Capital review - A monkey on our back

Ref #7876538 v1.1 MEMORANDUM FOR FSO FROM Financial Policy (Ian Woolford) DATE OF MEETING 14 March 2016 SUBJECT A monkey on our back? FOR YOUR Discussion We recommend that FSO:

1. Notes there are policy uptake and design decisions that have consequences that may test both our supervisory resources and – arguably – our philosophy.
2. Advise whether there is a desire for follow-up papers, such as a stocktake and assessment of approval-type policy, and whether the issues raised in this paper resonate or require any other sort of follow-up, be it as standalone papers or assessments against supervisory resourcing and priorities interweaved in policy recommendations. Background

1. First things first: why the title? This was actually one of two working titles, the other being “Something’s got to give” – equally cryptic but it didn’t grab the attention of PSD managers to the same extent, so it lost out. Both titles are also slightly provocative, and deliberately so. This reflects this paper’s intent, which is to prompt us to step back, assess the direction of policy and supervision, the nature and degree of resourcing in PSD, and the priorities assigned to various BAU and project work. (I will explain this in more detail below, and will refer to them collectively as ‘the trade-offs’.)
2. This paper was commissioned following a discussion at a recent PSD Managers away-day about growing concerns that some of us have about the balance of the trade-offs, and in particular the growing amount of a particular type of supervisory activity: ‘approvals’ or non-objections, although it can be cast even more broadly than this to include our philosophy and the direction of travel in Basel policies. As well as trying to promote a more birds-eye view of the trade-offs, the other intent is to gauge whether FSO is interested in taking the discussion any further, and if so, what sort of papers would support further more detailed decisions.
3. In short, the basic proposition is that while resourcing has grown in PSD, so has the scope and the scale of our work. Sometimes that growth in scope is because of external policy developments, sometimes because we have taken on new responsibilities (e.g. the insurance sector), and sometimes because of the prioritisation choices we have made.
4. This is not a paper whining that we are under-resourced or overworked – this is a paper that attempts to stand back and look at these developments against the backdrop of our stated

2 Ref #7876538 v1.1 philosophy (the three pillars and policy choices that are fit for New Zealand), and to ask, have we, through time, incrementally taken on more risk and responsibility through the way we regulate than our philosophy and supervisory resourcing suggests we should? That is, have we allowed a monkey on our back? The elements of the ‘trade-offs’ Scope and nature of policy 5. Over the last 15 years, there has been significant growth in what we do (helpfully presented in Appendix 1 in a table produced by Paula and Wendy). We have gone from regulating the banking sector only (and overseeing payments without any real powers) to also being responsible for NBDTs and insurers, AML supervision, and are seeking more powers in the payments area. 6. There has also been a sizeable increase in staff numbers, almost exclusively motivated by the growth in sector responsibilities. For example, staff numbers grew from 20 in 2000, to 54 in 2015. 7. There has also been growth in the volume of regulation. Obviously, when two prudentially unregulated sectors (NBDTs and insurance) came under our control, they did that for the main purpose of being prudentially regulated, and so regulation in these areas sprung up. We have a relatively simple prudential framework for the NBDT sector that isn’t obviously broken. The legislative framework for insurers is up for review, and important aspects of the solvency standards were recently reviewed. By design and necessity the solvency standards are much more complicated than capital requirements for the NBDT sector. This makes sense given the nature of their business, the traditional approach to analysing capital for risk for insurance, and the diversity in the sector, even though insurers don’t pose the same sort of systemic concerns as the banking sector. 8. On the banking side, the volume of regulation is unambiguously up. In crude terms, the handbook was a slimmer version in 2000, comprising BS1 through BS7. As it currently stands, it is a much more complicated document, and the chapters now run to BS19. BS2B alone is 203 pages long. The FSAP team will probably recommend adding another chapter or two at least (large exposures almost certainly, there will doubtless be others). 9. The external drivers for this growth have all been clear: Basel II, Basel III, and to a lesser extent the 2004 FSAP. 1 And more is to come. We will be bringing a paper to FSO in April updating the Committee on Basel’s five (!) current capital consultations, some of which are quite fundamental. One at least, proposes a move away from complexity by making the standardised approach more complex. The market risk framework is much more complicated than our current requirements, and possibly too complicated for our domestic banks to competently comply with (although we have yet to properly assess the requirements). The one theme

1 There have been internal drivers as well of course, perhaps crisis management-related ones being the main culprit.

3 Ref #7876538 v1.1 across all of the Basel proposals is more capital, and sometimes vastly more, and more complexity. These papers will all come through FSO in due course, but the point is, the drivers for expansion are predictable and relentless. 10. What is more interesting for the purposes of this paper, however, is not just the growth in banking requirements, but the nature of the requirements. What we have seen, particularly since the GFC, is a drive for more complexity (sometimes depicted as moving towards simplicity, albeit through more granularity as a means to more accurately reflecting risk). In part because of the complexity, in part because of deep regulatory concerns about the limits of self discipline, and in part because G20 regulators are typically resourced at much higher levels than we are, we have also seen a move towards what I would call ‘approval-based regulation’.2 This is where the monkey makes its entrance. Approvals 11. Capital instruments are one example of growing complexity. In order to ensure that all capital instruments can absorb losses on a going concern basis, the Basel guidance requires that AT1 and Tier 2 capital instruments have the potential to be converted into ordinary shares or be written off on the occurrence of certain triggers. The Reserve Bank has chosen to implement this requirement by requiring conversion or write off to be a term of contract of the instruments (Basel III also allows this requirement to be met through legislative provision). Given this approach the contractual terms of the instrument give the Reserve Bank or the statutory manager the power to trigger conversion/write-off, which may occur if the bank is ‘non-viable”. The Reserve Bank’s and statutory manager’s powers have therefore been extended through the terms of contract of capital instruments, which act as an extension of our failure management regime. The Basel III requirements also impose a number of other requirements for the Reserve Bank to make decisions in respect of the instruments; for example a bank must receive Reserve Bank approval before repaying a capital instrument. 12. One concern that we have with this approach is that due to the complexity of the instruments, and the interaction of the contractual terms with our legislation, there is a level of legal risk that the instrument may not convert or be written off as intended. For example, conversion requires actions to be taken by the Reserve Bank or a statutory manager; we have worked intensively with instrument drafters to ensure that the instrument correctly applies our powers and the powers of the statutory manager. Existing corporate law does not envisage the write off of instruments, and there are therefore some areas of legal risk that exist in respect of conversion. 13. We have undertaken a number of steps to mitigate the legal risks associated with these instruments. First our requirements impose a fall back requirement in the case of a failure to convert; that is if conversion fails the instrument must be written off. Second we have imposed a requirement that banks receive a notice of non-objection to treat the instrument as capital. This process has required a high level of resource within PSD, and is technically complex.

2 In our framework we have things that we have to “approve” and things where we provide a “non-objection”. Both are relevant to the arguments in this paper, and both essentially serve the same purpose.

4 Ref #7876538 v1.1 However, BSG considered that in a world where requirements are complex, and contractual terms impose powers or obligations on the Reserve Bank or statutory manager, a review by the Reserve Bank is necessary. That is, if we want to allow these instruments, we need to control for the incentives banks face and the legal risk associated with the instruments, and this requires us to get our hands dirty and accept there is a bit of a monkey on our back. For me at least though, it does beg the question about whether we are resourced appropriately to support a G20 bells and whistles framework. Or, whether a part of the answer is to think about more radical alternatives such as moving to a pure common equity capital framework (for example), or at least move perpetual debt to tier two, but require them to be convertible. The former would do away with complexity in this area at least, the latter mitigate the risk by increasing the quality of the tier one buffer. 14. Aside from the resourcing issue, one concern with this approach is that we have shifted to a regime where boards of banks place weight on approvals from us. We would be concerned if this undermined the level of due diligence undertaken by banks, and have gone to some lengths to develop mechanisms to ensure that banks, and bank boards, implement a comprehensive due diligence process in respect of the capital instrument, to minimise the extent to which our non-objection undermines internal processes. Boards however do place some reliance on the Reserve Bank non-objection, and board approval generally (although not always) occurs after the Reserve Bank has confirmed it has no objections to the terms. 15. To date FP has assessed all capital instruments. From 2016 the arrangement within PSD is that Financial Policy leads the work on the more complex instruments, and supervisors lead the work (with Financial Policy support) for repeat issues or less complex instruments. Some supervisors are uncomfortable with this arrangement, arguing that it requires specialist training/qualifications, and given the number of instruments there is still a significant resource strain from this process. 16. Outsourcing is another area of approval based regulation, particularly at the moment as we are poised to deal with the hump of outsourcing activity, relating to both in-train proposals, and the expected transition period. 17. The entire IRB framework is arguably the best example. The models are all approved (or rejected) by the Banking Steering Group, on the advice of Financial Policy. But the models presented and assessed are not the end of the story, and while we have sought to introduce elements to the framework to better incentivise self discipline (making non-adherence to approved models an enforcement issue for example), there are a myriad of ways in which banks can reduce capital to levels (far) below what we expected when we approved the model. It is extraordinarily difficult to monitor exactly what is going on, hence we get drawn into complex exercises like the benchmarking exercise to try and definitively ascertain the extent of compliance and capital adequacy. It is also why Basel see one answer as putting floors in, measured against a more complex standardised framework.

5 Ref #7876538 v1.1 18. It doesn’t stop here though. Appendix 1 also illustrates the extent of the main approval-based regulation, and therefore monitoring / enforcement required by supervisors, but it is not even a complete list. 19. Amongst the various approvals, there will be a variety of views about whether they are all necessary, or whether some could indeed be left to management and the board of a bank to attest to. For example, in my view, the benefit of a regulator approving risk management programme changes for insurers is marginal, whereas it is crucial for approving internal models or outsourcing. In simple terms we tend to think about how incentive compatible our requirements are, and whether the regulator and regulated incentives align or depart. Our incentives align in terms of BCP, good risk management and so on, it is not debateable that when it comes to capital or outsourcing to parents our incentives and objectives depart (or put another way, what is privately optimal is almost always less than what is socially optimal). Options 20. The trade-offs are about resourcing, the nature of our framework, and the prioritisation and design choices we make within the framework. All of which is set against the backdrop of compliance with international standards and our philosophical approach. 21. We articulate our philosophical approach as being relatively light-handed, focussing on three pillars (self, market, and regulatory discipline), being conservative, and following international standards where it suits but not where it doesn’t. On this latter point, we have tended to assess where it suits and where it doesn’t more in terms of the coherence of the approach advocated by the Basel Committee as a standalone proposition, as well as against our framework. What we haven’t done enough of, I would argue, is to sufficiently take account of the resourcing consequences or whether a simpler more conservative approach could achieve our soundness and efficiency objectives, even if it is at the expense of a move away from strict adherence to Basel. 22. At the heart of our early articulation of the three pillars was the concern about moral hazard that can arise from regulators controlling and signing off risks that regulated entities are better placed to manage. It was thought to be better to align self-discipline incentives by having very powerful enforcement penalties for the courts (prison sentences for directors) and tools (deregistering banks). The idea was more weight could be put on self-discipline because of the consequences of not following the rules. However, no enforcement action has ever been taken against a bank to my knowledge. To my mind, the extremely heavy emphasis on self-discipline was always a tad naïve, particularly so without any demonstration effect! In a nutshell – one can simultaneously place a significant emphasis on the role of directors as well as recognise that there are limits to self-discipline as a regulatory tool. 23. So if we are in a world where the demands on regulators and supervisors are growing, one option is to increase the number of resources to meet the growing demand. Another option is to cut your coat to suit your cloth, and prioritise. Focus on the stuff that really matters. This isn’t a paper arguing for increasing staff numbers especially in the current environment, and we have always been comfortable being relatively lighthanded, comforted by also being relatively

6 Ref #7876538 v1.1 conservative. Nor is it a paper asking FSO if we can prioritise, because at the working and PSD management level this is happening all of the time and we can point to numerous examples. We could, arguably, be more systematic and possibly even ruthless about prioritisation, but if we had a policy of discarding (or simply not doing) supervisory monitoring we are taking another monkey on our back if a significant issue emerged from the quarter we have deprioritised. 24. Especially if resources are fixed, the bigger questions are whether we have incrementally built (or are building) a framework that has strayed from our philosophical underpinnings, whether through the growing use of approvals/intrusion, or through offering the full suite of Basel policy and home grown options that may be individually coherent, but are unable to be fully serviced by increasingly stretched supervisors. Even with very technical issues, it is relatively easy for policy to crank the handle – but a key question is whether supervisors can meet the growing demands. 25. If you accept that we face an issue here, as PSD management do, we face a spectrum of options. One option is as new Basel policies roll out, we explicitly factor into our policy advice the implications for the ability of the policy to be adequately supervised. A variant of this is to also assess whether more could be sheeted back to directors, although if we do more of that, then I would argue we should have a hard discussion about the lack of enforcement over the years and the willingness to enforce in the future, as well as how aligned regulatory and director incentives are. Perhaps this is actually less of an option, and more of an action point for policy developers. As mentioned above, there will be plenty of opportunities to do this as Basel capital policies are assessed and we bring them through FSO Committee over the coming months. 26. A potentially more far-reaching radical option would be to do a stocktake of our approvals￾based policies – or even all of our policies – to assess how consistent they are with our regulatory philosophy. This could be stand alone, or more first principle thinking and options factored into policy reviews, such as the upcoming capital policy review. 27. First and foremost this note is designed to provoke a discussion, and to alert FSO Committee members to a growing feeling in PSD that as we continue to go down a more-or less complete alignment with Basel requirements there are consequences for how we approach our supervision, and balance the and allocate risks between regulated entities and ourselves. But the general points made in this paper are not just about banks and Basel, the same goes for insurance and IAIS, although this is a bit less so for NBDTs. 28. One final point is that our philosophical position is referred to several times in this paper, in a way that suggests there is common agreement within PSD management. At a high level I think that is right, but of course by its very nature a philosophical position can mean slightly different things to different people. It may be worthwhile having a more structured recap and discussion at FSO on this issue at least – and of course any follow papers or action points that emerge from the discussion of this paper.

7 Ref #7876538 v1.1 By the end of 2000 2005 2010 2015 2020 ? BS1 (Principles) BS1 (Principles) BS1 (Principles) BS1 (Principles) BS2 (Capital) BS2 (Capital) BS2A (Standardized) BS2A (Standardized) BS3 (Bank registration) BS3 (Bank registration) BS2B (IRB) BS2B (IRB) BS4 (Audit) BS4 (Audit) BS3 (Bank registration) BS3 (Bank registration) BS5 (AML & CFT) BS5 (AML & CFT) BS4 (Audit) BS4 (Audit) BS6 (Market risk) BS6 (Market risk) BS5 (AML & CFT) BS5 (AML & CFT) BS7/BS7A (Disclosure) BS7/BS7A (Disclosure) BS6 (Market risk) BS6 (Market risk) BS8 (Connected Exposures) BS7/BS7A (Disclosure) BS7/BS7A (Disclosure) BS9 (Significant Influence) BS8 (Connected Exposures) BS8 (Connected Exposures) BS10 (Suitability Assessment) BS9 (Significant Influence) BS9 (Significant Influence) BS10 (Suitability Assessment) BS10 (Suitability Assessment) BS11 (Outsourcing) BS11 (Outsourcing) BS12 (ICAAP) BS12 (ICAAP) BS13/BS13A (Liquidity) BS13/BS13A (Liquidity) BS14 (Corporate Governance) BS15 (Significant Acquisitions) BS16 (Capital recognition) BS17 (OBR) BS18 (Covered Bonds) BS19 (High-LVR restrictions) Insurance Licensing/ Supervision AML/CFT Supervision BS2B (IRB model changes) BS2B (IRB model changes) BS9 (Significant influence) BS9 (Significant influence) BS9 (Significant influence) BS10 (Suitability) BS10 (Suitability) BS10 (Suitability) BS11 (Outsourcing) BS11 (Outsourcing) DSS (rules + conditions changes) DSS (rules + conditions changes) BS15 (Significant acquisitions) BS16 (Capital instruments) BS18 (Cover bonds) IPSA s26 (Changes of control) IPSA s44 (Transfers & Amalgamation) IPSA s73 (RM Programme changes) IPSA s109 (Statutory Fund approval) NBDT change of ownership NBDT suitability assessment PSD Full-time equivalent staff 20 25 49 54 ? *Excludes a variety of policy initiatives or workstreams (eg: Deposit Insurance, OTC derivatives, Trans Tasman Banking Council (TTBC) etc.) at present ? Banking Supervision Handbook ? Monitoring/ Licensing Pre- GFC Bank Supervision Post-GFC Bank Supervision ?

• minimal supervisory engagements - more frequent supervisory engagements
• emphasis on disclosure statements - increased private reporting
• system-wide thematic reviews (housing and rural lending review & ICAAP review) Payment Systems Designated Settlement Systems (DSS); Part 5C of RBNZ Act 1989 Non-bank Deposit Takers (NBDTs) Licensing / Oversight Approvals/ Non-objections The Evolving Risk Profile of PSD Legislation* Reserve Bank of New Zealand (RBNZ) Act 1989 ? Insurance (Prudential Supervision) Act [IPSA] 2010 Anti-Money Laundering and Countering Financing of Terrorism (AML/CFT) Act 2009 Reserve Bank Amendment Act 2008. Part 5D of RBNZ Act 1989. Regulatory framework for Non Bank Deposit Takers (NBDTs)