Andrew Gracie: LAC and MREL - from design to implementation

Speech by Mr Andrew Gracie, Executive Director of Resolution of the Bank of England, at the BBA loss absorbing capacity forum, London, 23 July 2015.

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

Thanks for the opportunity to speak.

In many ways, resolution is the last major piece of the financial reform agenda and ensuring banks have sufficient loss-absorbing capacity is central to that.

Rules in this area remain to be finalised, but the components are coming into place. The binding requirement for European resolution authorities is, from early next year, to begin setting a minimum requirement of own funds and eligible liabilities (MREL) for all EU banks in line with the Bank Recovery and Resolution Directive (BRRD) requirements and European Banking Authority (EBA) Regulatory Technical Standards (RTS). You will have seen that the RTS have recently been published in an "EBA final" version. This will be a binding EU regulation once formally adopted by the Commission. For European G-SIBs, MREL requirements will need to be set consistently with the FSB's Total Loss-Absorbing Capacity (TLAC) standard. Consultation responses on this are being considered alongside the results of the impact assessment. The standard will be finalised by the time of the G20 summit in November.

So no definitive statements today, but I want to address some recurrent issues, including some points that have cropped up in the FSB and EBA consultations.

I want to talk about three big themes:

• MREL and TLAC: how far are they the same or different?
• The relationship between the regulation of banks' liability structures via TLAC and MREL and resolution planning.
• And why, if we are preparing firm specific resolution plans, have a TLAC standard at all?

Starting with MREL and TLAC, much has been made of the differences between them. Let me set these out in turn:

• TLAC sets a global standard for G-SIBs¹ while MREL is for all EU banks and investment firms.
• TLAC describes a Pillar 1 minimum requirement, while MREL is set by the resolution authority, bank by bank on the basis of a resolution plan.
• The TLAC requirement is set in terms of risk weighted assets (RWAs) and leverage while MREL is formally set in relation to total liabilities and own funds.
• The FSB consultation specifies the quantum and quality of TLAC a G-SIB requires and sets out how TLAC should be distributed within groups. While BRRD provides for more flexibility, certainly on whether liabilities need to be subordinated to count as MREL.

So much for the differences. But both are founded on the same principles and in the Key Attributes have the same DNA. Both are designed to achieve the twin objectives of:

• Maintaining critical economic functions in the event of failure; and
• Doing so, without recourse to public funds.

What this implies for the size of MREL and TLAC is the same. Resources should be sufficient to support the continuation of critical economic functions at the point of resolution. Based on an assumption that going concern capital is eroded on the way in to resolution, this implies that there is sufficient loss absorbency to convert to equity capital to stabilise the firm. Stabilisation for this purpose means having resources available to allow the firm to be reauthorised by all relevant home and host supervisors and to maintain market access.

All of these factors are set out in the EBA's technical standard. The same considerations are described in the principles that cover the FSB Term Sheet.

It is worth pausing to unpack what this is likely to mean for different sorts of firm and different group structures. Certainly for G-SIBs, the assumption lying behind the calibration of TLAC is that given the size and complexity of their operations and their activities cross-border it is hard to conceive of a significant restructuring at the point of resolution. Rather the accent is on stabilising the firm via bail-in² to restore solvency to buy time for an orderly restructuring and/or a solvent wind-down afterwards. This should be good for the system and, by preserving value, for creditors of the firm.

By contrast, a very small bank may need no MREL beyond its current capital requirements. If, at failure, there is no obstacle to putting the firm into insolvency and paying out covered deposits using the relevant deposit guarantee scheme (DGS) then no MREL is required beyond capital requirements as a going concern.

This goes too to the location of TLAC or MREL in groups. There needs to be externally issued MREL wherever, under the preferred resolution strategy, bail-in of external liabilities is likely to occur or where other resolution tools are to be applied.

The plan under a single point of entry (SPE) strategy is that only the entity at the top of group is put into resolution on the basis that the interdependencies within a group are so extensive that resolving the different operating companies within a group separately is either impossible or not credible. By contrast where a group is organised in a more modular fashion with limited or no interlinkages between the parts, then the resolution can have multiple points of entry (MPE). It has been convenient for exposition to set up this dichotomy between SPE and MPE. But in practice, many SPE firms will have around the periphery subsidiaries that could be resolved separately or allowed to enter an insolvency without adverse consequences for the resolution of the firm. And many MPE banks are, in practice, collections of SPE strategies. They may be set up to be resolved jurisdiction by jurisdiction. But within jurisdictions the resolution is likely to involve keeping together in resolution the legal entities within that jurisdiction.

All of this can be summed up with two observations and one implication. The first observation is that the resolution strategy will govern decisions on MREL. The second is that TLAC can and will be applied for G-SIBs through the setting of MREL requirements. In other words TLAC is just one particular expression of how we will set MREL for firms where, as authorities, we are likely to face no choice other than bail-in in order to continue the firm's economic functions.

The implication is that there needs to be some level of disclosure about resolution strategies and liability structures at the legal entity level. Otherwise investors will not be able to price the risks they are exposed to. The Basel Committee intends to consult on new Pillar 3 standards for this purpose by the end of this year. This is welcome. But it is welcome too that banks are starting to do more to provide this kind of information to investors.

Let me turn to my second theme - the relationship between setting MREL and TLAC and resolution planning. The TLAC standard describes an end-point. Beyond proposing a conformance date of 2019 at the earliest, it does not describe how the standard will be implemented and what will happen in transition. Similarly European resolution authorities setting MREL in the EU will need to consider how to go from MREL in current liability structures to an MREL that ultimately meets the BRRD objectives. The RTS allows transitional MREL to be set until 2020.

Work needs to be done to clean up existing liability structures to make them consistent with requirements. This is firm level work which will be addressed as part of the resolution planning process. Let me give you some examples of what I have in mind.

• My first is structured notes. Some respondents to the FSB consultation suggested the blanket exclusion of structured notes from TLAC should be relaxed. We are reflecting on that. But the current situation is that in many cases structured notes have been issued to meet the needs of often quite small pockets of investors. The result is that some firms have many thousands or even tens of thousands of notes outstanding each with different structures and embedded derivatives. Trying to value all these instruments and bail them in would be a barrier to making resolution work. And as such will need to be addressed as part of the resolution planning process for a particular firm.
• My second example is non-equity capital instruments issued out of operating companies. As I described earlier, SPE strategies are built around putting only the top company in a group into resolution and avoiding legal risk and the potential dislocation of systemic functions by keeping operating subsidiaries outside resolution. But there is a tension with this strategy where banks want to count non-equity capital instruments issued out of subsidiaries as TLAC or MREL. The instruments may meet the other requirements to qualify - minimum maturity, subordination etc - but if the only way that they can bear loss is by putting the subsidiary into resolution there is a question of compatibility with the resolution strategy. As part of BRRD we have in the EU a statutory power at the point of non-viability (PONV) to write-down regulatory capital instruments. But the same is not true outside Europe. As part of the resolution planning process we will need to identify these cases and how to address them.

Incidentally one advantage of this PONV write-down power is that in resolution we can ensure the order of depletion of liabilities follows the creditor hierarchy. That is, losses are first absorbed by regulatory capital instruments in the operating company where they occur. If the losses are sufficient to wipe out the capital, then the operating company is recapped by write-down or conversion of any other internal TLAC. This pushes losses up into the parent company or holdco where they are absorbed following the creditor hierarchy, capital first and then debt.

• My third example of issues that will be addressed through resolution planning is foreign law debt. We cannot count as MREL or TLAC instruments where the statutory powers of the resolution authority to bail them in are uncertain. BRRD is explicit on this requiring, under Article 55, all third country unsecured liabilities to carry contractural terms that provide for bail-in. Most existing debt instruments do not. As part of the resolution planning process, we will ensure that they do and that our statutory powers in resolution are enforceable.

These examples illustrate the basic idea that there is much to be done through the resolution planning process to make existing liability structures compatible with resolution. This won't be a matter of grandfathering. Indeed even if a particular liability meets the definition of MREL, on the face of the directive it does not mean it does not represent a barrier to resolvability. As resolution authorities we will address these barriers as part of our resolution planning work and find ways to remove them. Or else we will adjust MREL requirements for firms. In the FSB proposal this is the placeholder for a Pillar 2 element alongside the Pillar 1 minimum requirements set out in the term sheet.

My third theme is why then a standard - if we are going to be preparing resolution plans and setting MREL for individual firms why have a standard at all? There is an obvious level playing field consideration, especially among the G-SIBs that compete with each other and operate internationally. But there is a deeper reason that goes to making resolution - and indeed international banking - work.

The resolution naysayers point to the events of the crisis and the Lehman failure in particular and say that cross-border resolution will never work, national interest will always win out.

Clearly the whole enterprise of resolution planning is to face up to this challenge ex ante. The Key Attributes envisaged Cooperation Agreements (CoAgs) among the authorities to describe a common resolution plan for an individual G-SIB. But to work CoAgs need underpinning and this is where TLAC comes in. Not only does TLAC provide countries with confidence that G-SIBs are operating with sufficient resources to be resolved safely. But that confidence is provided in a concrete form. This is the internal TLAC downstreamed in SPE groups from the parent entity to material subsidiaries. This is intended to provide hosts with comfort that in the event of a resolution they will not be exposed as losses would be upstreamed to the holdco, enabling the hosted legal entity to continue to operate. Hosts would therefore not need to take the sort of unilateral actions to ring-fence assets that were so damaging during the crisis and have tended to fragment the international financial system since. In short we need cross border resolution to work through structural changes like TLAC for international banking to work.

And we need this to be internalised by the market as well. I mentioned before disclosure for investors. But it is as important for liability holders that are unlikely to be bailed in to understand that too. Part of the rationale for locating TLAC at holdco in SPE strategies is to keep the resolution away from the operating companies and to be able to give liability holders at opco a strong message that where losses and recap needs can be met within TLAC they will not be bailed in. This way we maximise the chances of achieving continuity in the CEFs of the firm and avoiding any amplification in adverse systemic consequences.

To sum up, MREL or TLAC is a watershed on the road to making resolution work. Finalisation of the TLAC proposal this year and joint decisions on MREL in resolution colleges for EU firms in early 2016 will clear the way for banks to engage investors on their capital planning and debt issuance for the year ahead. Changes to existing liability structures may take time. But the clarity this will bring will mean that we have a regime that is not only feasible - after all we have the legal powers under BBRD to bail in the things we need - but credible.

---