Capital review - Total loss absorbing capacity ('TLAC')

Ref #7740912 v1.0 MEMORANDUM FOR Financial Sector Oversight Committee FROM Susan Guthrie DATE 11 Oct 2018 SUBJECT Total Loss Absorbing Capacity (‘TLAC’) FOR YOUR Decision We recommend BSG:

1. Note “TLAC” refers to the equity and debt instruments issued by a bank that the authorities assume will be either loss absorbing on a going-concern basis or able to be successfully and quickly used to bail-in holders (i.e. impose losses on them) during resolution. The FSB announced the TLAC standard in November 2015 and interim policies are due to take effect on 1 January 2019.
2. Note BSG and FSO were briefed about TLAC in 2014. 1 In 2017 TLAC was included within the scope of the Capital Review and our preliminary analysis has now been completed.
3. Note Our view is that imposing minimum TLAC requirements on New Zealand banks, or just the large four banks, is unlikely to produce meaningful net benefits for New Zealand at the present time. However there are design aspects of evolving TLAC policies that we believe warrant FP continuing to monitor the development and implementation of TLAC policies in other jurisdictions.
4. Note The global community’s interest in, and development of, TLAC policies provides some potentially valuable lessons for New Zealand.
5. Note As one of many factors, we will consider the TLAC rationale when setting minimum capital requirements for systemic banks (in particular, we will consider whether it is reasonable to require banks to have shareholder equity equal to twice the level needed to absorb unexpected losses on a going-concern basis).
6. Agree • We will not develop a TLAC policy at this time but will continue to monitor TLAC policy developments globally. • We will investigate the pros and cons of requiring banks to regularly report the value and class of liabilities owed by New Zealand banks to related parties, with each class defined by contractual features that impact on the reliability and likely speed of bail-in.

2 Ref #7740912 v1.0 Introduction

1. “TLAC” refers to the sum of equity and debt instruments issued by a bank that the authorities assume will be either loss absorbing on a going-concern basis, or able to be successfully and quickly used to bail-in holders (i.e. impose losses on them) during resolution (and thus provide the capital base for the post-resolution resurrected bank). The FSB announced minimum TLAC requirements in November 2015 and interim policies are due to take effect on 1 January 2019 (with full compliance by 2022).
2. BSG and FSO were briefed about TLAC in 2014, when the policy was first being debated internationally.1 The central purpose of TLAC, identified in these early briefing papers, remains today, namely “to ensure that banks structure their liabilities in a way that supports resolution by bail-in.” 2 However, at the time, development of the standard was still at an early stage and “some details as well as conceptual issues” were yet to be resolved. Hence a full analysis could not be completed at the time and it was agreed that FP would maintain a watching brief.
3. Until recently there has been no detailed analysis of the appropriateness of a TLAC policy for New Zealand. However in 2017 TLAC was included within the scope of the Capital Review and our preliminary analysis has now been completed. TLAC is aimed at systemic banks
4. TLAC is all about providing the capital base for post-resolution resurrected banks. As bankruptcy, rather than resurrection, is a viable resolution option for non-systemic banks, the expectation is that only systemic banks will be “too-big’-to-fail’ and therefore resurrected. Thus the expectation is that only systemic banks will be required to have aggregate TLAC funding in excess of their minimum regulatory capital requirements.
5. Thus we see, for example, the Financial Stability Board’s TLAC standard, issued on 9 November 2015, only applying to G-SIBS. And we see the Bank of England (BoE) applying the EBA’s FSB-equivalent, the MREL, in such a way that non-systemic banks, while subject to MREL, only face a minimum MREL requirement equal to their minimum regulatory capital requirement (this means non-systemic banks automatically meet their MREL requirement when they comply with their minimum capital requirements). TLAC policies highlight the importance of contract terms
6. Not all unsecured, subordinated debt is considered a reliable basis for bail-in. The BoE, for example, does not accept, for MREL purposes, liabilities “where the value of the liability depends significantly on derivatives or embedded derivative components”. This 1 At that time the policy was called GLAC or “gone-concern loss-absorbing capacity”. Refer Gone￾concern loss-absorbing capacity (GLAC) – a primer. Paper to FSO. Financial Policy. August 2014. Documentum 5830963. As well, a description of the BoE’s policy was provided to FSO in February 2016 (‘The UK resolution framework”, Op policy. Documentum 6361294) 2 At that time the policy was called GLAC or “gone-concern loss-absorbing capacity”. Refer Gone￾concern loss-absorbing capacity (GLAC) – a primer. Paper to FSO. Financial Policy. August 2014. Documentum 5830963. As well, a description of the BoE’s policy was provided to FSO in February 2016 (‘The UK resolution framework”, Op policy. Documentum 6361294)

3 Ref #7740912 v1.0 is because, in the BoE’s view, instruments like this are too difficult to “value rapidly in resolution”. This view extends to the “structured notes” and “liabilities subject to set off or netting arrangements”.3 7. Thus, in jurisdictions where the TLAC standard (or similar) applies, systemic banks may have to put a greater emphasis on issuing reliably bail-in-able debt than they have in the past. A simple principle underpins the setting of TLAC minimums 8. The assumption underpinning the FSB’s TLAC standard is that G-SIBs can reasonably be required to have capital and capital-like (i.e. bail-in-able) funding equal to twice what they require to absorb losses on a going-concern basis. This is because systemic institutions need to emerge from resolution with market support, and this means they need to be capitalised adequately post-resolution (and the best guide to this is the minimum going-concern capital required of them currently).4 9. Thus, under the FSB’s TLAC, G-SIBs will initially be subject to a minimum requirement of 16% of RWA and 6% of leverage assets (twice the minimum BCBS regulatory total capital requirements), increasing to 18% and 6.75% respectively by 2022. 10. In the UK, the ‘double-up’ rationale is stated explicitly. As well as applying the TLAC minimums to G-SIBS, by 1 January 2020 domestically systemic banks will face a minimum MREL requirement equal to “[2 x P1] + [1 x P2A] or 2 x LR” (i.e. twice Pillar 1 capital plus any “Pillar 2A” requirement or twice the level implied by the leverage ratio). 11. In Switzerland G-SIBs face a minimum TLAC requirement of 28.6% of RWA or 10% of leverage assets, which is double the minimum Tier 1 requirements facing these banks. However domestically systemic banks receive a TLAC “discount”, being subject to a minimum TLAC requirement only 1.4 times as large as the minimum Tier 1 requirement (for these banks, the minimum TLAC required is 18.5% of RWA or 6.5% of leverage assets). 5 Subsidiaries and groups 12. TLAC policies policy deal explicitly with the issue of failing subsidiaries in the context of banking groups. 13. This aspect of the standards seems to have been the most problematic to design, particularly once cross-border banking groups are considered. Important aspects of MREL’s group rules were still being consulted on by the BoE in late October 2017, for example, two years after the initial MREL consultation (the later consultation addressed issues arising in the context of groups with multiple-point-of-entry resolution strategies). 3 BoE (2015) The Bank of England’s approach to setting a minimum requirement for own funds and eligibile liabilities (MREL), consultation on a proposed statement of policy. December 2015. 4 In order to have market confidence the resurrected entity is likely to need other support too, such as liquidity support. 5 Refer Ernst and Young, Swiss Regulatory Update 2018 banking Ernst and Young Swiss report

4 Ref #7740912 v1.0 14. The treatment of groups and subsidiaries is also the aspect of TLAC that is potentially of most interest to us, given the subsidiary status of our large four banks. 15. Not-with-standing the BoE’s most recent consultation, in the published principles underpinning the TLAC standard, and in the MREL, there is an underlying assumption that most groups will operate with a single-point-of-entry resolution strategy. This means that when a subsidiary is facing insolvency, the group as a whole, rather than the subsidiary, will be the ‘resolution entity’, with losses incurred by the subsidiaries transferred to the group via bail-in-able within-group instruments (known as ‘internal TLAC’). The losses will then be transferred from the group to external parties via the bail-in-able instruments issued by the group (‘external TLAC’). The capital needed to resurrect the subsidiary post-resolution is therefore expected to ultimately come from third parties, but to get there it must navigate internal group bail-in arrangements.6 “When losses arise in an operating entity such that it is no longer viable, the intragroup liabilities will be written down or converted to equity (in general without the use of stabilisation powers), passing the losses up to the holding company (because these liabilities of the operating company represent assets for the holding company). If these losses cannot be absorbed by the holding company it will be placed into resolution, and the externally-issued liabilities of the resolution entity will be written down and/or converted as necessary using the bail-in tool. This means that in resolution of the group, formal stabilisation powers should only be applied at the resolution entity level.”7 16. Even when the group and subsidiary reside within one jurisdiction there are complexities that appear to have led to quite prescriptive TLAC policies, suggesting that resource-intensive monitoring (by the bank and/or regulator) might be required. For example, in the UK, a subsidiary that is not systemic will nevertheless have a MREL requirement that exceeds its minimum capital requirement if it is a ‘material subsidiary’ of a domestic group that is systemic (the requirement will be for ‘internal MREL’). “Material subsidiary” has to be defined in order to give effect to the policy and thus far the BoE has, following the FSB, decided this means the subsidiary “represents 5% of the group’s RWA”, operating income or leverage assets, which is something the bank and supervisor would have to monitor. 17. The resource implications of monitoring compliance with prescriptive TLAC rules is one reason we are sceptical of the value of a TLAC policy for New Zealand at present (although we acknowledge that one of our recommendations – regular reporting by banks about their TLAC-like liabilities – will nevertheless introduce new compliance costs on banks). Cross-border banking 18. The group aspect is particularly important in a cross-border context and it is this aspect which concerns us the most about a TLAC policy in the New Zealand context. Consistent with their single-point-of-entry underpinnings, the TLAC and MREL policies appear to presume seamless coordination, or at least no conflict, between home and 6 Banks that are deemed ‘resolution entities’ face minimum ‘external’ TLAC requirements, and subsidiaries face minimum ‘internal’ TLAC requirements. 7 Bank of England (2015) The Bank of England’s approach to setting a minimum requirement for own funds and eligible liabilities (MREL). Dec 2015.

5 Ref #7740912 v1.0 host regulators. TLAC seems predicated on differences between the authorities having been settled in a binding way prior to a bank entering resolution. Anything less than this risks a disorderly resolution, irrespective of amount of TLAC – internal or otherwise – issued by the failing bank. “Host authorities must have confidence that there is sufficient loss-absorbing and recapitalisation capacity available to subsidiaries in their jurisdictions with legal certainty at the point of entry into resolution. Without such confidence, host authorities could demand extra resources to be ring￾fenced in their own jurisdictions either ex ante or ex post in a resolution. The adverse consequences of such actions, including global fragmentation of the financial system, and disorderly resolutions of failed cross-border firms, should be avoided. To implement an orderly resolution, there must be sufficient flexibility to use loss-absorbing capacity within a G-SIB where needed. This means that there will need to be a credible mechanism by which losses and recapitalisation needs may be passed with legal certainty to the resolution entity or entities.”8

19. The BoE is explicit about this assumption. “In the case of banking groups with cross-border operations, the group will have a group resolution strategy that is agreed between the resolution authority of the parent company of the group (the ‘home’ authority) and the relevant authorities of the material subsidiaries located in other jurisdictions (‘host’ authorities). 9

20. It seems to us that more likely than not, in times of stress, home and hosts interests will depart and even pre-positioned agreements, made between regulators prior to a bank failing, will come under pressure. For example, a local subsidiary may be systemic to the host but not material to the parent group. In this case it may not be in the best interests of the group to resurrect the subsidiary, and thus bail-in third parties. Will the home regulator require the group to resurrect the subsidiary nevertheless? The host regulator will definitely seek this outcome but it does not obviously serve the purpose of the home authority.

21. Issues like this are perennial ones in the trans-Tasman banking relationship and have been the subject of much analysis and discussion. While we are open to pre￾positioning where it seems sensible, and to the possibility of single-point-of-entry resolution, it is our view that we need to retain the full suite of resolution options, including multiple-point-of-entry resolution and stand-alone options. Hence we have OBR and the outsourcing policy, both of which are compatible with either single or multiple-point-of-entry resolution.

22. Notwithstanding the debate about the merits of single versus multiple-point-of-entry, the possibility of within-group bail-in (delivered by ‘internal TLAC’) is relevant for our large banks. However even if our banks met the internal TLAC requirement it is not clear that their resolution would be more orderly as a result. Such a policy would,

8 Financial Stability Board (2015) TLAC principles and term sheet. Nov 2015 9 BoE (2017) Internal MREL – the Bank of England’s approach to setting a minimum requirement for own funds and eligible liabilities (MREL) within groups, and further issues. Oct 2017. OIA s6(a)

6 Ref #7740912 v1.0 however, add to the complexity of the prudential regime with all the added compliance and monitoring costs that implies. 23. It is still early days for TLAC. Jurisdictions are grappling with how to operationalise the policy for groups that are likely to pursue multiple-point-of-entry resolution, and policies might emerge that are a good fit with the New Zealand context. 24. As well, conceptually at least, ‘internal’ TLAC promotes the idea that groups (and thus potentially home regulators) should take responsibility for resurrecting failed subsidiaries, and should be willing to commit to act in a particular way, well in advance of any subsidiary failing. In this respect, a TLAC policy might become a mechanism by which we could (in theory at least) get the Australian authorities to pre-commit to particular actions to resolve the New Zealand subsidiaries at the Australian group level. 25. The BoE seems to view the pre-positioning delivered via internal TLAC as a very positive outcome. “Placing sufficient internal MREL resources at subsidiaries (referred to as ‘pre￾positioning’) should reduce the risk that the wider group is unable or unwilling to commit the funds to support it. By doing so, it provides greater clarity and confidence to that subsidiary’s financial counterparties, customers and creditors about how losses of the company will be addressed by the group resolution strategy. Similarly, pre-positioning provides assurance to overseas authorities that subsidiaries in their jurisdictions will be recapitalised. Aligning incentives between authorities in this way helps international co-operation in a cross-border resolution. In doing so, it offers comfort for host authorities who might otherwise set higher capital requirements… For branches of foreign banks that operate in the United Kingdom and which perform critical functions, the Bank [of England] will look at the adequacy of the group resolution plan in such cases, covering the amount and distribution of the group’s loss absorbing resources and how the branch will be treated…In cases where the Bank [of England] has concerns about the resolution plan, it will engage with the home authority to deal with the issue.” 10 26. Because policies are still being developed, and because of the potential benefits, we believe there are grounds for continuing to monitor the development of TLAC policies globally. LESSONS FROM TLAC POLICIES 27. The global community’s interest in, and development of, TLAC policies provides some potentially valuable lessons for New Zealand, namely: i) Underpinning the FSB’s TLAC policy, and versions of it such as the EBA’s MREL, is the view that, in the event capital is fully eroded in a systemic subsidiary, it would be reasonable for the resolution authority to require the group to provide sufficient capital (e.g. via being bailed-in) to resurrect the post- 10 BoE (2017) Internal MREL – the Bank of England’s approach to setting a minimum requirement for own funds and eligible liabilities (MREL) within groups, and further issues. Oct 2017.

7 Ref #7740912 v1.0 resolution entity as a going-concern. Importantly, this bail-in of the parent or other group entities can reasonably be required to occur before bail-in losses are imposed on other unsecured, subordinated creditors. This is entirely consistent with the Reserve Bank’s position on resolution, namely we first look to the group to support the subsidiary. ii) Also underpinning the TLAC standard and similar policies is the view that within-group bail-in of a failing subsidiary should occur first, before any subsequent steps (e.g. the group as a whole being placed in resolution). iii) It would seem unwise to assume all unsecured subordinated debt can be relied upon as a basis for bail-in in the context of resolution. Various contractual features can make some subordinate debt instruments difficult to value quickly, meaning they are of no use when resolving systemic banks. We noted earlier that unsecured subordinated debt that is subject to netting, for example, or has a value that relies on derivatives, does not qualify for MREL status in the UK. iv) It would seem important to regularly monitor the quantity of bail-in-able debt on issue by systemic banks. This would mean, for example, having banks regularly report the value of their unsecured subordinated liabilities that can be relied upon to be bailed in, and those that can’t (those that can’t be relied on will include debt that is dependent on derivatives, or embedded derivative components, for their value, structured notes, or debt that is subject to netting arrangements).11 The Reserve Bank does not appear to currently collect this data (or, if it does, it is not part of the regular statistical surveys). v) Underpinning the FSB’s TLAC standard and similar policies is the view that systemic banks can reasonably be required to have capital and capital-like (i.e. bail-in-able) funding equal to twice what they require to absorb losses on a going-concern basis. This is because systemic institutions need to emerge from resolution with market support, and this means they need to be capitalised adequately post-resolution (and the best guide to this is the minimum going￾concern capital required of them currently). This view underpins the G-SIB policies being adopted globally, for example, and also the policies for domestically systemic banks in some countries (e.g. the UK) but not all (e.g Switzerland’s use of a factor of 1.4 for domestically systemic banks). vi) Regulators could require banks to meet their TLAC requirement entirely with shareholder equity (as proposed by the Federal Reserve Board of Minneapolis, for example). That is, assuming the TLAC requirement is set at twice the regulatory capital minimum, requiring banks to meet their TLAC requirement entirely with shareholder equity would mean banks would become subject to regulatory action (not necessarily resolution) once regulatory capital fell below the TLAC minimum.12 11 BoE (2015) The Bank of England’s approach to setting a minimum requirement for own funds and eligibile liabilities (MREL), consultation on a proposed statement of policy. December 2015. 12 We have not looked in detail at how such a policy would work, but the range of options available to the regulator when the TLAC requirement was breached could in theory include suspension of dividends and being required to raise new capital rather than moving immediately to resolution.

8 Ref #7740912 v1.0 Recommendations 28. The FSB’s TLAC standard takes effect from 1 January 2019 but it only applies to G￾SIBs. However other jurisdictions are developing similar policies for domestically systemic banks. Hence TLAC policies are still evolving internationally. 29. TLAC’s application to subsidiaries, and the clear message that parents should be the first port of call when a subsidiary falters, resonates with our own view on resolution. Hence there are aspects of TLAC that have conceptual appeal and we recommend keeping a watching brief on TLAC policies adopted in other jurisdictions. 30. There is an inevitable intermingling of TLAC and resolution and to the extent any further analysis of TLAC is subsequently undertaken, it will be done in conjunction with the further development of resolution policies. There is, however, an immediate short￾term step which we believe is worth taking – and that is to ask the banks for more information about their liabilities, focusing on liabilities grouped into classes based on the ease with which holders can be bailed-in, and for asking the large four banks for information about the extent to which parents or other group members are holders of easily bailed-in liabilities. 31. In summary, we recommend: • Not developing a TLAC policy at this time but continuing to monitor TLAC policy developments globally. • Investigating the pros and cons of requiring banks to regularly report the value and class of liabilities owed by New Zealand banks to related parties, with each class defined by contractual features that impact on the reliability and likely speed of bail-in.