Benoît Cœuré: Interview in Risk Magazine

Interview with Mr Benoît Cœuré, Member of the Executive Board of the European Central Bank and Chairman of the Committee on Payments and Market Infrastructures (CPMI), in Risk Magazine, conducted by Ms Cécile Sourbes on 10 May and published on 19 May 2016.

The views expressed in this speech are those of the speaker and not the view of the BIS.

Central bank speech  | 
23 May 2016
PDF version
 |  5 pages

What was the outcome of the questionnaires that the CPMI and the International Organization of Securities Commissions (Iosco) sent to central counterparties (CCPs) in 2015?

The work done by CPMI-Iosco is a response to the request made by the G20 ministers and governors to ensure that CCPs are safe and resilient. The work done on risk in CCPs has been set as a priority this year by the Chinese presidency of the G20. Risk in CCPs is an important piece, and maybe even the main missing piece, of the post-Pittsburgh regulatory agenda.

A great deal of work has been done to make the derivatives market safer, make central clearing mandatory and improve the resilience of infrastructures - in particular through the implementation of the CPMI-Iosco Principles for financial market infrastructures (PFMI). Now, it's time to wrap it up, take stock of what has actually been done and consider whether this is enough or whether we need to do more. So it's an important milestone.

To answer your question, yes, CPMI-Iosco sent a series of questionnaires to several CCPs last year which covered both recovery planning and risk management practices more generally. It was a way to assess their compliance with the PFMI and the recovery guidance released in 2014 by CPMI-Iosco. We have looked at the responses in depth and this has given us a good insight into how CCPs have implemented the PFMI. We are now taking stock of this material and we are also working with the Financial Stability Board (FSB). The outcome of this work will appear in two "twin" reports that will be published in early summer, presumably in July. Those reports will come right in time for the G20 summit in Hangzhou on 4 and 5 September.

Are there any areas that require further guidance?

The first report will review the implementation of the PFMI across 10 derivatives CCPs. We are very positive about the work that has been conducted by CCPs, but we've identified shortcomings when it comes to addressing certain topics, including financial resources and liquidity resources. We have also found a number of shortcomings in terms of recovery planning.

This is not about rewriting the PFMI - there is unanimity that the PFMI are the right approach as they stand. It's about making sure that the PFMI are implemented fully and consistently across all jurisdictions. And the outcome of this work will show there is still more to be done.

This is why we will publish a second report, which will provide additional guidance to support the implementation of both the PFMI and the recovery guidelines released in 2014. There will be a public consultation process on this second report in order to gather input from the industry, and the final guidance will be published at a later stage, either late this year or in early 2017.

In terms of the PFMIs, what are the areas where further guidance is needed?

Governance, financial resources and stress testing are all included. We will also have further guidance on margining and, to some extent, on collateral management.

The report will be quite broad. It will send a positive message that the PFMI are being taken seriously by the clearing community, but it will also highlight a number of gaps that must be carefully examined.

The CPMI-Iosco guidance on the recovery of financial market infrastructures is relatively recent and only dates back to 2014. Why is there a need for additional guidance at this point?

Recovery planning is a work in progress and it's a completely new area. It has been implemented differently by CCPs in different jurisdictions, for obvious reasons. CCPs have different business models, they face different risks, their membership and ownership are different and, even more importantly, they operate under different regulatory and legal frameworks. So CCPs move at different paces depending on the difficulties they face.

Our report is more about making sure that the existing guidance is properly and consistently implemented across markets and jurisdictions, based on what is actually being done by CCPs, than about issuing additional guidance, since it only dates from 2014.

I must clarify that this report will cover recovery only. It's not about resolution. Resolution is being discussed by the FSB and they work under a different timeline. We do work in close partnership with them. Resolution comes after recovery, so it makes sense for CPMI-Iosco to address recovery and then, later this year, the FSB will come up with its own guidance on resolution.

Some questions remain when it comes to CCP resolution. Will the FSB be prescribing any specific rules on that?

It's not my role to comment on the work the FSB is conducting, but let me stress the strong interaction between the recovery and resolution discussions. One of the first issues they will address is the question of who is in charge and what the resolution point is. When is the right time to go into resolution?

Good risk management starts in the CCP, even before the recovery point, of course, and that's why the implementation of the CPMI-Iosco guidance is so important. And the PFMI are stringent enough, in particular through the Cover 1 and Cover 2 requirements, to make sure that even recovery is considered a tail event and that resolution is even less likely to materialise - let's call it a "tail-of-tail" event. But this is obviously not an excuse for not having the right arrangements because resolution can still happen and because the allocation of costs in resolution shapes the incentives in recovery.

As far as the resolution point is concerned, an authority should in my view be able to decide that a CCP needs to go into resolution even before the recovery waterfall has been exhausted, for two reasons. First, there might be a public interest in a CCP going into resolution rather than waiting for the recovery instruments to be exhausted. The resolution authority speaks for the public, starting with the indirect members and investors, and society more broadly if there are possible systemic consequences. Second, authorities don't want to step in when financial resources are already exhausted.

Another key interaction between the daily management of the CCP, recovery and resolution has to do with cooperation. As you know, the PFMI include a responsibility for cooperative oversight. There are discussions within the FSB about setting up cross-border crisis management groups for CCPs, where relevant authorities will cooperate if resolution becomes necessary.

Of course, the two discussions - between CCP supervisors and overseers as well as between relevant authorities if resolution becomes necessary - need to be consistent because we cannot afford to have two different groups of people taking different decisions. This starts with having a common approach to deciding on which CCPs are systemic in more than one jurisdiction. We are doing some work within the CPMI-Iosco, at the request of the FSB, to agree on a common methodology to define what makes a CCP systemically important.

What kind of criteria are you looking at?

It's an ongoing discussion within CPMI-Iosco and we'll then report the outcome of our work to the FSB, prospectively also around July. The objective is to look at the potential cross-border systemic risk implications of CCPs throughout their potential lifecycle, although the measures to address these implications may be different under business as usual, emergency and resolution conditions.

One recovery tool that is highly controversial is the use of initial margin. Market participants and some regulators believe that haircutting initial margin could create the wrong incentives, since member firms might run for the exit to protect their initial margin, making the task of rescuing a CCP even more difficult. Do you understand those concerns?

I fully appreciate the concerns. The whole point of the recovery and resolution toolkit is to identify a set of instruments that are effective in meeting the potential financial shortfall, but that also set the right incentives for proper risk management of the CCPs and for the clearing members.

Setting the right incentives for central clearing is extremely important. The G20 decided in 2009 to make central clearing mandatory. This has been extensively discussed by the regulatory community and our conclusion is not that the G20's decision was wrong. It is correct because mandatory clearing reduces the overall level of risk in the system, with ancillary benefits such as more efficient use of high-quality collateral. But, of course, it also pools risks in CCPs, and that is why the work on risk in CCPs is so important, and has to be conducted in a way that sets the right incentives. This is true for the CCP rulebooks but also for the banking regulation. That's why we are in close contact with the Basel Committee on Banking Supervision about the parameters of the leverage ratio, because we want to be sure that there are no disincentives for mandatory central clearing. We are now at a point where we should look across different elements of regulation and make sure we are consistent.

On your question, there is a broad consensus at the global level, as well as at the European level, on the toolbox that can be used - cash calls, variation margin gain haircutting, partial tear-ups and, most controversial of all, initial margin haircutting.

It's true that initial margin haircutting is problematic in several respects. It can make the task of rescuing CCPs even more difficult and it also raises legal issues in some jurisdictions which require the initial margin to be insolvency remote. That's probably what makes initial margin haircutting the least desirable instrument, but since we are still at an early stage in the discussions, I don't want to exclude it completely. I'm just saying that this tool would have to be used as a last resort, in a capped way, and for events that are beyond the "extreme but plausible" market conditions that we consider for stress testing.

Some regulators say that, regardless of the problems with initial margin haircutting, this tool should still be made available to CCPs in case they need to gain quick access to liquidity to replenish their default waterfall. Do you have sympathy for this argument?

I don't want to prejudge the outcome of the discussion. There are a number of shortcomings that we still need to investigate more deeply, especially around the legal issue of bankruptcy remoteness and the kind of uneven playing field that we will create if we have an instrument that has a different legal treatment in different jurisdictions. One major concern for us is to be internationally consistent. So if an instrument happens to create an uneven playing field across jurisdictions, it's probably not the right instrument.

Stress testing for CCPs is also high on the regulatory agenda. The CPMI is conducting some work on this. How is it progressing?

The CPMI has already expressed public support for exploring a framework for regulatory stress testing, and this is part of the CCP work plan that was released in September 2015. This work plan is a joint endeavour by different standard-setting committees, including Iosco, the FSB, and the Basel Committee on Banking Supervision. A framework for consistent and comparable stress testing could serve two key objectives: it could help authorities better understand the macro-prudential risks that could materialise, and it could help them assess the financial resources of individual CCPs.

On the first aspect, I believe that supervisory stress tests will bring maximal value to the stakeholder community - both the supervisors and the CCPs themselves, as well as their clients - if they are designed in a way that accounts for the inter-connectedness and the potential spill-over of risk. It's obvious that risk can spread from one CCP to another, from CCPs to the clearing members and then bounce back to the CCPs.

Making sense of these interconnections is a big task. Some would say that the cliff is just too high and would tend to give up. I would say it's a big and ambitious project, so the earlier we start, the better.

What is the timeframe for this project?

CPMI-Iosco are discussing it. We will probably come up with some initial thoughts that we will present to G20 ministers and governors this summer. And I would very much like them to take the project on board and make it a priority of the financial regulatory work stream.

We need to be realistic about it, though. It's only starting and it will extend beyond the Chinese presidency of the G20. This means the German presidency - which will run through 2017 - will need to take up this project. But I think there is a lot of value in trying to build a "macro" approach to supervisory stress testing.

One reason it's so important is because it's one major component of what is increasingly being called a macro-prudential approach to CCPs. CCPs are, essentially, systemic actors. We made them systemic for good reasons, because they help reduce the level of risk in the system. But still, it creates new forms of inter-dependency and a potential spill-over of risk in the system.

What is the macro-prudential approach to CCPs?

The macro-prudential approach to CCPs potentially includes different dimensions.

It starts with good risk management in CCPs under the PFMI. It then includes a better understanding of the inter-dependencies - on which the FSB is currently working - and a better understanding of the CCP practices that can create pro-cyclicality in the system, e.g. margining practices.

Margining is actually a good example of that last point since in some situations there can be a contradiction between the stability of the CCPs and financial stability at large. Imagine a situation where a CCP facing a particular type of risk needs to substantially increase its margin. This is clearly in the interest of a CCP's good risk management, but it may also increase risk in the wider system. The key is to avoid pro-cyclicality.

It then extends to this notion of supervisory stress testing with a macro dimension. Developing a framework for supervisory stress testing should therefore be a priority. It will take time, but it will go a long way to support a better understanding of the macro-prudential risks of CCPs.

Are other supervisors keen to embrace this idea?

I think they all agree that the project makes sense and is useful. They also agree on the fact that supervisory stress testing should be consistent internationally, so its methodology needs to be discussed at the global level. I'm not saying that we should have a big coordinated supervisory stress test at the global level. That would be an incredibly complex project. The implementation needs to be decided by supervisors themselves but there is a clear understanding and agreement that we should have a common concept.

On that note, I would like to welcome the first EU-wide stress test conducted by the European Securities and Markets Authority (ESMA).

The outcome of the EU stress test showed that the daily stress testing CCPs were conducting on themselves was not as tough as the minimum shocks set out by ESMA. Is this surprising or even worrying?

When it comes to compliance with the regulatory requirements, this is the kind of thing we have been looking into within CPMI-Iosco and it's what the first of the "twin" reports I mentioned earlier will be about. The experience shows that there are some shortcomings with regards to regulatory expectations. There has been room for interpretation of the PFMI, which in some cases may create an uneven playing field. That's why we need additional guidance. This additional guidance is not about raising the bar. It's about making it clear where the bar is.

Was this EU stress test urgently needed, in your view?

It's not a question of urgency. It is about starting to develop an instrument that, in my view, will become a normal supervisory instrument, as is already the case in banking supervision.

Stress tests on banks used to be big, extraordinary events, but are increasingly becoming the bread and butter of banking supervision. It started in the United States with the Comprehensive Capital Analysis and Review (CCAR), and Europe is following a similar route.

We are not yet at this stage when it comes to CCPs, and the whole concept is still in its infancy. But what ESMA has done is very important in terms of progressing along the learning curve. I also see a connection between supervisory stress testing and the discussions on financial buffers in recovery and resolution. The big focus for the supervisory community right now is to know whether CCPs need more financial resources in recovery and resolution. That's one big question mark. But I think this question is asked in too-general terms because it very much depends on each CCP. They have different business models, face different risks and evolve in different environments.

So I personally wish we could move gradually to an approach that is closer to the one we have in banking supervision, if we use stress testing to inform supervisory requirements. I would be very much in favour of the concept of a "pillar II" for CCPs. We would have pillar I requirements that would be standardised across CCPs, and we would have a pillar II, which would be decided by the competent supervisor and informed by the outcome of the stress test. To me, that's the way forward. It will take some time to take shape and, in the meantime, we should restrain ourselves from approaches to financial resources of CCPs that are too simplistic and standardised.