Response to Consultative Document 10 November 2014 on loss-absorbing capacity of GSIBs
Norges Bank and Finanstilsynet's letter of 30 January 2015 to the Financial Stability Board.
Norges Bank and Finanstilsynet welcome international initiatives towards ending the "too big to fail" (TBTF) problem and avoiding taxpayer-supported bail-outs of large financial institutions. FSB's proposal for an international standard on a minimum requirement for total loss absorbing capacity (TLAC) at global systemically important banks (GSIBs) is an important step in this direction. This proposal should be considered part of and supported by other measures to mitigate the TBTF problem, such as strengthened capital and liquidity requirements. Furthermore, in order to facilitate the resolution process, adequate resolution planning of each GSIB is essential.
We consider the proposed cross-border TLAC allocation to represent an important step towards an agreed framework for crisis resolution of large, systemically important cross-border banking groups. The proposed TLAC standards could also be applied to domestically systemically important banks.
In order to obtain a simple and transparent set of regulatory rules, one should avoid exceptions and variations in rules designed to satisfy specific jurisdictions or specific groups of banks. Attempts to meet such special needs not only undermine the purpose of the regulation, but can also add considerably to its complexity.
Norges Bank and Finanstilsynet strongly support the efforts to strengthen the resolvability of GSIBs through a minimum requirement for their loss absorbing capacity. For the credibility of orderly resolution of GSIBs, it is important that all eligible liabilities in the TLAC requirement can easily be written down or converted to equity over a short period, usually during a weekend.
We consider the calibration of the size of the TLAC as somewhat low, in particular taking into account the need for a buffer against future losses when a bank reopens right after it has been resolved through bail-in. For the bank to regain trust in the market, unsecured short-term creditors need to be assured that the reopened bank is creditworthy and that there are no more hidden losses. To avoid government guarantees in future bank resolutions, it is essential that the resolution authorities bail-in sufficient capital to ensure that resolved banks have an extra buffer against potentially hidden losses. The proposed range for Pillar 1 appears to be based on losses in banking crises over the last couple of decades. Many of these bank problems were then resolved with the help of government guarantees. Hence, from a prudential perspective the proposed range for Pillar 1 TLAC (16 – 20% of RWA without regulatory buffers) seems to be on the low side.
Norges Bank and Finanstilsynet support that TLAC requirements should be benchmarked against both total un-weighted exposures and risk weighted assets (RWA). However, as RWA are of less relevance once a bank fails or is likely to fail, it would be inappropriate to only benchmark TLAC towards this measure.
We also support the proposal that all TLAC eligible liabilities should be contractually junior to non-eligible liabilities in the creditor hierarchy, but senior to subordinated debt. Without such subordination of eligible liabilities, a resolution authority may not be able to bail in all of them without violating the "no creditor worse off" principle.
Norges Bank and Finanstilsynet agree that TLAC should be coordinated with Basel III in such a way that capital used to meet the CET1 buffer requirements cannot count as TLAC. This should ensure that TLAC requirements do not have any undesired interference with the countercyclical buffer. If the CET1 buffers are included in the TLAC, the countercyclical buffer might not work as intended, when the macro-prudential authorities decide to reduce it. In such a hypothetical situation, banks that have met their TLAC requirement only with a margin might not be willing to relax their lending practices as intended by the macro-prudential authorities. Nevertheless, if CET1 buffers are included in the TLAC, we assume that the calibration of the minimum Pillar 1 TLAC requirements should be increased accordingly.
We support the proposal to locate TLAC at resolution entities in group structures, and in addition to require a significant part of the TLAC to be pre-positioned in material subsidiaries. The appropriate size and location of TLAC is crucial for timely and effective crisis resolution. We consider the TLAC proposal to be a valuable supplement to earlier FSB guidance on resolution strategies, i.e. single-point-of-entry (SPE) or multiple-point-of-entry (MPE).
Pre-positioning of TLAC on the balance sheet of material subsidiaries should ensure host authorities that sufficient loss-absorbing capacity will be available in the local subsidiary to secure an orderly resolution process. If, in addition, the TLAC is in the form of bail-in-able debt to the parent bank, this should facilitate a quick recapitalisation of the subsidiary and enable it to continue to operate as a going concern. Pre-positioning will reduce the incentives for host authorities to impose pre-emptive ring-fencing measures in a crisis. The suggested range of pre-positioned TLAC should, in our view, be set closer to the upper limit of the indicated range (of 75-90% of the TLAC requirement that would apply on a stand-alone basis). Norges Bank and Finanstilsynet support a process whereby the actual figure will be set by the relevant host authority in consultation with the home authority.
It is essential that the internal TLAC is readily available and can be converted and written down without legal challenge. On-balance sheet TLAC will be far superior to inter-group guarantees. Collateralized guarantees, as suggested in the term sheet, could easily be subject to legal disputes.
The suggested criteria for material subsidiaries seem appropriate. Agreeing on which subsidiaries are material according to these criteria will be critical for the successful implementation of the proposed TLAC proposal.
Successful resolution of cross-border banking groups will also require adequate funding arrangements, ref. the earlier FSB guidance note on effective resolution strategies (July, 2013). Thus, an assurance that sufficient liquidity will be available to a material subsidiary in an SPE resolution is complementary to the capital assurance provided under the current TLAC proposal. Such parent liquidity assurances should become part of the COAG's operational resolution planning.
Finally, we will stress the importance of transparency in banks' creditor hierarchy once bail-in and TLAC are introduced. It is essential that all investors in bank liabilities know not only the seniority of their claims, but also the size of their tranche as well the size of all capital and liability tranches with lower seniority. Home and host authorities have a joint responsibility to make clear the creditor hierarchy in their respective jurisdictions, including the ranking of deposits and scope of depositor preference, if applicable, ref. Key Attributes (KA 7.4). Without such information, investors will not be able to price the risk of the various bank debt instruments correctly.
For further details, see the enclosed answers in the Q&A sheet.
|Ida Wolden Bache
Acting Executive Director
Norges Bank Financial Stability
|Erik Lind Iversen
Deputy Director General
Answers to questions from FSB in Consultative Document 10 November 2014 on loss-absorbing capacity of GSIBs
Calibration of the amount of TLAC required
1. Is a common Pillar 1 Minimum TLAC requirement that is set within the range of 16 – 20% of risk-weighted assets (RWAs), and at a minimum twice the Basel III leverage requirement, adequate in the light of experiences from past failures to support the recapitalisation and resolution objectives set out in this proposal? What other factors should be taken into account in calibrating the Pillar 1 Minimum TLAC requirement?
- Taking into consideration that the CET1 held to meet Basel III buffer requirements cannot be included in TLAC, the Pillar 1 range may seem adequate. A study of banking crises since 1988, conducted by the Basel Committee, suggests that a loss absorbing capacity of 24% of RWA would have been sufficient to cover 95% of the cases. However, Norges Bank and Finanstilsynet would point to at least two arguments why the suggested range is probably on the low side:
- After a bank has been resolved through bail-in and reopens, it is important that unsecured short term creditors of the reopened bank trust that there are no more hidden losses from which they may suffer. To ensure that, resolution authorities need to bail-in an extra buffer beyond what seems to be expected losses at the point of time when the bank enters into resolution.
- In several of the banking crises covered by the Basel Committee study, explicit as well as implicit government guarantees may have limited the recorded bank losses. The perceived absence of supportive government measures, going forward, thus indicates that expected losses conditional upon a banking crisis may be higher.
2. Does the initial exclusion of G-SIBs headquartered in emerging market economies (EMEs) from meeting the Common Pillar 1 Minimum TLAC requirement appropriately reflect the different market conditions affecting those G-SIBs? Under what circumstances should the exclusion end?
3. What factors or considerations should be taken into account in calibrating any additional Pillar 2 requirements?
- A bank's business model
- Complexity of group structures
- Lack of progress in developing acceptable recovery plans
- Difficulties in developing resolution plans
- A low margin of bank's own funds relative to their capital requirements (pillar 1 as well as pillar 2)
- A large share of liabilities that will be economically or legally difficult to bail in
- Rapid growth in a banking group's lending
Ensuring the availability of TLAC for loss absorption and recapitalization in the resolution of cross-border groups
4. Should TLAC generally be distributed from the resolution entity to material subsidiaries in proportion to the size and risk of their exposures? Is this an appropriate means of supporting resolution under different resolution strategies? Which subsidiaries should be regarded as material for this purpose?
- We agree with the proposal to allocate TLAC in cross-border groups according to the size and risk of material subsidiaries' exposures. We also agree with the suggested criteria for material subsidiaries, but note that the practical implementation of the proposal hinges on agreement between home and host on which subsidiaries are "material". The proposal will be a critical supplement to the agreed framework for crisis resolution (SPE or MPE), as it gives substance to these general principles. Cross-border groups should be explicit about whether their prevailing resolution method is MPE or SPE, and both supervisory colleges and crisis management groups should agree further on the level, location and composition of TLAC.
5. To what extent would pre-positioning of internal TLAC in material subsidiaries support the confidence of both home and host authorities that a G-SIB can be resolved in an orderly manner and diminish incentives to ring-fence assets? Is a requirement to pre-position internal TLAC in the range of 75 - 90% of the TLAC requirement that would be applicable on a stand-alone basis, as set out in the term sheet (Section 22), appropriate to satisfy the goals of the proposal and ensure that TLAC is readily and reliably available to recapitalize subsidiaries as necessary to support resolution? Can this pre-positioning be achieved through other means such as collateralized guarantees?
- Prepositioning TLAC as suggested could potentially bridge the gap between home and host authorities in SPE resolution models, as it would give host authorities the required assurances about parent support. This will certainly be opposed as a measure to fragment the group, but we should note that the counterfactual – without such measures – could easily be a messy "grab for assets" and ring-fencing in the early stages of an approaching crisis. Thus, we find the proposal well balanced and support the mechanism. We would add that on-balance sheet TLAC will be far superior to collateralized guarantees.
Determination of instruments eligible for inclusion in external TLAC
6. Are the eligibility criteria for TLAC as set out in the term sheet (Sections 8-17) appropriate?
- The eligibility criteria set out are broadly appropriate. However, Norges Bank and Finanstilsynet consider the two last paragraphs of Section 13 problematic. Both these paragraphs constitute special exceptions to the general rule and may undermine the purpose of a TLAC requirement. Furthermore, such exceptions will no doubt contribute to complicating financial regulation and thus making comparisons of banks across various jurisdictions more difficult.
7. What considerations bear on the desirability of an expectation that a certain proportion of the common minimum Pillar 1 TLAC requirement consists of (i) tier 1 and tier 2 capital instruments in the form of debt plus (ii) other eligible TLAC that is not regulatory capital?
- As long as the bail-in of all debt instruments that are part of TLAC is credible, it should not matter for a bank's loss absorbing capacity whether the TLAC requirement is met only with own funds or also with some debt that is not part of own funds. Based on this, we are sceptical to demand that a certain proportion of the common minimum Pillar 1 TLAC requirement consists of debt not part of own funds. However, banks that choose to meet the TLAC requirement only through own funds and banks that also use other eligible debt to meet the TLAC requirement, should be subject to the same types of intervention if they suffer similar losses on their assets.
8. Are the conditions specified in the term sheet (Section 8) under which pre-funded commitments from industry-financed resolution funds to provide resolution funding contribute to TLAC appropriate?
- Norges Bank and Finanstilsynet do not think this derogation is appropriate. It may serve to undermine the purpose of the TLAC requirement by not making investors in bank debt take their fair share of losses in a bank in which they have invested. Furthermore, we note that such commitments cannot be used in the European Economic Area according the Bank Recovery and Resolution Directive (Directive (EU) 2014/59).
9. Is the manner in which subordination of TLAC-eligible instruments to excluded liabilities is defined in the term sheet (Section 13) sufficient to provide certainty regarding the order in which creditors bear loss in resolution, and to avoid potentially successful legal challenges or compensation claims? Where there is scope for liabilities which are not subordinated to excluded liabilities to qualify for TLAC, are the transparency and disclosure requirements set out in section 13 and 24 sufficient to ensure that holders of these instruments would be aware of the risk that they will absorb losses prior to other equally ranking but excluded liabilities? If not, what additional requirements should be adopted?
- As regards Section 13: The first two paragraphs are clear. However, the second to last paragraph that allows for non-subordination of liabilities eligible for TLAC in jurisdictions where liabilities not eligible for TLAC also are statutory excluded from bail-in, is problematic. Also in this case, liabilities that are eligible for TLAC should be subordinated to those excluded from bail-in. In that way, one can avoid NCWO problems. The last paragraph is however not sufficiently clear to make holders of the relevant instruments aware of the risk that they will absorb losses.
- Section 24 is sufficiently clear.
Interaction with regulatory capital requirements and consequence of breaches of TLAC
10. Do you agree that the TLAC requirement for G-SIBs should be integrated with Basel III such that the minimum TLAC requirement should be met first, and only after TLAC is met should any surplus common equity tier 1 (CET1) be available to meet the Basel III buffers?
- We agree that TLAC should be coordinated with Basel III in such a way that capital used to meet the CET1 buffer requirements cannot count as TLAC. This ensures that TLAC requirements do not have any undesired interference with the countercyclical buffer. If the CET1 buffers are included in the TLAC, the countercyclical buffer may not work as intended, when the macro-prudential authorities decide to turn it off. In such a hypothetical situation, banks that met their TLAC requirement only with a margin might not be willing to relax their lending practice as intended by the macro-prudential authorities. Nevertheless, should the CET1 buffers be included in the TLAC, we would suggest that the calibration of the minimum Pillar 1 TLAC requirements should be increased accordingly.
11. What disclosures (in particular in terms of the amount, nature and maturity of liabilities within each rank of the insolvency creditor hierarchy) should be required by resolution entities and material subsidiaries to ensure that the order and quantum of loss absorption in insolvency and resolution is clear to investors and other market participants?
- It is essential that all investors in bank liabilities know not only the seniority of their claims, but also the size of their tranche as well the size of all capital and liability tranches with lower seniority. Home and host authorities have a joint responsibility to make clear the creditor hierarchy in their respective jurisdictions, including the ranking of deposits and scope of depositor preference, if applicable, ref. Key Attributes (KA 7.4). Without such information, investors will not be able to price the risk of the various bank debt instruments correctly. Hence resolution entities should publish such information regularly.
Limitation of contagion
12. What restrictions on the holdings of TLAC are appropriate to avoid the risk of contagion should those liabilities be exposed to loss in resolution?
- The suggested restrictions are important and adequate.
13. Should G-SIBs be required to conform with these requirements from 1 January 2019? Why or why not? What, within the range of 12 to 36 months following the identification as a G-SIB, should be the conformance period for banks identified as G-SIBs at a future date?
- Although implementation of TLAC clearly requires some transition due to issues of grandfathering and difficulties in renegotiating existing debt contracts, it might be desirable to shorten the conformance period a little. The last question points to the desirability of increasing the scope of TLAC i.e. let it also apply to domestically systemically banks.
Market impact and other aspects
14. How far is the TLAC proposal, if implemented as proposed, likely to achieve the objective of providing sufficient loss-absorbing and recapitalization capacity to promote the orderly resolution of G-SIBs?
- The TLAC proposal, if implemented at the higher end of the suggested range, should go a long way towards providing sufficient loss-absorbing and recapitalisation capacity at global, cross-border systemically important banks.
15. What will be the impact on G-SIB's overall funding costs of the adoption of a Pillar 1 Minimum TLAC requirement?
- We assume that banks funding costs may increase somewhat. This will mostly be due to less implicit guarantees for banks, and as such the increase in funding costs is welcome.
16. What will be the impact on the financial system and its ability to provide financing to the real economy?
- The proposal will contribute towards reducing the TBTF problem. In that way it also contributes to the banking system's ability to financing the real economy on a sounder basis.
17. Do you have any comments on any other aspects of the proposals?
- See the main letter.
 See BCBS Calibrating regulatory minimum capital requirements and capital buffers: a top-down approach, October 2010.
 Experience from Denmark in handling Amager Bank certainly points in this direction.