance while making future crises less likely? Over the past decades, many answers have been found and turned into actual policies. Basel III, for instance, has been a big step forwards; it provides a global set of standards to make banks safer and sounder.

But no bank is immune to difficulties, of course. So what do we do if a bank gets into trouble? And what if it’s a large, cross-border bank? Before the crisis, these questions did not receive enough attention. And this brings us to one of the core lessons from the crisis. Yes, making banks safer and sounder helps; it reduces the probability of default. But a default might still happen, so we must be able to mitigate its impact. All banks, including global ones, must be resolvable. When push comes to shove, even large cross-border banking groups must be able to fail. They must be able to fail without triggering another global crisis. This is the bottom line, if you will; this is the goal.

The financial crisis – lessons on crisis management

Ten years ago, we were far away from this goal. In 2008, large cross-border banks got into trouble left, right and centre, and no one had a proper plan for dealing with them. Both banks and supervisors haphazardly stumbled into the crisis.

That said, preparing for a crisis is a huge challenge. Handling a cross-border bank in distress is a complex task. Take Lehman Brothers as an example. In 2008, it had 8,000 subsidiaries and affiliates all over the world. When the parent filed for bankruptcy in the United States, it triggered another 80 insolvency proceedings for subsidiaries in 18 different countries.^[1] As we all know, this made for a rather messy failure with dire consequences.

To add to the challenge, no two cross-border banks are alike. So there is no blueprint to help authorities chart a course through a crisis. This, in turn, means that rules can only take us so far, and judgement becomes quite important.

But besides the technical complexity, there is one core question that needs to be answered when a bank fails: who pays the bill? As the saying goes, success has many fathers, while failure is an orphan. Banks are no exception. When a bank does well, many people take credit for its success and demand their fair share of the profits. But when a bank fails, no one wants to take the blame, let alone foot the bill.

This is important on two levels. First, there is the issue of bail-in versus bail-out. Which stakeholders will bear the burden of the failure? And second, there is the issue of home country versus host country. In which jurisdiction will the burden be borne?

During the crisis, large banks that got into trouble were often bailed out with taxpayers’ money. The goal was to keep these banks afloat so that they would not bring down the entire system. So once the banks fell ill, their medical bill was paid for with public money – and sometimes their funeral was paid for too. This created all the wrong incentives for banks and put an undue burden on society. Something had to change.

Likewise, it became clear that home and host countries needed to rethink their treatment of cross-border banks. Banks that take on business and risks at the global level cannot be taken to a national sickbed when things go wrong. So the two core questions are: how do we ensure that cross-border banks can fail without bringing down the financial system and without burdening taxpayers? And how can we ensure that banks that are global in life are global in death?

Global in life, global in death – how far have we come?

Over the past ten years, a great deal has been done to answer these questions. A comprehensive toolkit has now been developed at the global level, which was a major advance.

Moving from bail-out to bail-in, for instance, was crucial in order to correct banks’ incentives, preserve financial stability and protect taxpayers. The new tools – total loss-absorbing capacity (TLAC) and the minimum requirement for own funds and eligible liabilities (MREL) – play core roles here.

But whether the bail-in approach can really succeed depends on four things. First, the market for bail-inable debt must be large enough, and all banks must be able to tap it. Second, bail-inable debt must be distributed across group entities based on needs. Third, bailing in creditors must not cause a chain reaction of financial trouble, so attention must be paid to who actually holds bail-inable debt. And fourth, those who might be bailed in must be able to understand how the new system works.

This last point is crucial, in my view, as it requires us supervisors to act. We have to become even more transparent. We have to better explain our principles and policies. And we will do so, because only then will investors and creditors be able to anticipate our actions, develop trust and contribute to more stability.

And as a more general point: the toolkit needs to be implemented in all countries. If this does not happen, cross-border banks will continue to pose a problem. And here we still have some way to go.

This brings us to home and host countries and their relationship. There are two basic options. First, one could assume that, in a crisis, cross-border cooperation tends to break down and banks die a national death. Under this assumption, authorities would have to require banks to segment their business along national lines, so that they could be resolved with local legal tools.

Second, we could assume that there is scope for home and host authorities to cooperate, not only in good times but also in bad times. Under this assumption, we would need to develop tools for jointly managing cross-border banks that have got into trouble.

It was the second option that policymakers began to pursue right after the crisis. And they have indeed made progress over the past ten years: recovery and resolution plans, resolution colleges, the powers to make banks resolvable and internal TLAC have all become part of the toolbox.

But in spite of these global efforts, there seems to be a lingering fear that, at the end of the day, banks might still be very much national in death. Thus, countries put up fences in order to shield local shareholders, creditors and taxpayers. And here, I’m talking about measures that exceed the prudential standards that were agreed upon at the global level.

Such ring-fencing could either be a permanent policy or an ad hoc one. It could either mean that foreign-owned banks always have to hold certain amounts of capital and liquidity inside the country, or it could mean that they only have to do so in times of crisis.

Ad hoc ring-fencing is often borne out of uncertainty. But it actually creates uncertainty in and of itself. Home and host authorities might have agreed to cooperate in a crisis. But when the day comes, such agreements may or may not hold, and ring-fencing may or may not happen. In the end, it’s a prisoner’s dilemma, so it is quite likely that ring-fencing will happen. “Mutually assured fragmentation”, some have called it.^[2]

This phrase indeed captures the main consequence – ring-fencing stands in the way of integrated markets. Cross-border banks are divided up along national lines, with some entities of the group ending up on one side of the line and others ending up on the other side. So cross-border banks are not seamless structures.

And this comes at a cost. First, there are short-term costs. Ring-fencing and the lack of trust it implies can amplify the costs of any crisis. As in the actual prisoner’s dilemma, cooperation might not be the dominant strategy, but it would make everyone better off.

Second, there are the long-term, structural costs. Capital and liquidity are most efficiently allocated when they can flow freely. Ring-fencing obviously blocks such a free flow, even within a single banking group, meaning that cross-border banking groups are less able to provide an intragroup market for capital. This might, in turn, force some of the group’s entities to revert to local markets, at higher costs.^[3]

At the same time, cross-border banks might have to hold more capital than would otherwise be necessary.^[4] The catch is that this does not even make them more resilient. In fact, it can have the opposite effect: ring-fencing might make cross-border banking groups less resilient as their entities become less able to support each other in times of crisis. And this idea can be taken one step further. If there’s ring-fencing, the banking sector as a whole becomes less able to act as a shock absorber. When the banking sector is segmented along national lines, a shock that hits one country will have to be absorbed within that country, putting a huge burden on its economy. But if the banking sector were more integrated, the same shock would be diffused across borders and carried by many shoulders. Such private risk-sharing can act as a very effective buffer against crises.

In this context, the Vienna Initiative is a case in point. In 2008, there were concerns that cross-border banks might pull out of central and eastern European countries. Given these banks’ importance for the region, their departure would have spelled disaster for the affected countries. This was when cooperation happened just when it was needed the most. Facilitated by bodies such as the International Monetary Fund and the European Bank for Reconstruction and Development, a meeting was held in Vienna that included relevant banks and their supervisors: most of the participating banks then committed to continuing to support their subsidiaries even if market conditions deteriorated. Host authorities agreed to refrain from ring-fencing and this greatly helped dampen the crisis in central and eastern Europe.^[5]

The European banking union – on the right track

The Vienna Initiative shows that cooperation is possible. It shows that a lack of trust can be overcome. Of course, this initiative sprung up on the spot during a severe crisis. But can we agree to cooperate even before a crisis happens?

This is what the European banking union is aiming to achieve. It seeks to hardwire trust and cooperation into a legal framework. So, where do we stand in Europe? Has the banking union helped to bring about banks that are European in both life and death? Has it helped to generate trust?

Well, we have come a long way.

First, European banking supervision ensures that banks are supervised according to the same high standards across the euro area. Most importantly, we have harmonised the main tool of banking supervisors, the Supervisory Review and Evaluation Process. All over the euro area, supervisors now apply the same tool in the same manner. So right from the start, countries have much less scope to ring-fence. European banking supervision helps to level the playing field for banks. At the same time, banks can more easily operate across borders as they no longer have to deal with different supervisory regimes.

Second, there is the Single Resolution Mechanism, the second pillar of the banking union. It ensures that banks can be resolved at the European level. So in Europe, we have set up an institutional framework that clearly defines how we work together – in good times and in bad times.

However, all this has not taken us past the finish line. The European banking market remains segmented along national lines; it is not a domestic market from the banks’ point of view, and it is not a single jurisdiction from the regulators’ point of view.

Some 30 years have now passed since the single passport for banking services was launched, the idea being that banks would set up branches in other countries or provide services remotely, without local establishments. In fact, we saw very little of that. Given the importance of being close to clients, having local knowledge and benefiting from well-established brands in retail banking, most banks opted to enter the market in other Member States by acquiring local banks – subsidiaries are thus dominating the scene.

This may be changing, though. In the future, I think that new technologies will reduce the relevance of branch networks and make it easier for banks to offer services remotely, also across borders. But at the same time, we have to deal with the underlying sources of fragmentation; and these are both a result of and a problem for crisis management.

First, there is “legacy fragmentation”. The national legal systems have not yet caught up with the banking union. And this makes it harder to manage crises at the European level. Think of bank insolvency laws. In the euro area, insolvency laws are still very much national and thus very diverse. And here is the problem: if the resolution of a failing bank is not deemed necessary in the public interest, that bank becomes subject to liquidation under national laws. In this case, banks belonging to cross-border banking groups die a national death – or rather many national deaths given the legal differences across countries.

The first-best solution would, of course, be to move on and harmonise national insolvency laws. But as you know, the first-best solution is not always the most feasible one. And in any case, what is deemed “first-best” lies in the eye of the beholder. So, I would not count on harmonised insolvency laws coming any time soon. Still, a debate has started about giving the Single Resolution Board an administrative liquidation tool. This is a debate I very much welcome.

The second type of fragmentation is true ring-fencing. Take for example the capital and liquidity waivers for cross-border intragroup exposures. Given the new European approach to supervising and resolving banks, such waivers would make sense within the banking union. But many people think otherwise, arguing, for instance, that without a fully integrated European safety net, including a single deposit guarantee scheme, national taxpayers might still find themselves picking up the tab. This means that, even within the banking union, countries still feel the need to lock in capital and liquidity.

What else can be done?

So the key question is: how can we increase trust at both the European and the global level? How can we foster cooperation, and how can we reduce the need to ring-fence? In other words: how can we get from “mutually assured fragmentation” to “mutually assured cooperation” or a “fully integrated single jurisdiction” in the banking union?

In Europe, one obvious step would be to have more formal reassurances about the institutional instruments at our disposal in times of crisis. Agreeing in principle on a common backstop to the Single Resolution Fund has been a step in that direction. And a European deposit insurance scheme would be another step in the right direction. Against this backdrop, I find the current debate rather lopsided as it keeps revolving around the sharing of risks. But there is more to this issue than that – European deposit insurance would also help to build trust and thus reduce risks. This point of view is often lost.

Another step we could take is to acknowledge that the concerns of host authorities might be justified. For example, we could take them into account when setting prudential requirements for entire banking groups. If a local risk does not diversify away or net out in consolidation, it needs to be captured in group requirements. This might help to limit the risk at the national level and reduce the need to ring-fence.

But we also need to prepare for crises, of course. So we need to agree on how to deal with cross-border banks that have got into trouble – and stick to such agreements when times turn bad.

In my view, recovery and resolution plans could become key in this regard. They could help home and host authorities to strike credible agreements on how to work together in a crisis, before one actually occurs. That’s why supervisors and banks spend much effort on improving the usability of these plans in a crisis.

To be most effective, each banking group must have just one plan, and that plan must cover the entire group. It also must contain sufficient information on the relevant entities within the group. Only then can it serve as a tool for further integration. In Europe, we have moved in this direction, also following a recommendation issued by the European Banking Authority in 2017.^[6]

And we should not forget the option to make intragroup financial support agreements part of banks’ recovery plans. This would help to map out how group entities could support each other in a crisis while taking into account local needs and restrictions. The challenge would be to better integrate the assessment of capital and liquidity needs for all the group entities with the safeguards provided for in recovery plans. There should be other ways of providing reassurance to local authorities than trapping capital and liquidity in each Member State.

Recovery plans are great tools, but they need to be applied in a credible manner. And even when they are, some banks may still end up in resolution. So the next step is to ensure that the links between recovery and resolution plans are fleshed out; the steps that take us from recovery to resolution must be clear to both home and host authorities and to the banks themselves. This will maximise the chances of a smooth resolution process.

As for resolution plans, the good news is that, in the banking union, single point of entry is the method of choice. This shows that, at least in good times, there seems to be some trust in the European approach to resolving banks.

This does not make the plans themselves credible, of course. They will only be credible if they are detailed enough to guide actions in a real-life crisis. Authorities should focus on the relevant questions. What are the critical functions of the bank? How exactly is the bank or the group connected to critical market infrastructures? How interconnected are the various business lines? Banks should approach these questions and their planning with a stronger focus on operational details.

Recovery and resolution plans can become key tools for authorities to lay out how they will approach cross-border banks that have got into trouble. The foundations have been laid. Now it is up to banks and supervisors to seize the opportunity and agree on how to tackle a crisis – in a credible, cooperative manner.

Conclusion

Ladies and gentlemen,

The financial crisis dealt a blow to global banking. It became clear that no one was prepared to deal with large cross-border banks in a crisis. Trust vanished, cooperation broke down and national approaches often prevailed.

Since then, much has been done to solidify the foundations of a global approach to banking. In Basel, countries from around the world agreed on common standards on how to regulate banks and handle crises. This is a major achievement when we consider how different the schools of thought are around the world. Crisis management is a case in point: there are very diverse views on what it should achieve and how it should be done – just compare the United States and Europe.

At regional level, however, it should be easier to make progress. And indeed, Europe has taken quite a few positive steps: since the crisis, both the supervision and the resolution of banks have been elevated to the European level. We have hardwired cooperation into the legal framework. But countries are still ring-fencing their national banking sectors, fearing that banks are still national in death. This is blocking the path towards a truly domestic European banking market.

It is common wisdom that trust is the glue that holds the banking sector together – even more so in a cross-border setting. This makes it an issue at both the European and global level, and we must attempt to build more trust, for one by devising credible institutions and binding rules. And also by talking to each other, understanding each other and emphasising common goals. This, by the way, is what distinguishes our situation from the real prisoner’s dilemma: we can communicate and find the best solution for everyone involved. Cooperation is key.

Thank you for your attention.