Sabine Lautenschläger: The European banking sector - the big challenges
Speech by Ms Sabine Lautenschläger, Member of the Executive Board of the European Central Bank and Vice-Chair of the Supervisory Board of the European Central Bank, at the IIF European Chief Risk Officer Forum, Frankfurt am Main, 20 November 2017.
Have you ever heard of Metéora? Metéora is a group of rock pillars in Greece at the very top of which several monasteries were built centuries ago. As you can imagine, it is quite hard to reach these monasteries. In fact, until the 1920s or so, the only way to get there was by sitting in a basket which the monks then pulled up.
Now the story goes that, one day, a tourist got very nervous half-way up the cliff. He noticed that the rope attached to the basket he was in looked quite old and rather frayed. So he asked the monk who was accompanying him if they ever replaced the rope. "Of course we do, ..." the monk replied, "... whenever it breaks."
I don't have to tell you that this is not how risks should be managed. It is our job, the job of risk managers and supervisors, to be a tad more forward-looking. We have to ensure that the rope is replaced before it breaks. And we have to ensure that proper safety measures are in place should the rope break unexpectedly.
But in reality, it's a bit more complex, of course.
With regard to banks it's not just a rope that might break. There are many different risks that have to be well managed to ensure the well-being of banks. Some risks are quite likely to materialise, others less so. We all have a list of risks in our minds. And I am sure that these lists look quite similar. They might include interest rates that could be too low for too long, a sudden reversal of risk premia, geopolitical uncertainties, non-performing loans and cyber risks.
Banks must monitor and manage all these risks. However, we don't need to discuss all of them today. So, in my speech I will focus on two of them: non-performing loans and Brexit. In addition to that, I will discuss a third topic which is closely connected to risk: capital buffers, and how to calculate them using internal models.
Non-performing loans - the need to act
In a sense, non-performing loans, or NPLs for short, are risks that have already materialised. Things have already gone wrong. Some creditors cannot service and pay back their loans. The resulting NPLs pose a risk - for the affected banks and for the economy.
Only sound and healthy banks with clean balance sheets can reliably finance economic growth. But balance sheets that are weighed down by NPLs are a huge burden for banks. NPLs are a drag on profits; they might also weaken trust in the banks; and they keep banks from lending to the economy.
That's why it is so important to resolve the problem. And we are on the way to doing so. Since 2015, the ratio of NPLs in the euro area has gone down from around 7.5% to around 5.5%. In absolute terms, we are talking about a decrease of about €200 billion. So, things are moving forward. However, in some parts of the banking sector, the ratio is still too high.
I realise, of course, that eliminating NPLs is easier said than done. First, it is difficult to sell them - there is hardly any market for them. Second, it is sometimes difficult to resolve them in court. While foreclosure procedures in some countries last less than a year, in other countries they can take six years or even longer.
These are hurdles but they must not serve as excuses. On the contrary, if a bank is not able to recover value from the collateral for a non-performing loan within an adequate time frame, it is questionable whether this collateral should be acknowledged as risk coverage.
So, the banks need to do more. They must devise ambitious, realistic and credible plans to get rid of their non-performing loans. And now is the time to do it. The economy in the euro area has been growing for more than four years now; we are experiencing a broad-based, solid recovery. This is an ideal backdrop for banks to tackle their NPLs.
We published qualitative guidance to banks on that very subject earlier this year. It sets out the ways in which banks are encouraged to deal with non-performing loans. On that basis, we have scrutinised the reduction plans devised by the banks. And we have already started to provide feedback.
We have been assessing the banks' strategies with three questions in mind: how ambitious are they? How realistic are they? And is the governance in order? Regarding their ambition, unsurprisingly the reduction targets set by banks vary greatly, from 3% to 83% over three years. As you can no doubt imagine, we think that some banks' targets are not ambitious enough.
To assess how realistic the banks' strategies are, we look at the timelines, for instance. In general, we don't like strategies that simply defer reductions to later years. We also look at potential tools to be used, such as cures, sales or write-offs. Here, we expect the strategies to be well diversified, not just focusing on a single tool.
And there is another thing we need to take into account. How realistic a strategy is also depends on provisioning. Does the foreseen level of provisioning support the strategy? Could there be a disconnect between portfolios which are meant to be sold and the associated projected provisioning? Given current market prices, these two elements might not always match.
That said, we have observed a trend of increasing coverage.
But dealing with NPLs also requires us to look to the future. Once the balance sheets have been cleaned up, they should stay like that.
That is why we have published a draft addendum to our guidance, which is currently the subject of a public consultation. This draft provides transparency for the banks as we lay out what we expect from them in general in terms of prudential provisioning for future NPLs.
We are aiming for a structured dialogue. The starting point for this dialogue is defined by our general expectations and whether the bank concerned deviates from them. When the bank provides reasoning and evidence to back up this deviation, there is no need to consider supervisory action.
But there might be cases where the reasoning and evidence is not convincing at all, and where we might still regard the provisions as insufficiently prudent. In such cases, we will consider taking supervisory action under Pillar 2.
And in that case, we would have a broad range of measures at our disposal. We might, for instance, require the banks to report, in detail, on the evolution of their NPLs; to collect different information; to re-evaluate the collateral taken in to cover the loans; or to apply specific adjustments, such as deductions and filters, to the calculation of own funds.
Such actions are nothing new: they are part and parcel of a supervisor's daily work.
We have done a lot in respect of non-performing loans. But that's our job. It is our job to address vulnerabilities in the banking sector. And who would deny that non-performing loans are among the biggest vulnerabilities? We have to deal with them. And more than that, our actions complement the NPL action plan agreed on by European finance ministers in the summer.
After all, it's not just up to banks and supervisors to deal with NPLs. Governments can and should help too, with a view to repossessing the collateral. In some countries there are no specialised courts or judges to deal with insolvencies, for instance. Also, there are often no time limits for insolvency procedures. All these factors and others slow down the resolution of NPLs, and they could all be addressed by national governments.
With a level playing field in mind, we should aim for a situation where court resolution of NPLs works equally smoothly in all countries, and where out-of-court settlements are common tools.
Ladies and gentlemen, non-performing loans have slowly built up over the years. So there was hardly an element of surprise. Other risks are different. They linger out of sight for a while, and then they emerge suddenly and unexpectedly. Brexit is the most recent example.
Brexit - the clock is ticking
As regards Brexit, the main issue for banks has been clear from day one: market access. Once the United Kingdom has left the EU, banks operating from the UK might lose the EU passport. And with it, they might lose access to the European market.
And by the way, that is also an issue for European banks that do business in the UK. They keep the EU passport, of course, but might lose access to the UK market.
So, once the UK has left, banks that operate from there might have to seek another route into the European market. And the most obvious one is to become part of the European market, that is, to relocate. Likewise, European banks that operate in the UK might have to find a way to maintain market access.
And that is a pressing issue, as the deadline is just over 16 months away. So banks should be taking expeditious steps now. But are they?
Well, the picture is a bit mixed. First, there are those banks which access the European market from the UK - the "incoming" banks, from our point of view. They are making progress in preparing for the post-Brexit world, although details and decisions are often still missing. That said, there are still some banks, mostly smaller ones, which seem to be delaying their final decision on relocating.
Second, there are those banks which access the UK market from the EU - the "outgoing" banks, from our perspective. Most of these banks have also made plans, which is good. Still, some of these plans remain a bit too high-level, in our view. So, more work needs to be done, even if the circumstances might be challenging.
My message to all affected banks is this: don't procrastinate. No one will wait for you. When Brexit happens you will either be prepared, or not. I advise you to be prepared.
And there are a lot of things that need to be considered. Settlement finality, for instance, might be affected by Brexit, because the protection of the EU Settlement Finality Directive regarding the unwinding of transfer orders will no longer apply across the Channel.
Another issue is the continuity of contracts, that is, the ability of market participants to continue servicing existing contracts, in particular derivatives contracts, without appropriate permissions in place. Banks should therefore include this topic in their contingency plans.
And finally, banks should assess what Brexit means for their recovery plans. After all, it could raise barriers and make certain recovery options less feasible.
But Brexit is not just something the banks have to think about. It also raises a lot of questions for us supervisors. And right from the start, we have worked hard to answer all these questions and map out our approach to Brexit.
We have to think about cross-border banking groups, for instance. In the wake of Brexit, new banks will be set up in the euro area. And they will most certainly be part of banking groups which are headquartered outside the EU, in a third country.
This raises the question of how autonomous the new euro area entities will be. Will they become well-established banks? Or will they end up as shell companies, which are overly reliant on other group entities from outside the EU? And to be very clear: that's not what we want to see.
We expect banks to manage some of their risks locally. They need local staff, and they need local infrastructure. In respect of market risk, banks must be able, in the medium to long term, to trade locally on a permanent basis, and they must have local risk committees. Likewise, they will need to trade and hedge risks not just with other group entities but with diversified counterparties.
Against that backdrop, we are currently working on a booking model assessment framework, which will set out what we expect from banks. Naturally, our expectations will follow the principle of proportionality. In other words, large banks that are highly interconnected and conduct complex capital market operations will have to meet higher expectations.
But it's not just booking models that play a role here. I already mentioned the need for local staff. It seems, however, that some incoming banks have something in mind which is referred to as "dual-hatting". It means that staff members carry out functions in more than one entity of the group. Here, we have some concerns, in particular if dual-hatting includes important functions, extending to that of the CEO. Wearing two hats might cause conflicts of interest and limit the time that can be spent on each. This is even more of an issue when the relevant staff members are physically located outside the euro area. In our view, certain key roles should not be part of such dual-hatting arrangements.
In other words, euro area entities must be sufficiently independent from group entities that are located outside the EU. We will not accept shell companies; that's for sure. We do not see them as banks.
And we won't accept more inventive set-ups either. In order to avoid having to move staff and assets to the Continent, some banks seem to be considering the idea of establishing a subsidiary in the euro area, which would then set up a branch in the United Kingdom. This branch would use the subsidiary's EU passport to enter the European market from the UK. We have serious misgivings about this. First, the primary reason for establishing a branch should be to serve the market of the country where the branch is based. Then, there are legal issues. Can a third-country branch use its parent company's EU passport to access the European market from that third country? That cannot be in the spirit of European law. And the subsidiary might just serve as a shell company. As I said, we do view these kinds of design critically.
Internal models - setting capital buffers
Ladies and gentlemen, I have described a small part of our risk landscape. As you know, the current overall risk landscape is much wider than that. To deal with risks, banks need to hold adequate capital buffers.
These buffers serve as a universal protection against potential losses. A bank's risk-weighted assets should be the basis on which the capital requirements are calculated.
It is the banks who set the risk weights, using their own internal models. So the models have to work properly. And there has to be a level playing field regarding the supervisory approval of these models. You know just how much the processes for approving internal models differed across euro area countries in the past.
That's why we are now assessing all major internal models used by those banks directly supervised by the ECB. This targeted review of internal models, TRIM for short, is a huge project; we are planning to do about 200 on-site investigations. It is probably the biggest project since the health check for banks which we conducted in 2014.
What do we want to achieve with TRIM? Well, there are four goals.
First, we want to ensure that the models used by banks comply with regulatory standards.
Second, we want to harmonise the way in which we treat internal models from the supervisory perspective. There should be no unwarranted or undue heterogeneity in capital requirements for banks with similar risk profiles, just because they are headquartered in different jurisdictions within the euro area.
Third, we want to ensure that the results of internal models are driven by risk and by risk only. They should not be materially driven by modelling choices.
And fourth, as a result of all this, we want to confirm that capital requirements are indeed adequate.
So, once we have finalised TRIM, and once the banks have addressed all potential findings, capital requirements will be calculated in a more consistent way across institutions.
That said, it will take us a while to finalise TRIM. As I mentioned, it is a huge project. Still, we are on track - thanks to the good cooperation with banks and the dedication of our teams.
We started our actual reviews of internal models in April of this year. By mid-2018, we will have conducted 120 on-site inspections. We have already launched more than two-thirds of these. As for credit risk, our focus lies on models that cover exposures to retail customers or small and medium-sized enterprises. In addition, we also focus on market risk and counterparty credit risk models.
But we will also look beyond that. From mid-2018 into 2019, we will assess models that cover low-default portfolios. These portfolios include specialised lending, as well as lending to institutions and to medium-sized as well as large corporates. We will provide more detailed information on this part of TRIM in due course.
We have already started to follow up on the cases of non-compliance that we have identified so far in the general set-up of internal models. In particular, we have found issues in model governance that need to be remedied immediately. These are now being addressed in a thorough and consistent manner.
And we are working on an update of our guide for TRIM.
This update is based on the comments we received on the first version of the guide, which was sent to the banks and posted on our website in February.
It is also based on insights from our on-site missions and a horizontal analysis of their outcomes. We will, of course, seek feedback from banks on our updated guide. The part of the guide that is dedicated to general aspects of model governance, for instance, will soon be published for consultation.
Ladies and gentlemen,
Let us return to Metéora. In the 1980s, one of the monasteries served as a location for a James Bond movie, which inevitably involved some stunts. Roger Moore later confessed that he resorted to moderate drinking in order to overcome his fear of heights. Probably not the best approach from the viewpoint of a risk manager.
But the really dangerous scenes were done by a professional stuntman, Rick Sylvester, who, at one point, had to jump off a high cliff. But did he resort to moderate drinking to calm his nerves? No. Did he just jump off the cliff like that? No. Instead, the stunt team very carefully prepared the jump, putting in place all sorts of safety measures, mitigating all the risks. Mr Sylvester jumped, the stunt went well and the movie became a box-office success, with prudent risk management playing a hidden but vital part.
Thank you for your attention.