The banks and the market

Speech by Sabine Lautenschläger, Member of the Executive Board of the ECB and Vice-Chair of the Supervisory Board of the ECB, Lecture at the Florence School of Banking and Finance, Florence, 15 March 2018

Almost ten years ago to the day, the world of finance was shaken by a strong tremor. On 16 March 2008, JP Morgan offered to buy the investment bank Bear Stearns. And it was time. Bear Stearns was in deep trouble: its share price had fallen by two-thirds since the beginning of that year; it was bleeding money; and it was basically shut out of the funding market. At the time, the failure of Bear Sterns was considered a major event. No one knew that the big quake was still to come.

It happened half a year later, on 15 September 2008, when another investment bank failed: Lehman Brothers. It pushed the financial system close to the abyss. Trust evaporated, markets dried up, and around the globe, banks were sent reeling.

All in all, it was the biggest financial crisis in decades. It dragged down the real economy, which fell into what we now call the Great Recession. In the euro area, economic output went down by almost 4.5% in 2009, while unemployment jumped by more than a quarter to 9.6%. Three years later, it reached 12%. In Italy, for instance, output fell by more than 5% in 2009, while unemployment peaked at almost 13% in 2014.

These figures are alarming, but they tell us little about the people who lost their jobs, their homes and their future. Taking all this suffering into account, the crisis turned into a tragedy.

As for its causes, I admit that the story I’ve just told is far too simple. Of course, it was not the failure of a single investment bank that caused all this mayhem. The full story is much longer and much more complex.

However, no one can deny that banks played a crucial role. It became obvious that they can pose a huge risk to financial stability and the economy. But at the same it’s an inescapable truth that we need banks. Or, to be more precise, we need the services they provide.

Why we need banks

Who does an investor turn to when she needs money to start or expand a business, when she needs to hedge risks or wants to acquire another company? Who does a saver turn to when he wants to deposit money? Who does a family turn to when it needs a loan to buy a home? Here in Europe, all of them would most likely turn to a bank. Banks provide a vital service to the economy. Without them, the economy would not work.

All this leads us to a crucial question: how can we ensure that banks reliably serve the economy? Long taken for granted, this has become the holy grail of banking regulation and supervision.

And it has become the ultimate goal of all the relevant reforms that have been undertaken since the crisis. And much has been done. We have revamped the global standards for capital and liquidity – Basel III. Rules for banks are now much stronger than they were before. And the same is true for supervision. Just think of European banking supervision, which was set up in 2014. All this will help to make banks more resilient, so that they can reliably serve the economy.

But there is one thing which neither rules nor supervisors can do. They cannot prevent each and every bank from failing. And in my view, we should not even try to. The possibility of failure is inherent to any kind of business, and it is a core element of well-functioning market economies.

No, it is not the job of supervisors to keep individual banks alive. We, in particular, must not lower our standards to cater for the weakest bank. It would also be wrong to assume that each bank, even the weakest, just needs time to sort itself out. It would be wrong to assume that the local economy just needs to grow for a few more years for the bank to recover. The result would be weak banks that stagger on, zombie-like, barely able to survive in good times. In short: such supervisory forbearance distorts the market and weakens the entire banking system. Supervisors would become part of the problem.

Obviously, we shouldn’t overdo it either. If justified, banks should always have reasonable time to comply with new rules and standards. Think, for instance, of the transitional phases for the Basel framework. But the standards themselves should not be weakened.

All this leads us to an interesting conclusion. Obviously, it was a problem that banks failed during the crisis. But taking a long-term view, I would argue that it was also a problem that many weak and under-capitalised banks that did not have a viable business model did not fail. They did not fail prior to the crisis – even though they probably should have – and they could not fail during the crisis either.

But not letting banks fail creates some huge problems further down the road. Let’s discuss why banks should be able to fail, but often were not able to do so.

Why banks must be able to fail (but often were not allowed to)

Why was it that in the past banks were often not allowed to fail? The answer is simple: the failure of a bank can trigger a financial earthquake.

Banks engage in maturity transformation: they accept short-term deposits and make longer-term loans. Banks depend on public trust. If that trust disappears, they may suffer a bank run. And that can affect the stability of the whole system. This happens because banks are highly interconnected – either directly or indirectly. The failure of one bank may trigger the failure of other banks.

Such contagion can occur by banks being exposed to each other. When a bank defaults on its debt, creditor banks suffer direct financial losses. But contagion can also be more subtle. The failure of a bank may cause investors and depositors to wonder about the health of other banks. Trust then evaporates and funding sources dry up. Even sound banks might find themselves in the midst of a liquidity crisis. This is what happened in the financial crisis.

Either way, a bank’s failure can infect the entire banking system. And in this respect some banks are considered to be riskier than others. I am talking about banks that are big, complex, interconnected, and which cannot easily be replaced. Their failure can disrupt financial stability big time. Such banks are systemically important. If they fail, the financial system can grind to a halt. And not only that: the flow of credit stops, investment and consumption stall, and jobs are destroyed. The social costs can be huge.

To prevent all this from happening, governments have repeatedly bailed out systemically important banks with public money. Even medium-sized banks were bailed out for fear of contagion. But the short-term gain of preserving financial stability comes at a great cost in the long term.

The belief that banks would be bailed out whatever the cost created bad incentives. Banks got used to an implicit and costless government guarantee, which kicked in when things went wrong. Sound and prudent management was often considered less important. Banks could boost their profitability by investing in riskier assets. That way, they could reap higher returns, knowing that the risks would be borne by others, by taxpayers.

And neither shareholders nor investors provided the necessary checks and balances. They lacked incentives to monitor banks’ risks. They often expected outsized returns, knowing that the banks would need to take equally outsized risks to earn these returns. But they didn’t care too much; investors too lived in a world where profits were gained by the few but losses were imposed on the many.

Funding costs of systemic banks were too low, which allowed them to grow even larger, making them “too big to fail”. At the end of the day, banks and their investors had little incentive to act in a sustainable and forward-looking manner.

So banks must be able to fail for all the reasons I just mentioned.

Making banks fail again

When’s the right time?

Against this backdrop, the aim of all the regulatory reforms was not just to make banks more resilient. They also aimed to make banks fail again. But don’t get me wrong here. Failure is the least desirable option, of course. Above all, we need healthy and well-functioning banks.

But still, sometimes banks do fail. And to ensure that they can do so in an orderly manner, a European resolution mechanism has been devised. It defines the tools and the institutional set-up. It forms the second pillar of the banking union alongside European banking supervision.

But again, resolving a bank is just the very last resort. Thus, the new framework becomes applicable long before a bank fails. It requires banks, for instance, to draw up recovery plans. In these plans, the banks set out how they would overcome severe crises using their own resources.

Now, recovery plans are drawn up when the going is good. When times are tough, the new framework offers some more tools, including early intervention measures. Supervisors can use them if they see that a bank is about to get into trouble. The idea is to be intrusive and to make the bank officially aware of the supervisor’s serious concerns. This helps to get the bank back on track.

Early intervention measures can be an important tool to prevent bank failures. In practice, however, it is challenging to use them, for two reasons.

First, because they overlap with some of our standard tools. To give you just one example, asking a bank for a capital plan can be either a standard supervisory tool or an early intervention measure. This overlap is a challenge as it often prevents us from applying early intervention tools. We are bound to use the least intrusive tool, which is always the standard one. We always have to take into account the costs and benefits of our measures.

Assume that the supervisor, as an early intervention measure, asks the bank to draw up a capital plan. The bank might then be required to disclose this request to the markets as it shows that the supervisor is seriously concerned. This might erode trust in the bank and, in the worst case, trigger a bank run. This would bring about the very crisis that the supervisor wanted to prevent in the first place. If an early intervention measure does more harm to the bank than it does good, supervisors might be obliged not to apply the measures.

Instead, they could ask for the same thing, but frame it as a standard tool. In that case, the bank would still have to draw up the capital plan, but would probably not need to disclose the request to the markets. Hence, the standard supervisory tools can be considered as less intrusive than the early intervention tools.

So the overlap between the two tools should be removed. There should be a clearly defined set of standard tools and early intervention measures and a clearly defined escalation process leading from one to the other.

The second reason why it is challenging to use early intervention measures is because those measures are defined in the Bank Recovery and Resolution Directive, or BRRD. And as the BRRD is an EU directive, it has to be transposed into national law. The ECB must thus exercise its powers based on the national transpositions; and these vary from country to country. This creates uncertainty and inconsistency. What we thus need is a harmonised legal basis from which we can apply early intervention measures.

Now, let’s go one step further. Imagine our bank gets into more serious trouble. In this scenario, there is one tool that helps to create a bit of breathing space for the bank and the authorities to sort things out – a moratorium. A moratorium allows the authorities to suspend all the activities of a bank for a short period of time. In a sense, it freezes time.

A moratorium helps to manage a crisis in an orderly manner. It gives the bank and its shareholders time to find a solution. This is particularly useful during a sudden liquidity crisis. However, there is a practical problem. Moratoriums are not part of the European toolbox. In some countries they do exist; in others, they don’t. And in any case, they are not harmonised. We need harmonised, short-term moratoriums in all euro area countries.

If, for instance, a cross-border bank gets into trouble, the supervisor in that bank’s home country can impose a moratorium. But what if the bank has a subsidiary in a country where only a court could impose a moratorium? What if the bank has a subsidiary in a country where a moratorium cannot be imposed at all? This is a serious stumbling block for crisis management at European level. This obstacle has to be removed.

Let’s return to our bank and assume that nothing helps and that the situation goes from bad to worse. At some point, the bank will have to be declared “failing or likely to fail”, to use a bit of jargon. This is mostly done by the supervisors, and it is the most intrusive thing we can do. Declaring a bank “failing or likely to fail” has far-reaching consequences. It is a difficult decision which requires a delicate balancing act.

As long as there is a reasonable chance of the bank recovering, a “failing or likely to fail” decision would put an undue financial burden on the shareholders. But if the decision is put off for too long, the costs of an imminent failure start to mount up. This would hurt creditors and depositors. In extreme cases, it might even become a threat to financial stability. So, the “failing or likely to fail” decision must neither be taken too early, nor too late. And in any case, it must be taken with great care.

There are four general cases which call for a “failing or likely to fail” decision. Some of these cases are quite clear-cut and can be objectively assessed. Others are a bit less clear-cut – particularly when they refer to things that are likely to or might happen in the near future. Thus, there is some room for discretion when determining whether a bank is failing or likely to fail. After all, each bank fails in its own way, so each case is different, and each decision needs to account for these differences. We need to take all facts into account and assess them case by case, and not against a single yardstick. In short, we must be proportionate.

Imagine a bank breaches capital requirements. Then we have to ask how severe that breach is and whether it might just be temporary. If it’s just a marginal breach which could be healed within a reasonable period of time, it might not be proportionate to declare the bank “failing or likely to fail”.

We also need to take into account that a bank depends on the markets. So how do they react? A bank in trouble might quickly lose the trust of the markets. This could lead to a bank run, which requires supervisors to act differently.

So, to sum up: “failing or likely to fail” decisions are not taken automatically but are based on expert judgement. At the same time, they have to be taken case by case and in a proportionate manner. All in all, they are the most difficult decisions a supervisor has to take. And I can assure you that we do not take them lightly.

How to resolve a bank

Imagine that we have declared a bank failing or likely to fail. What happens next? Well, at that point the resolution authority takes over. In Europe, this is the Single Resolution Board, or SRB for short.

The SRB then has to answer two basic questions. First: are there any measures, private or otherwise, that could restore the viability of the bank? If the answer to this question is “no”, the second question is: should the bank be resolved or liquidated?

And the answer to this question depends on a number of things. The objective should always be to ensure that the financial system is not disrupted. If the bank performs critical functions, these need to be continued one way or another. In addition, measures need to be taken to prevent the bank from “infecting” other banks or the markets in general. And finally, public funds and the bank’s insured depositors, for instance, need to be protected. So, it is crucial to conduct what we call a “public interest test”.

If all these goals can be met by liquidating the bank, then that’s what will happen under national insolvency law. And this is something of a weak spot, because these laws still differ significantly across the euro area. A minimum level of harmonisation is thus called for. In particular, the triggers for “failing or likely to fail” decisions and for liquidations need to be aligned. If they are not, banks that are deemed failing or likely to fail according to European rules might not be deemed as such under national insolvency laws. This is obviously a huge stumbling block to the harmonised treatment of bank failures.

If liquidating the bank has the potential to disrupt the financial system and therefore the public interest test is positive, the bank will be resolved at the European level instead. The SRB has various tools at its disposal for this and they are laid out in the relevant legislation. All these tools ensure that a bank can fail in an orderly manner without disrupting the financial system. And this is the overarching goal.

Who foots the bill?

Ladies and gentlemen, I have talked about the decision to declare that a bank is failing or likely to fail. And I have talked about what happens after such a decision is taken. The next thing I will talk about is money. Bank failures are costly. So who foots the bill?

“Not the taxpayers” is the first and most important answer I can give you. After all, the main objective of the new framework is to spare taxpayers. Instead, it should be the owners and creditors who bear the losses. I have already explained the reasoning behind this. But how does it work in practice?

The first thing to mention is that banks now have to hold more equity than they did before the crisis. This means that each bank has a larger buffer to absorb losses. And these losses are borne by those who provided the equity, the shareholders. So, higher equity means that the owners of a bank bear a larger share of any losses incurred. At the same time, failure becomes less likely.

And if the bank does fail and needs to be resolved, there is the bail-in. The bail-in is one of the SRB’s resolution tools, and it ensures that the burden of failure is indeed borne by the owners and creditors of a bank. The bail-in establishes what you might call a hierarchy of loss-bearers. First in line are shareholders, followed by holders of subordinated debt, followed by other creditors.

The question is, of course: how much can be bailed in? And the answer is: everything! Well, in theory. In practice, there is a safeguard: the Single Resolution Fund, or SRF. The SRF is financed by banks from all 19 euro area countries. It was set up in 2016, and by 2023, it should amount to about €55 billion.

In a resolution, the SRF can contribute up to 5% of the bank’s balance sheet. But it can only do so after shareholders and creditors have borne losses amounting to at least 8% of the bank’s total balance sheet. So, shareholders and creditors have to bear a large part of the losses first, and then the banking industry as a whole can step in through the SRF.

The basic idea of bail-in is quite clear. But, as is so often the case, the devil is in the detail. However, I will spare you that particular discussion and turn to three more general questions related to bail-ins.

A bail-in ensures that losses are borne by the owners and creditors of a bank, by those who took the risk and earned the return. So the first question is: what if they were not aware – or were not made aware – of the risks they were taking? For it is sometimes the case that retail investors are among those who are bailed in. These are not professional investors but ordinary people who were trying to build up funds for retirement, for instance. For such people, a bail-in is not just a financial loss; it is a personal tragedy.

We must avoid such tragedies. But we must do so in a way which is in line with the basic idea of bail-in. And this is as much a question of consumer protection as of financial education. In terms of consumer protection, we need disclosure requirements which are clear and easy to understand so that all investors are aware of the risks they are taking. In terms of financial education, we must ensure that all investors have a basic understanding of finance, and in particular of the relationship between risk and return.In both areas, I see a lot of room for improvement in Europe.

And there’s more we can do. We could set a minimum threshold of €100,000 for investing in subordinated debt instruments. This too would help to steer retail investors away from risky financial endeavours. Here, we would need a European approach, of course. If each country were to go its own way, the market for the relevant financial instruments would become too fragmented.

The ECB welcomes current proposals which foresee the introduction of non-preferred senior debt instruments and a general depositor preference rule. This would further facilitate bail-ins.

The second question is: what happens if the Single Resolution Fund runs out of money? What if there’s a major crisis, and the €55 billion is not enough to cover the losses? Well, in such a situation we would need a common European backstop for the SRF. And we need it quickly, and in any case before 2023, which is when the SRF will reach its target volume. European leaders committed to this when setting up the SRF, and they should stick to their commitment. The backstop could take the form of a direct credit line from the European Stability Mechanism to the SRF. The important thing is that it can provide both solvency and liquidity support.

And now to the third and final question: has public support for failing banks really become a thing of the past? “Almost” is the answer to this question. And the rules are now such that any public support for a bank in fact triggers a “failing or likely to fail” decision.

There are a few exceptions, however. One of these is known as precautionary recapitalisation. It allows governments to recapitalise banks using public funds. But the scope for this kind of state aid is very narrow.

  • First, it is only available to solvent banks.
  • Second, it can only be used to cover capital shortfalls which were identified in the adverse scenario of a stress test.
  • And third, it can only be of a precautionary and temporary nature. In other words, it must not be used to cover losses which the bank has already incurred or is likely to incur.

I will highlight just one issue here which relates to the role of the ECB. With regard to the first condition, it is up to the ECB to confirm whether the bank in question is solvent. It has some discretion here, as the European legal framework does not offer a specific definition of solvency. So it is crucial that we have a consistent definition of solvency – a definition that not just covers a point in time but takes a forward-looking perspective. Even if a bank is solvent today, it might not be so in the near future.

In the end, it is the European Commission that grants or denies permission for precautionary recapitalisation, as state aid rules apply. Among other things, this entails burden-sharing. Shareholders, as well as holders of hybrid and subordinated debt, must contribute to reducing the capital shortfall. This ensures that the burden of failure is still borne by those who took the risk and reaped the return.

Spotting problems and keeping them from mounting up

Given the limited time we have, I have given you a very brief overview of what resolving a bank entails. It is a gruesome and complex event, and one which is best avoided, but not at all costs.

To avoid failure, banks and supervisors must spot problems early on and address them quickly.

Since the crisis, stress tests have become an important tool for helping us to do this. Stress tests basically assess how banks would fare in various hypothetical scenarios involving different future paths for the economy and the financial markets.

Usually, there is a baseline scenario, where things develop as expected. And then there is an adverse scenario, where they don’t: inter alia, GDP, unemployment and property prices are stressed two years ahead in this scenario. So you can see that the key idea of a stress test is to assess how banks would cope if the going gets tough; it’s all purely hypothetical, of course.

Stress tests allow us to make a consistent assessment. Each bank has to apply the same scenarios and calculate their impact in the same way. This means that the results can be compared across banks.

And stress tests provide a glimpse of the bank’s risk management in action. How quickly can it process the scenarios and assess their impact on [the state of] the bank? This is important information about the quality of governance. But the main focus of stress tests is on capital. Do the banks have enough capital to absorb any losses in either scenario? And the crucial question that follows is, of course: what if they don’t?

Well, let’s take a brief look at the upcoming stress test. This year, the European Banking Authority, EBA for short, will conduct a stress test involving the largest banks in the EU. This test includes 37 banks from the euro area which are directly supervised by the ECB. To obtain a fuller picture, we will, in parallel, stress-test most of the remaining large banks in the euro area, using the same scenarios and methods as the EBA.

In respect of capital shortfalls, the stress test is not a pass-or-fail exercise. The banks do not have to meet certain thresholds. Instead, we supervisors will decide on a case-by-case basis whether banks have a capital shortfall, taking into account their individual risk profiles and their individual capital needs.

At the end of the day, stress tests may deal with hypothetical scenarios, but their results can have very real consequences for the banks. They inform the work of supervisors, and they inform market expectations when the results are published, as they will be in the case of the EBA’s stress test. They are also a moment of truth for the banks.

More generally, stress tests help supervisors and banks to spot and remove weak points in banks’ resilience early on.

If this is not done, problems can quickly mount up. A current example are bad loans, or to use a more technical term, non-performing loans, NPLs for short. Banks in the euro area carry a total of €760 billion in such loans on their balance sheets. And this is an issue.

Having large stocks of NPLs on their balance sheets prevents banks from doing their job. They have a hard time financing our economy. ECB internal research has shown that banks with high stocks of NPLs indeed lend less than others.

It is thus important that banks resolve their NPLs quickly and decisively. And this is not just for the sake of the economy and the greater good. Failing to deal with NPLs also hurts the banks themselves. High levels of NPLs require constant attention. And this takes time and energy. So getting rid of NPLs would enable managers and staff to use their time more productively. They could improve business models, shore up profits and make their bank more resilient.

So, failing to deal with NPLs hurts the economy and of course it hurts the banks. That’s why we consider NPLs a top priority. Getting rid of NPLs by restructuring, working out or selling them requires a joint effort by banks, politicians and supervisors. Banks have to put in place the right strategies and execute them. Politicians have to improve the judicial and legal systems to allow a swift resolution of NPLs.

And we supervisors? Well, last year we published guidance on how to deal with NPLs. On the basis of that guidance, we scrutinise the banks’ strategies and provide feedback. So we are keeping a close eye on how banks are addressing the problem.

And today, we have published an addendum to our guidance. This addendum will help to ensure that NPLs remain a problem of the past once they have been dealt with. To that end, it lays out how we expect banks to provision for new NPLs. More specifically, it refers to exposures that are newly classified as non-performing as of 1 April 2018. The addendum is not a binding document, but it specifies our expectations and will thus serve as a starting point for our dialogue with each bank.

Conclusion

Ladies and gentlemen,

Rest assured that I have not exhausted the topic of my speech. But I may well have exhausted you. So I will conclude quickly and painlessly.

First, banks are important for the economy, particularly here in Europe. That said, it might be good to add a bit more capital market to the mix; but that is a topic for another day.

Second, banks must be able to fail. If they are not, the economy will suffer in the long run.

But if banks fail, they must do so in an orderly manner. If not, financial stability will suffer in the short run. The new European resolution framework helps to make orderly failure possible.

Third, failure is the least desirable outcome, of course. We need banks which are profitable and resilient in the first place. Stronger regulation and supervision can do their bit. In the end, however, it is up to the banks themselves to ensure they lead long and healthy lives.

And with these three messages this long speech comes to an end.

Thank you for your attention.

Media contacts