M-Capital-Ratio

M.A. Lawler They said he was a good communicator (well obviously he met the primary requirement, (financial excellence) and he is, the new Governor (who cares enough to take an income cut to take on the task). Slowly his wish to communicate is filtering through to those who should be informing the owners of RBNZ that the public can indeed understand - but as long as writers use only the code of the profession, in this case economics and financial management - they cannot. The non-technical summary: How much capital is enough is short and clear. Thank you. I may venture into the Consultation paper, knowing in advance the gist of what I read, ie, the whys, the wherefores - and where the gains and losses of the change lie. The article is a perfect example of democracy in action. More please. Thank you. Ann Lawler

Mark Lyons Subject: Re Proposed Capital Changes for New Zealand banking system Hi, I will make this brief and you have my permission to publish this as a submission. Australian Shareholders in Banks are in no mood for this shit. Mark Lyons OIA s9(2)(a) OIA s9(2)(a)

From: Martien Lubberink Sent: Friday, 14 December 2018 2:41 PM To: Subject: RE: many thanks for your time and the time of your team yesterday Great proposal! Martien OIA s9(2)(a) OIA s9(2)(a)

From: To: Capital Review; Ian Woolford Subject: Bank capital review Date: Friday, 17 May 2019 1:21:50 PM Attachments: Cover_letter_RBNZ.pdf Comments_Paper_4_Lubberink.pdf Dear Ian, I appreciate the opportunity to provide feedback on RBNZ’s Capital Review Paper 4. Please see my accompanying comments. Kind regards, Martien Lubberink OIA s9(2)(a)

Martien Lubberink May 17, 2019 Ian Woolford Financial System Policy and Analysis Department Reserve Bank of New Zealand PO Box 2498 Wellington 6140 Dear Ian, I appreciate the opportunity to provide feedback on RBNZ’s Capital Review Paper 4. Please see my accompanying comments. Kind regards, Martien Lubberink OIA s9(2)(a)

How Much Capital is Enough? Comments on the RBNZ 4th Capital Review Paper DR. MARTIEN LUBBERINK∗ May 17, 2019 ∗Author. OIA s9(2)(a)

c Martien Lubberink (2019) Comments on RBNZ’s Capital Review Paper 4. Contents 1 Executive Summary: A reserve bank that does not want to supervise 3 2 Main Comments 6 2.1 The Governance of the process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 2.2 A narrow focus on CET1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 2.3 The RBNZ should consider introducing a Pillar 2 framework. . . . . . . . . . . . . . . . 15 3 Minor Comments 18 4 Answers to specific questions 19 5 Boxed materials 21 6 Bibliography 24 2

c Martien Lubberink (2019) Comments on RBNZ’s Capital Review Paper 4. 1 Executive Summary: A reserve bank that does not want to supervise

1. The current prudential regulation renders the New Zealand financial system vulnerable to the impact of significant shocks. The RBNZ relies on self-discipline and market discipline. Compared to other supervisors, the New Zealand Reserve Bank takes a rudimentary, almost Spartanic approach to bank supervision. Pillar 1 dominates, there is no Pillar 2 to speak of. Because of a fear for moral hazard, supervision was not meant to be intrusive. There is no deposit insurance system. Thus far, this has had no adverse impact on the financial system. I guess we have been lucky. However, the current approach to supervision is out of date and unsustainable in its current form. In some way, New Zealand banking can be compared to cycling in the Netherlands. Since the 1970’s the Netherlands has created a large cycling infrastructure.1 As a result, cycling in Holland has been very safe. If you visited Amsterdam, you must have noticed that there is hardly any traffic enforcement. Dutch cycling also relies on self￾discipline and monitoring (other cyclists, cars, pedestrians). However, this system now starts to show cracks. The number of bicycle casualties is on the rise, for two reasons. First is the ageing population: vision loss and hearing loss do affect older adults, impairing their ability to monitor traffic. Second is technological advancements: battery-powered bicycles have become increasingly popular. Combined, these two factors make cycling in the Netherlands today more lethal than driving cars, which was never the case.
2. The New Zealand approach to prudential supervision, shaped by a reliance on self-discipline and market discipline, has become an anachronism. The GFC demonstrated the limits of market discipline and depositor monitoring. The post-GFC era shows the limits of self-discipline: poor conduct is a growing problem. In light of these developments, the RBNZ approach to prudential supervision increasingly reflects a set it and forget it attitude. Like cycling in the Netherlands, the New Zealand financial system, once deemed very safe (see Calomiris and Haber (2014)), has become vulnerable.
3. Breaches of capital requirements will become visible quickly. The prize-winning Financial Strength Dashboard allows the New Zealand public to identify weak banks and act accordingly. Prompted by social media, retail investors fearing the OBR haircut will move money abroad, to well-managed and resilient European, Far East, or North American banks. Money that was deemed to be sticky in the past has become hot. 1See the BBC, 8 August 2013: Why is cycling so popular in the Netherlands? 3

c Martien Lubberink (2019) Comments on RBNZ’s Capital Review Paper 4. 4. Meanwhile, new risks are emerging. They affect the resilience of the New Zealand financial system. Operational Risk has become much more important: examples are cyber security risk, the risks associated with money laundering, and other conduct risks, such as Libor related scandals, as well as scandals that led to the Royal Commission into Misconduct in the Banking, Superannuation and Financial Services Industry.2 A recent article in the Financial Times showed that banks’ risks managers are most concerned about cyber security risk, and that the attention for operational risk had increased by seven percentage points over the last year. Quoting Paul Sharma: “There is definitely more of a focus on conduct rather than capital.”3 5. No simple solutions in banking. Andrew Haldane some years ago championed the virtues of the Leverage Ratio. Haldane (2012) claimed that the Leverage Ratio would be a ‘simple ratio’, thus render￾ing it superior to the risk-based ratio. However, Haldane failed to acknowledge the era of accounting scandals. About ten years before his speech, companies like Enron and WorldCom demonstrated that the simplicity of accounting is deceptive. Enron kept information off-balance, which rendered the reported values of assets and leverage unreliable. In other words, the idea that there exists a simple measure of solvency is misleading.4 The same applies to the different ways of calculating Risk Weighted Assets. Metrobank, the UK contender bank, which lost a whopping 75 percent of its market value because it did not get banking regulations right, demonstrates that the Standardised Approach is not a panacea for small banks that are encouraged to compete against established rivals. Likewise, the idea of a “wall of equity” suffers from the same problem. Deutsche Bank in February 2016 demonstrated that investors will get very nervous once a bank approaches a breach of its capital requirements. It is not obvious that requirements of 4 percent, 8 percent, or 18 percent will calm investors once they anticipate a breach. 6. The nordic banks have recently demonstrated that high levels of capital will not prevent a bank from getting into trouble. Dankse Bank, plagued by money laundering problems that wiped out half of its market value, may have suffered from a supervisor who applied a set it and forget it approach.5 7. Lastly, regarding the writing down or writing off of CoCo capital, the RBNZ prefers a high trigger. However, the idea that a high CoCo trigger will contribute to financial stability may suffer from the same keep it simple, set it and forget it mindset. 2See the link for the final report of the Royal Commission here. 3See FT.com, 25 March 2019: Conduct replaces capital under the spotlight. 4See an excellent comment on this point by Bloomberg’s Matt Levine New Leverage Rules Find Banks Some More Assets. 5See Reuters, 30 April 2019: Supervisors ignored Russian warnings over money laundering at Danske: document. 4

c Martien Lubberink (2019) Comments on RBNZ’s Capital Review Paper 4. 8. The 4th Capital Review Paper: more of the same. Against this backdrop of an increasingly vulnerable banking system, emerging risks, supervisory complacency, and the longing for the simple world of yesteryear, the RBNZ presents a proposal that can be best characterised as more of the same. 9. Briefly summarised the main elements of the RBNZ proposal RBNZ (2019) are: • Tier 1 capital requirements at 15 percent of Risk Weighted Assets (RWAs), • a 1 percent surcharge for systemically important banks, • a structure of regulatory capital where Common Equity Tier 1 (CET1) dominates, • a more prominent role for a conservation buffer, which comprises solely of CET1, • a significantly diminished role for non-equity capital, • a more prominent role of the Standardised approach of calculating risk weights, and • a significantly diminished role for the Internal Ratings Based approach of calculating risk weights. 10. More: Total capital ratios for New Zealand banks will be 17 percent overall, and 18 for systemically important banks, up from 10.5 percent. The main capital component driving this increase is the conser￾vation buffer, which adds 5 percent to the current buffer of 2.5 percent. The D-SIB and Countercyclical Buffer contribute 2.5 percent to the increase in capital.6 11. The same: As under the current rules, these are all Pillar 1 requirements, meant to cover credit risk, market risk, and operational risk. 12. However, the RBNZ proposal fails to acknowledge the vulnerabilities of the current system: i) better-informed and more savvy retail investors and depositors; ii) the growing importance of op￾erational risks, iii) the complexities of banking and the na¨ıve belief in simple solutions. Instead of acknowledging operational risk and liquidity risk, RBNZ (2019) focuses on a risk that is largely under control: credit risk. 13. More of the same is not an improvement, it leaves the current financial system vulnerable to the impact of significant shocks. The governance of own funds regulation at the Reserve Bank is such that a next official may dial the ratio requirements down to levels that are comparable to other countries. I doubt if such a move will contribute to safety and stability, hence my submission. 14. My comments therefore focus on the structure of capital. The RBNZ should let go of the idea of the “wall of equity” and accept layers instead. Layers in the form of, for example, contingent convertible capital. But also “layered” degrees of capital disclosures, thus following the example of Europe’s Pillar 2 framework, which is now split into two parts, of which the “guidance” part allows the supervisor to do its job: supervise. This instead of the current approach which seems to be inspired by Madame de Pompadour’s famous adage: Apres nous, le d ` eluge. ´ 6The box on page 22 present a more detailed view on the ratios. 5

c Martien Lubberink (2019) Comments on RBNZ’s Capital Review Paper 4. 2 Main Comments 2.1 The Governance of the process 15. The New Zealand public deserves to be properly informed. The quality of the banking system is the result of the combined quality of the banks and the regulator and supervisor. The governance of policy initiatives like the one on bank capital is important. See Mortlock (2019a,b,c) and my earlier post in interest.co.nz for more comments along these lines. 16. A lack of coordination. Worryingly, there are few signs of coordination between the Reserve Bank and Treasury (e.g. regarding resolution and recovery, or regarding Phase 2) or between the Reserve Bank and APRA, despite Governor Orr asserting that the RBNZ is in a “constant dialogue with our Australian colleagues at APRA. On this particular topic we have shared many conversations, papers, and briefings over a period of many months.” See, for example, my comment on the capital tug-of-war on page 8. 17. The poor governance raises questions about the sustainability and credibility of the proposals. Will a next RBNZ official introduce another capital review? Will investors and depositors trust the ensuing framework? 18. The research the Reserve Bank relies on should be unbiased. The selection of studies the Reserve Bank relies on appears to reflect a passionate desire to please those with a particular view. For example, the RBNZ largely ignores papers like DeAngelo and Stulz (2015), Gropp et al. (2016), Jorda et al. (2017), Dautovic¸ (2019) and Berndt et al. (2019). Gropp et al. demonstrate the adverse effect of changes in capital requirements. They find that their sample banks increase capital ratios by reducing their Risk￾Weighted Assets and not by raising their levels of equity. Jorda et al. , in a rigorous study examining the entire history of advanced economies between 1870 and 2013 and for the post-WW2 period find that higher capital ratios are unlikely to prevent a financial crisis. Dautovic¸ finds that “increased risk taking is compensating the positive results on solvency arising from higher shareholders’ capital so that the net effect on banks’ probabilities of default is insignificant.” Berndt et al. find that banks’ creditors (i.e. not the government) now appear to expect to bear the losses of failing large, systemically important banks. 19. The box on page 21 shows additional evidence on Reserve Bank’s choices for literature, specifically on the use of internal models. 6

c Martien Lubberink (2019) Comments on RBNZ’s Capital Review Paper 4. 2.2 A narrow focus on CET1 20. The reliance on Pillar 1 Common Equity capital is an important weakness of the Reserve Bank’s proposed capital framework. First, the volume of required additional equity is large. This raises questions about how the equity capital requirements will be met. I suspect that funding in the form of common equity will be difficult (if at all possible) in practice. Second is the absence of layers that can act as shock absorbers. International evidence shows that layers such as CoCos, Tier 2, or Senior Non-Preferred (SNP) can contribute to a resilient banking system at relatively low cost. Third is that the higher New Zealand capital requirements will likely prompt a response of the Australian supervisor. It’s not that overseas supervisors will sit still and ignore the higher capital requirements imposed on their subsidiaries by a host supervisor. Excess capital of New Zealand subsidiaries may be repatriated, thus affecting their ability to absorb losses. 21. The required additional amount of equity is large. Though estimates vary, the additional amount of Common Equity required to satisfy the RBNZ proposals is about $25bn, based on my own calcu￾lations, assuming a one percent management buffer. This for sure is a large amount. To illustrate, the $25bn is about the same amount of equity reported by UK’s Nationwide Building Society, or the market capitalisation of Allied Irish Banks. The $25bn is also the combined market value of Contact Energy, Meridian Energy, and Genesis Energy combined. In other words, if you have $25bn to spare, you can use that to start an entire new bank or buy a significant chunk of the NZ energy sector. 22. In practice, the NZ market may be too small to support $25bn fresh bank equity capital. Susan Guthrie’s memo of 12 September 2017 hand-waves some important comments on the prudential and practical difficulties of funding additional equity capital. For example, the memo suggests New Zealand banks can issue shares: “it would seem preferable to require registered banks (that can issue ordinary shares) to list a portion of their issued shares on NZX. This would offer potentially relevant information and help develop a deeper and more diversified local equity market.”7 Paragraph 137 of RBNZ (2019) also suggests the new capital requirements can be funded by issuing shares. 23. Given the volume of the capital requirements alone, listing shares on the NZX is a stretch. In theory doable, but in practice costly and unlikely realistic. 24. The Big-4 parent banks may not come to the rescue willingly, they are constrained by rules on double leverage. Susan Guthrie, in the same memo, thinks that the risk of double leverage “arguably 7See the memo here: Next steps for the definition of capital. 7

c Martien Lubberink (2019) Comments on RBNZ’s Capital Review Paper 4. over-states the case given that shareholders in the parent would be fully compensated for any additional risk the parent was exposed to by funding their investment in other ways” (§73). The box on page 23 of this submission shows the risks of double counting of capital. 25. The GFC taught us that the problem of double leverage is important. Most prudential supervi￾sors restrict the amounts a bank can borrow to fund equity investments in other banks. It is not obvious that APRA or other home supervisors will happily accept their banks to recapitalise New Zealand sub￾sidiaries using borrowed money. 26. Then there is the capital tug of war. Related to the issue of double leverage is the treatment of holdings by banks in other banks at solo level (below consolidated level). This is a problem requiring coordination between home and host supervisors. Host supervisors (the RBNZ being one given that it hosts the Big-4 banks) want as much capital as possible in their own country. Early on this year, Governor Orr fired a shot across the bow “Group capital requirements are not relevant for New Zealand creditors and/or taxpayers. Only capital in the New Zealand subsidiary can be relied on. This is set by the RBNZ’s requirements, not APRA.” 8 Orr is in good company, see the excerpt of the comment letter from the Polish FSA on the Financial Stability Board’s TLAC proposals in Figure 1. 9 Figure 1: We want capital here! 27. Beggar thy neighbour. The Polish FSA wants the subsidiaries of large foreign banks to locate as much a possible capital in Poland. However, home supervisors may retaliate. Bank of England’s Sam Woods expressed the worry that Brexit and the subsequent loss of passporting rights would prompt European host supervisors to bolster capital requirements for hosted bank subsidiaries.10 UK home banks would then be forced to issue shares to stem the prudential qualms of EU bank supervisors. Alternatively, the BOE could force bank subsidiaries to repatriate excess capital.11 8See RBNZ letter of 10 January 2019 here. 9See the link to the comment letter here. 10See the speech by Sam Woods, 4 October 2017: Geofinance. 11Paul Tucker stresses the need for supervisors to cooperate in this speech: The Resolution of Financial Institutions Without Taxpayer Solvency Support: Seven Retrospective Clarifications and Elaborations. 8

c Martien Lubberink (2019) Comments on RBNZ’s Capital Review Paper 4. 28. Home supervisors may change rules to force New Zealand subsidiaries to repatriate capital. APRA may increase the risk weight applied to solo holdings: from the current weight of 400 percent to a significantly higher percentage (or require full deduction).12 This would limit the loss absorbing capacity of Australian-owned bank subsidiaries. In addition, if APRA were to force the NZ subsidiaries to repatriate capital, then this would, on the margin, benefit subsidiaries owned by EU banks (Rabobank), given that the risk weight under the European implementation of Basel III is only 250 percent.13,14 Figure 2: Great capital plans for the two of us! 29. Issuing shares on the NZX will attract a penalty: the minority interest haircut. The RBNZ pro￾posal (RBNZ, 2019) and Susan Guthrie’s memo of 12 September 2017 suggest that New Zealand banks could issue shares to meet the higher capital requirements. However, issuing shares out of subsidiaries will introduce a minority interest at the consolidated level. The minority interest may or may not qualify as regulatory capital at the consolidated level, where the current rules governing the prudential treatment of minority recognises only the minority interest net of any surplus of Common Equity Tier 1 of the subsidiary. The surplus is Common Equity Tier 1 of the subsidiary minus the minimum Common Equity Tier 1 requirement of the subsidiary and the capital conservation buffer. Under Basel III rules the surplus would be the amount of CET1 over 7.0% (4.5% minimum requirement plus 2.5% capital conservation buffer). That surplus would be haircutted.15 12See APRA Prudential Standard APS 111, Attachment D - Regulatory adjustments, Paragraph 8 (f): “Equity exposures held in overseas deposit-taking institutions . . . that are subsidiaries of the ADI, after deduction of any intangibles component, must be risk-weighted at 300 per cent if listed and at 400 per cent if unlisted.” 13See my blog post on the capital tug-of-war here: Double leverage, a regulatory tribulation. 14A playful spreadsheet to learn about intra-group capital can be found here. 15See APRA Prudential Standard APS 111, Attachment C, Paragraph 4. “The amount that may be included in Regulatory Capital at Level 2 is the: (a) total amount of the capital attributable to third parties; less (b) any surplus capital amount above the minimum regulatory requirements, as calculated in paragraphs 5, 6 and 7 of this Attachment.” 9

c Martien Lubberink (2019) Comments on RBNZ’s Capital Review Paper 4. 30. Awareness of the minority interest haircut is low, but effects can be significant. RBNZ seems unaware of the rules on minority interests. The consultation papers and documents of the information release mention the word-pair ‘minority interest’ twice - in the initial ‘issues’ paper of May 2017 (RBNZ, 2017d). However, the effects can be significant. For example, in 2017, ABN AMRO admitted to having misread the Basel III rules on minority interests and was forced to take a e3.43bn haircut on its capital instruments.16 31. The RBNZ should consider CoCos as an attractive alternative for equity capital. Given the practical limits on equity funding, CoCos could come to the rescue. The second Capital Review paper (RBNZ, 2017b) as well as the associated Response to submissions RBNZ (2017a) present a case against contingent convertible capital instruments (Additional Tier 1, or CoCos). According to the RBNZ, the main argument to remove CoCos from the capital framework is the lack of loss-absorbing effectiveness. 32. CoCos could cover the out-sized demand for going-concern bank capital. There is a healthy investor appetite for CoCos. Glover and Beardsworth (2019), for example, show that the market for Contingent Convertible securities (CoCos) has grown significantly under Basel III rules, in particular in Europe. See Figure 3. Figure 3: A growing CoCo market, source Bloomberg. 33. The market for CoCos is now well-developed, liquid and deep. There is a CoCo index (Glover, 2014). Contractual issues have been largely ironed out. AT1 instruments have become standardised, thanks, for example, to the EBA’s efforts to monitor these securities.17 CoCos can be sold in offshore markets, which adds to their attractiveness.18 16See the ABN AMRO press release EBA interpretation impacts certain capital ratios of ABN AMRO Group here. 17See the link to EBA Additional Tier 1 instrument monitoring here. 18See for example the Lloyds AT1 issuance of 2 October 2018, listing particulars. 10

c Martien Lubberink (2019) Comments on RBNZ’s Capital Review Paper 4. 34. CoCos can thus facilitate monitoring of joint stock and non-joint stock companies. Figure 4 demonstrates that CoCos are now held by professional investors. The upper graph shows the distribution by investor type of a recent AT1 issuance by Societ´ e G´ en´ erale. The lower graph shows that AT1 capital ´ of ANZ and CBA track equity share price: the correlation coefficient between series are 0.58 and 0.74 respectively. Note the downward trend after the publication the Royal Commission report. (a) Societ´ e G´ en´ erale 1.25Bn NC5 AT1 of October 2018. ´ (b) ANZ Equity and average CoCo prices. (c) CBA Equity and average CoCo prices. Figure 4: Sophisticated CoCos. 35. CoCos absorb losses in going concern. Evidence from research and recent international cases demonstrates that CoCos do work. They absorb losses in going concern as well as in gone concern. Lubberink and Renders (2018) demonstrate going concern loss absorption: In the lead-up to the imple￾mentation of Basel III, European banks repurchased debt securities that traded below par. Banks engaged in these Liability Management Exercises (LMEs) to realise a fair value gain that prudential rules exclude from regulatory capital calculations. The LMEs enabled banks to augment Core Tier 1 capital, given that 11

c Martien Lubberink (2019) Comments on RBNZ’s Capital Review Paper 4. alternative methods to increase capital ratios were not feasible in practice. Using data of 720 European LMEs conducted between April 2009 and December 2013, Lubberink and Renders show that poorly capitalised banks repurchased securities and seamlessly converted e41,55bn of European bank hybrid capital into e32.44bn equity. The cost of executing LMEs, however, is not trivial. Banks lost about e9.1bn in premiums to compensate holders of hybrid debt for surrendering their fixed income invest￾ments. If prudential rules allowed banks to recognise the gains on a poorer credit standing without the need to execute LMEs, then this amount could have been added to Core Tier 1 as well.19 36. A low trigger facilitates CoCos gone-concern effectiveness. Regarding the proper level of trigger￾ing CoCos, the Reserve Bank claims that a low trigger e.g. 5.125% or 7%, doesn’t work.20 The Reserve Bank claims that “The . . . literature suggests that the sort of contingent debt that has been required under Basel III, has emerged in the market place, and is likely to emerge, will not be triggered in time to pro￾vide going-concern capital.” (RBNZ, 2017a, §81). And “The recent experience of Spanish bank Banco Popular highlights this concern. Despite having issued going-concern contingent instruments (some with 7% triggers) none of the instruments became loss absorbing prior to the bank being deemed non-viable.” (RBNZ, 2017b, §185). 37. The Basel III rules do not confine the enactment of conversion to breaching a numerical trigger value: “The trigger event is the earlier of: (1) a decision that a write-off, without which the firm would become non-viable, is necessary, as determined by the relevant authority; and (2) the decision to make a public sector injection of capital, or equivalent support, without which the firm would have become non-viable, as determined by the relevant authority” (BCBS, 2011). The first clause enables a supervisor to require a bank to convert CoCos to stay afloat. 38. The Banco Popular case then demonstrates the second clause of BCBS (2011) at work: ‘Equiva￾lent support’ came from Santander, which enabled the firm to continue to function as a bank without affecting financial stability. Daniele Nouy, chair of the supervisory board at the European Central Bank ` confirmed the CoCos did their work: “modestly . . . I would say that we have passed the test.” (Nouy, 2017). Andreas Dombret, former Member of the Executive Board of the Deutsche Bundesbank chimes in “There is one thing that I would like to stress from the outset which is all too easily forgotten when assessing the recent crises: neither in the case of the Spanish bank nor the Italian banks have we seen 19See for a publicly available copy of the research, this link. 20According to Basel III rules, the trigger level for a write-down or a conversion of CoCos must be at least 5.125% Common Equity Tier 1 (BCBS, 2017a) 12

c Martien Lubberink (2019) Comments on RBNZ’s Capital Review Paper 4. lasting spillover effects or negative repercussions in the markets. Branches opened their doors again on the next working day following the resolution - admittedly under a new name but otherwise it was business as usual. The fact that it all went largely unnoticed is the most remarkable thing about it, and that should by no means be taken for granted” (Dombret, 2017). 39. Many of the Reserve Bank’s worries about CoCos are unjustified. Referring to Italian banks, the RBNZ worries about CoCos because they may prompt moral hazard behaviour, may force a bail￾out instead of a bail-in, or may affect retail investors. According to the Reserve Bank, these cases demonstrate that “contingent debt sold to retail investors did not, in fact, absorb bank losses” (RBNZ, 2017b, §33). “Recent cases in Europe are consistent with contingent debt being incapable of providing going-concern capital” (RBNZ, 2017a, §81). These worries are not justified by the facts. 40. For example, the RBNZ fails to acknowledge that the bail-in rules of the European Bank Recovery and Resolution Directive (BRRD) had entered into force only on 1 January 2016 (EC, 2014, Art 130). This is important, because Italian retail investors bought significant amounts of subordinated bank capital in the years leading up to the entry into force of the BRRD bail-in rules. For example, in 2011, Italian households held e35bn in subordinated bonds, about 9.4% of their bond holdings. 41. On the first day of 2016, many Italian investors would not have known about this sudden change in risk. Hence, it is unfair to expect retail investors to have fully adjusted to bail-in rules on the first of January 2016. And yet, the RBNZ seems to do just that. In fact, the RBNZ conflates EU State aid rules and compensation for mis-selling securities: “The Monte dei Paschi di Siena example shows that contingent debt issued to retail investors may not, in fact, be loss-absorbing - tax payers may end up bailing out investors.” (RBNZ, 2017b, §34) 42. Is it fair to impose losses on investors who could not know of bail-in rules? The European Commis￾sion offers a clear answer: “In situations where banks that have mis-sold financial instruments have left the market, it is up to Member States to decide whether to take exceptional measures to address social consequences of mis-selling as a matter of social policy. This falls outside the remit of State Aid rules.” (EC, 2017). 43. In the case of Veneto Banca and Banca Popolare di Vicenza, the European Single Resolution Board ruled that there was a need to rely on the BRRD EC (2014). The Single Resolution Board, the European body that deals with banks entering resolution, decided that the application of national law would be sufficient to achieve an orderly liquidation. Hence, in the case of the two Venetian banks, one cannot 13

c Martien Lubberink (2019) Comments on RBNZ’s Capital Review Paper 4. claim that the CoCos they issued failed to absorb losses under European bail-in rules, as these rules did not apply (Micossi, 2019). 44. Lastly, EU bank rules are not about saving tax payers at all times. State aid and precautionary recapitalisations are legitimate measures under European bank rules, albeit that conditions apply: the institution must be deemed solvent, the public funds may not be used to offset existing or anticipated losses, and losses will have to be covered by the bail-in of shareholders and creditors (at least 8% of the bank’s liabilities) (EC, 2017, Dombret, 2017). 45. In sum, the notion that recent Italian cases demonstrate the lack of loss absorbing effectiveness of CoCo capital, which underpins the Reserve Bank’s case against Basel III compliant Additional Tier 1 securities, is hardly compelling. In fact, the RBNZ fails to acknowledge important EU regulations that aim to protect both taxpayers and retail investors. Calling that a failure is inappropriate. 14

c Martien Lubberink (2019) Comments on RBNZ’s Capital Review Paper 4. 2.3 The RBNZ should consider introducing a Pillar 2 framework. 46. The RBNZ proposal RBNZ (2019) relies predominantly on Pillar 1. This pillar is about publicly known capital requirements. Two recent cases demonstrate that i) a reliance on publicly known, high, capital requirements does not necessarily contribute to financial stability and ii) there is a demand for capital requirements that are agreed between banks and their supervisors and are not publicly visible. These requirements are known as Pillar 2 capital requirements. 47. Pillar 2 is important in many countries outside New Zealand. See Figure 5. Graph (a) shows Pillar 1 and Pillar 2 requirements of European banks (thanks to the EBA’s short-lived initiative to make banks disclose full capital requirements) are significantly higher than Pillar 1 requirements. Graph (b) shows the frequencies of reported capital ratios and required capital ratios for 122 countries. The graph shows that reported capital ratios are significantly higher than the required ratios, with means of 17.32 percent and 9.38 percent respectively. Given that Pillar 2 requirements are not disclosed in many coun￾tries outside Europe, the inference form the graph is that Pillar 2 requirements largely explain the gap between reported and required capital ratios. (a) EU Pillar 1 & Pillar 2 requirements. (b) Reported versus required capital ratios. Figure 5: Pillar 2 requirements. 48. Both graphs demonstrate that Pillar 2 requirements are used in many countries. They have gained prominence in Europe (Bevilacqua et al., 2019, Melis and Weissenberg, 2019). The two examples below show why. I hope these examples inspire the RBNZ to adopt a similar approach. 49. Example 1: High capital requirements per se will not calm investors and depositors. A case in point is Swedbank, which for its 2017 year-end reported a CET1 ratio of 24.6 percent. This is high for sure. However, Sweden’s oldest bank recently lost about half of its market capitalisation when it became 15

c Martien Lubberink (2019) Comments on RBNZ’s Capital Review Paper 4. publicly known that it channelled about $95 billion kronor ($15bn) in laundered Russian funds through its Baltic operations.21 The Swedbank example thus demonstrates that requiring a high capital ratio will not stop investors and depositors becoming nervous. It is therefore not obvious that applying high ratio requirements to all banks will improve matters. Likewise, Spanish Bank Banco Popular reported healthy capital ratios before its demise in 2017, i.e. a CET ratio of 12.13 percent at the end of 2016. 50. Example 2: Investors, depositors get nervous when a bank approaches a breach of require￾ments. On 8 February 2016, Deutsche Bank issued a press release that spooked financial markets.22 It assured investors that the bank would continue its payments on Additional Tier 1 capital instruments. Such an announcement is bad news of course. I mean, why state the obvious: that the bank will pay its fixed income investors? Apparently, the payment was not that obvious because Deutsche Bank was get￾ting dangerously close to breaching its capital requirements. Deutsche Bank’s disclosed capital require￾ment, as determined by the European Central Bank was 10.76%. Moreover, the bank was loss-making, and its CET1 ratio hovered about 2.5 percent over its requirement.23 In this context of poor performance and a CET1 ratio closing in on its requirement, the announcement of 8 February 2016 dragged down Deutsche Bank’s share price. It also affected share prices and prices of other banks’ CoCos, see Figure 6. Figure 6: Sliding CoCos 51. Too much disclosure can put financial stability at risk. Not helping was that prior to the Deutsche Bank announcement, banks were required to disclose their full capital requirements, including the re￾21Note that the Swedish approach to capital buffers is more flexible than the approach of other European countries, see: Scope Group, 16 February 2016: AT1 risk repricing in Europe - driven by perception of higher coupon risk. Any yet, investors became nervous. 22See the link to the announcement here. 23The Deutsche Bank case of February 2016 February is well-documented. See for example Hale and McCrum (2016), Cotterill (2016). 16

c Martien Lubberink (2019) Comments on RBNZ’s Capital Review Paper 4. quirements that are agreed between a bank and supervisors by way of the supervisory review process (SREP). These SREP capital requirements are known as Pillar 2 capital requirements, they are bank￾specific and cover risks that are not covered by Pillar 1. Unlike in the past and unlike in many countries outside Europe, Pillar 2 requirements are generally not disclosed. They are kept out of the public eye to not affect financial stability. However, in a drive to improve transparency, the EBA had decided that the publication of Pillar 2 requirements was good (EBA, 2015, §20 − §21). With the benefit of hind￾sight, however, EU bank regulators realised that publication of the full capital requirements, i.e. Pillar 1 and Pillar 2 didn’t make sense. Shortly after the events of February 2016, European regulators started changing their approach to Pillar 2 requirements. They realised that a similar incident during less san￾guine times could have much more serious effects on markets and could potentially prompt depositors to run. Note also that the incident was about skipping a coupon payment, which is fully legitimate un￾der Basel III rules. Who would have thought that a relatively minor event would roil the markets so dramatically? 52. Pillar 2 Guidance - a buffer that requires supervisory attention. From 2016 on, the ECB splits Pillar 2 capital into two parts, one called P2G (for ‘Guidance’), the other called P2R (for ‘Require￾ments’).24 The Pillar 2 Guidance part is a “soft” buffer. A bank that dips into this buffer will prompt the supervisor’s attention, but actions will be focused on increasing capital in a discrete and bespoke way. Investors and depositors will not know. This should prevent a repeat of the Deutsche Bank incident. A failure to meet the Pillar 2 Guidance part will not result in automatic action of the supervisor and it will not be used to determine any restrictions on earnings distribution. Figure 7 shows the change. The SREP 2016 approach excludes Pillar 2 Guidance from the calculation restrictions on earnings distributions (the MDA), it is the part above the red line: Figure 7: SREP 2015-2016 24See the 2016 SREP methodology booklet here and the 2015 SREP methodology booklet here. 17

c Martien Lubberink (2019) Comments on RBNZ’s Capital Review Paper 4. 3 Minor Comments 53. The Reserve Bank proposals pay little attention to Operational Risk and the Asset Quality of banks. To illustrate, the RBNZ proposal (RBNZ, 2019) mentions “operational risk” twice. The word￾pair “asset quality” does not feature in the RBNZ proposals or the documents of the information release of 25 January 2019. 54. The Reserve Bank appears to move away from policies that adequately deal with operational risks. Despite announcing an exposure draft for Operational Risk in 2020Q1, for now I rely on the recent joint FMA-RBNZ study into bank conduct and culture. That report could have recommended that banks that engage in misconduct should meet higher capital requirements. This could be achieved with a Pillar 2 capital add-on for Operational Risk. However, the study only features a hand-waving remark about the difficulty of quantifying conduct and culture risk (FMA and RBNZ, 2018, page 6). Perhaps. But then again, the Basel Committee offers the tools to do just that. The document “Basel III: Finalising post-crisis reforms” explains how to determine the minimum capital requirements for operational risk (BCBS, 2017c,b). Moreover, the Reserve Bank’s response to the consultation on the calculation of risk weighted assets shows that it wants to follow the Basel Committee: “All banks will calculate the RWA arising from operational risk in the same way, using the Basel Standardised Measurement Approach.” (RBNZ, 2018). This is tricky, because it prevents responding to operational risk incidents in a timely fashion – bearing in mind that the Basel Standardised Measurement Approach relies on past data, not current or anticipated developments. 55. The lack of attention to asset quality is a worry, as a poor quality will undermine bank capital levels. Lots of examples on this, so I won’t dwell too long. Nevertheless, a case in point is the Spanish bank Banco Popular, it’s demise being the result of poor asset quality.25 25See documentation on the role of the supervisor in Popular’s demise here. 18

c Martien Lubberink (2019) Comments on RBNZ’s Capital Review Paper 4. 4 Answers to specific questions Regarding my views and your questions:

1. “We welcome feedback on our proposed calibration of the capital framework and any related issues raised in this paper” and “We welcome feedback on the risk appetite framework outlined in the consultation paper.” For the calibration the Reserve Bank relies on a pre-GFC Basel Committee document (BCBS, 2005).26 The document is an explanatory note on credit risk modelling using the Internal Ratings￾Based (IRB) approach. The irony! But seriously, the RBNZ calibration is about credit risk. Regarding other risks the Reserve Bank makes only a few assumptions, it does not model them. “We adopted the assumption that the current quantum of capital held against operational and market risks is appropriate going forward, and thus focused, in the quantitative modelling, solely on credit-related losses.” This is a serious caveat: how sure is the Reserve Bank that a pre-GFC explanatory note is still relevant today?
2. “We seek specific feedback on the proposed Tier 1 capital requirement of 15 to 16% (which includes the proposed prudential capital buffer of 9 to 10%).” See my discussion in Subsection 2.2. My suggestion would be to align with the Basel III definition of own funds, with designated roles for CET1, Additional Tier 1, and Tier 2 capital.
3. “We seek your views on the proposed additional capital requirement for banks that may be deemed ‘systemically important’ in New Zealand (D-SIBs).” Such additional capital requirements should be part of a comprehensive approach to bank recovery and resolution. The RBNZ could take a leaf from the European Bank Recovery and Resolution Directive (EC, 2014, BRRD). It offers the tools to deal with troubled banks. Moreover, the EU approach to gone concern loss-absorbing capital is bank-specific. MREL is a Pillar 2 instrument, whereas FSB’s TLAC applies to systemically important banks without differentiation, it is a Pillar 1 instrument (FSB, 2015). These policy tools should also be accompanied by proper evidence for their need.
4. “We seek your view on the proposed Countercyclical Capital Buffer (CCyB).” The CCyB is one of the least successful policy initiatives on bank capital. It’s hardly in use, and if it is used, then it’s watered down by reciprocity requirements. See the Figure 8 below as well as my blog post Are UKs countercyclical capital buffers high? I also wonder how the rating agencies will respond if a CCyB requirement decreases during a crisis. It is not obvious that they will not downgrade banks.
5. “We seek specific feedback on the role of Tier 2 capital.” My suggestion would be to align with the basic Basel III definition of own funds. Perhaps comple￾ment Tier 2 with Senior Non-Preferred (SNP), which is less costly than Tier 2 capital and achieves the same goal: loss absorption in gone concern. See Figure 9 below. 26See the Background paper of 3 April here. 19

c Martien Lubberink (2019) Comments on RBNZ’s Capital Review Paper 4. 6. “We seek specific feedback on the proposed consequences for banks (e.g. dividend restrictions) that breach the prudential capital buffer and the concept of an ‘Escalating Supervisory Response.” See my discussion in Subsection 2.2 and Subsection 2.3. My suggestion would be to deploy a comprehensive Pillar 2 framework. This gives a supervisor more discretion to meet public expectations (i.e to actively supervise). 7. “We seek specific feedback on the leverage ratio.” See my comments on Haldane (2012) on page 4 of this submission. I am sceptical about the use of the leverage ratio. It does not really adjust for risk. The Leverage Ratio may work under U.S. accounting rules. I am unsure if the Leverage Ratio would work in an IFRS environment. 8. “We seek specific feedback on the output floor and recalibration of the IRB approach.” The RBNZ floor plans are at odds with the second of its six high-level principles: “Capital require￾ments should be set in relation to the risk of bank exposures.” The idea of a level-playing field (for whom?) seems politically motivated. Moreover, using floors for competition policy purposes appears to be beyond RBNZ’s mandate. (a) CCyB compared against Common Equity Tier 1. (b) CCyB compared against other buffers. Figure 8: The use of Countercyclical Capital Buffer, bring a microscope. Figure 9: Parameters of Tier 1, Tier 2 and Senior Non-Preferred (SNP). Source: Societ´ e G´ en´ erale. ´ 20

c Martien Lubberink (2019) Comments on RBNZ’s Capital Review Paper 4. 5 Boxed materials Box 1: How the choice of studies supported RBNZ’s decision to limit the use of IRB models. This box shows how the RBNZ used a selection of studies to support policy choices made regarding the use of internal models. For example, the consultation paper (RBNZ, 2017c) refers to a study that covered a sample of 115 banks from 21 OECD countries, which were approved for adopting the F-IRB or the A-IRB approach during 2007-2010 (Mariathasan and Merrouche, 2014). The study concludes that “the evidence supports [the] hypothesis of risk-weight manipulation under Basel II’s IRB, and that the banks’ profit maximising motive is likely to be among the non-fundamental determinants of risk￾weights identified in EBA (2013).” However, the sample period covers the years in which Basel II had just entered into force. In addition, international supervisors only started looking into the use of internal models from 2011 on.a In this context, studies may be more likely to find instances of hard-to-explain use of internal models: Supervisors may have not been fully trained in the use of internal models or supervisors had other priorities than monitoring the use of internal models. A second limitation of Mariathasan and Merrouche is the choice of sample countries. The study relies on OECD countries, however, corporate governance practices differ considerably across countries (La Porta et al., 2000), and therefore some of the results of Mariathasan and Merrouche may be biased against the use of internal models. More examples of research that uses potentially non-representative data: a working paper by (Cizel et al., 2017) uses data from the years 2007–2012. Plosser and Santos (2014) uses data from 2010–2013. There is evidence, however, that offers a more sanguine view on the use of internal models. A recent study of the European Banking Authority on the consistency of internal model outcomes shows that matters regarding the use of internal models are improving (EBA, 2019). The EBA study shows that the majority of risk-weights variability is explained by fundamentals. Moreover, according to the EBA, monitoring activities by European supervisors are . . . “increasingly noticing issues identified by the EBA benchmarking exercise. The same con￾clusion holds for institutions’ internal validations. This is reassuring and indicates that the increased regulatory and supervisory attention paid to internal models is contributing to the consistency of the RWA of internal models.” The EBA likely refers to the ECBs TRIM initiative, the Targeted Review of Internal Models, a project the ECB started in December 2015 to assess whether the internal models currently used by banks comply with regulatory requirements, and whether their results are reliable and comparable.b Academic evidence also shows benefits of the internal ratings-based approach. Cucinelli et al. (2018) find “that IRB banks were able to curb the increase in credit risk driven by the macroeconomic slowdown better than banks under the standardized approach. This suggests that the introduction of the internal ratings based approach by Basel II has promoted the adoption of stronger risk management practices among banks, as meant by the regulators.” aSee “Annex 1: List of existing studies” in BCBS (2013): the majority of the studies are from 2011 or 2012. bSee the link to ECBs TRIM here. 21

c Martien Lubberink (2019) Comments on RBNZ’s Capital Review Paper 4. Box 2: The ratios proposed by the RBNZ are higher than the current requirements, which are 4.5 percent CET1, 1.5 percent Additional Tier 1, and 2 percent Tier 2. With a Conservation Buffer re￾quirement of 2.5 percent (Barker, 2015), the minimum total capital requirements is 10.5 percent. The RBNZ proposal augments these requirements by imposing an additional 5 percent requirement for the Conservation buffer. Large banks will have to meet a one percent D-SIB buffer for systemically im￾portance. The table below compares the ratio requirements with those of Basel II and Basel III. The last row shows the actual ratios at December 2018. It is clear that the RBNZ proposal relies much more than Basel III on Common Equity. For example, there is no room for Additional Tier 1 capital (the requirements for CET1 and Tier 1 are the same: 6 percent). And where Basel III features country and bank specific layers for the Conservation Buffer, the Buffer for Systemically Important banks, and the Countercyclical Buffer, the Reserve Bank relies very much on the Conservation Buffer only. That buffer then should meet a 7.5 percent requirement, which crowds out the buffer space for D-SIB and Countercyclical Buffer. As a consequence of one dominant buffer, the capital framework becomes rigid and it rules out flexibility that a bank may need in times of crisis. Capital Ratios Common Equity (in % of RWA) CET1 Cons. SIB C-Cycl. Tier 1 Tier 2 Total Total Capital Buffer Capital* Capital* Capital* Capital** Basel II 2.0 4.0 4.0 8.0 – Basel III 4.5 2.5 ≤5.0 ≤2.5 6.0 2.0 8.0 18.0 RBNZ proposes 6.0 7.5 1.0 1.5 6.0 2.0 8.0 18.0 Banks meet per December 2018: 11.7 13.9 1.1 15.0 *Excluding buffer requirements, **including buffer requirements The table also shows that New Zealand banks are short of the proposed capital requirements, albeit it that the ratios reported by individual banks vary considerably, see Figure 10. Note that proposed capital requirements do not factor in any current management buffers. Compared to international banks, the reported ratios are at the lower end of the spectrum. The latest Basel III monitoring report shows that Group 2 banks reported a CET1 ratio of 16.0 percent (BCBS, 2018).a But then again, higher risk weights depress the ratios of New Zealand banks by about 300 to 600 basis points. Note also that the Table 1 above excludes any additional loss absorbing requirements, such as TLAC requirements of the Financial Stability Board (FSB, 2015, APRA, 2018). Figure 10: NZ CET1 capital ratios. aGroup 2 banks are smaller than Group 1 banks. These banks are thus more representative for comparisons with New Zealand banks. 22

c Martien Lubberink (2019) Comments on RBNZ’s Capital Review Paper 4. Box 3: The double counting of capital, a prudential no-no. The diagram below helps to illustrate the problem of the double counting of capital. This is best explained by using an insurance subsidiary that is fully owned by a parent bank, though the explanation can be generalised to bank subsidiaries. The diagram shows the balance sheet of a bank that fully owns an insurance subsidiary; the balance sheet of the insurance subsidiary; and the balance sheet of the consolidated group. Looking at the balance sheet of the parent bank it may appear at first that there is 20 of equity supporting banking risks. However, this is not the case. The bank’s equity of 20 is reliant upon the existence its investment of 10 in the equity of the insurance subsidiary. What this means in practice is that any losses suffered by the insurance subsidiary will lead directly to losses in the banks equity. In other words, the 20 of equity on the balance sheet of the parent bank is supporting both the banking risks and the insurance risks. Figure 11: Double Counting of Capital By looking at the consolidated balance sheet of the group it is also clear that true amount of capital supporting both banking and insurance risks is 20. As a consequence, it would be double counting capital to recognise 20 of capital supporting the banking risks, because in reality 10 of this is supporting the risks in the insurance subsidiary. If left unaddressed, the parent bank can infinitely expand the number of subsidiaries, e.g. each subsidiary can invest in a next subsidiary, be it a bank or an insurance company. Such an expansion would be prudentially unsound. With a solvency ratio of 20 110 ≈ 18%, investors in the parent bank will be fooled into thinking that it is well-capitalised, while in fact, the parent bank’s equity of 20 would support an almost infinite amount of assets. To address the double counting problem described above, prudential rules require the banking parent to deduct its investments in its insurance subsidiary. This approach says that while capital is invested in a bank’s insurance subsidiaries, it is exposed to losses in the insurance subsidiaries, and so it should be excluded from the capital considered available to support the losses in the bank. Under this approach, the bank in the example above would be considered to have 10 of capital (= 20 − 10). The insurance subsidiary would also have 10 of capital. Therefore the total capital recognised across both the bank and the insurance subsidiary would be 20, which is equal to the capital in the consolidated group. It should be noted that the consolidated accounts automatically show the correct amount of capital sup￾porting the risks of a group. The double counting problem thus is in particular a problem when a bank’s holding in another financial institution is not consolidated. Basel III rules therefore require the deduction of holdings in the capital of subsidiaries that are not consolidated. An alternative to deduction is to apply a high risk weight to investments in the financial industry, which some prudential regulators require at the parent level. (Source: BCBS Definition of Capital Subgroup meeting notes, 28 October 2011.) 23

c Martien Lubberink (2019) Comments on RBNZ’s Capital Review Paper 4. 6 Bibliography APRA (2018, November). Increasing the loss-absorbing capacity of ADIs to support orderly resolu￾tion. APRA: Commonwealth of Australia. https://www.apra.gov.au/sites/default/files/increasing the loss-absorbing capacity of adis to support orderly resolution.pdf. Barker, F. (2015, June). The Reserve Bank’s application of the Basel III capital requirements for banks. RBNZ Journal 7(5). https://www.rbnz.govt.nz/-/media/ReserveBank/Files/Publications/Bulletins/ 2015/2015may78-5.pdf. BCBS (2005). An Explanatory Note on the Basel II IRB Risk Weight Functions. Basel: Bank for International Settlements. https://www.bis.org/bcbs/irbriskweight.pdf. BCBS (2011). Final elements of the reforms to raise the quality of regulatory capital issued by the Basel Committee. Basel: Bank for International Settlements. www.bis.org/press/p110113.htm. BCBS (2013). Regulatory Consistency Assessment Programme (RCAP) Analysis of risk-weighted assets for credit risk in the banking book. Basel: Bank for International Settlements. https://www.bis.org/ publ/bcbs256.pdf. BCBS (2017a). Basel III definition of capital - Frequently asked questions. Basel: Bank for International Settlements. www.bis.org/bcbs/publ/d417.pdf. BCBS (2017b). Basel III: Finalising post-crisis reforms. Basel: Bank for International Settlements. https://www.bis.org/bcbs/publ/d424.pdf. BCBS (2017c). High-level summary of Basel III reforms. Basel: Bank for International Settlements. https://www.bis.org/bcbs/publ/d424 hlsummary.pdf. BCBS (2018). Basel III Monitoring Report. Basel: Bank for International Settlements. https://www.bis. org/bcbs/publ/d449.pdf. Berndt, A., D. Duffie, and Y. Zhu (2019, March). The Decline of Too Big to Fail. Stanford: Australian National University. https://www.darrellduffie.com/uploads/pubs/BerndtDuffieZhuTBTF.pdf. Bevilacqua, M., F. Cannata, S. Cardarelli, R. A. Cristiano, S. Gallina, and M. Petronzi (2019). The evolution of the Pillar 2 framework for banks: some thoughts after the financial crisis. Rome: Banca d’Italia. http://www.bancaditalia.it/pubblicazioni/qef/2019-0494/index.html. Calomiris, C. W. and S. H. Haber (2014). Fragile by Design: The Political Origins of Banking Crises and Scarce Credit. 41 William Street Princeton, New Jersey 08540 USA: Princeton University Press. Cizel, J., H. A. Rijken, E. I. Altman, and P. Wierts (2017, March). Assessing Basel III Capital Ratios: Do Risk Weights Matter? Tilburg: Tinbergen Institute. http://pages.stern.nyu.edu/∼ealtman/Assessing% 20Basel%20III%20Capital%20Ratios.pdf. Cotterill, J. (2016). The CoCo that popped. Financial Times. February 10, 2016. Cucinelli, D., M. L. D. Battista, M. Marchese, and L. Nieri (2018). Credit risk in European banks: The bright side of the internal ratings based approach. Journal of Banking & Finance 93, 213–229. Dautovic¸, E. (2019, February). Has Regulatory Capital Made Banks Safer? Skin in the Game vs Moral Hazard. SSRN eLibrary (ID 3332501), 1–62. 24

c Martien Lubberink (2019) Comments on RBNZ’s Capital Review Paper 4. DeAngelo, H. and R. M. Stulz (2015). Liquid-claim production, risk management, and bank capital structure: Why high leverage is optimal for banks. Journal of Financial Economics 116(2), 219–236. Dombret, A. (2017, August). Failing or likely to fail? Putting the European banking union to the test. EBA (2013). EBA interim report on the consistency of risk-weighted assets in the banking book. London: European Banking Authority. EBA (2015, December). Opinion of the European Banking Authority on the interaction of Pillar 1, Pillar 2 and combined buffer requirements and restrictions on distributions. London: European Banking Authority. EBA (2019). EBA releases its annual assessment of the consistency of internal model outcomes. London: European Banking Authority. EC (2014). Directive 2014/59/EU of the European Parliament and of the Council of 15 May 2014 establishing a framework for the recovery and resolution of credit institutions and investment firms and amending Council Directive 82/891/EEC, and Directives 2001/24/EC, 2002/47/EC, 2004/25/EC, 2005/56/EC, 2007/36/EC, 2011/35/EU, 2012/30/EU and 2013/36/EU, and Regu lations (EU) No 1093/2010 and (EU) No 648/2012, of the European Parliament and of the Council. Brussels: Eu￾ropean Commission. EC (2017, June). European Commission - Fact Sheet State aid: How the EU rules apply to banks with a capital shortfall - Factsheet. Brussels: European Commission. http://europa.eu/rapid/press-release MEMO-17-1792 en.htm. FMA and RBNZ (2018, November). Bank Conduct and Culture. Wellington: Financial Markets Au￾thority and the Reserve Bank of New Zealand. https://www.rbnz.govt.nz/-/media/ReserveBank/Files/ regulation-and-supervision/banks/Banks-Conduct-and-Culture-Review.pdf. FSB (2015, November). FSB issues final Total Loss-Absorbing Capacity standard for global systemically important banks. Basel: Financial Stability Board. http://www.fsb.org/2015/11/ tlac-press-release/. Glover, J. (2014). Bank of America Compiling Index for CoCo Bond Performance. Bloomberg. January 10, 2014. Glover, J. and T. Beardsworth (2019). Contingent Convertibles. Bloomberg. February 19, 2019. Gropp, R., T. C. Mosk, S. Ongena, and C. Wix (2016, March). Bank Response to Higher Capital Requirements: Evidence from a Quasi-Natural Experiment. SSRN eLibrary (ID 2877839), 1–82. Haldane, A. G. (2012). The dog and the frisbee. Basel: Bank for International Settlements. https: //www.bis.org/review/r120905a.pdf. Hale, T. and D. McCrum (2016). Deutsche’s riskiest bonds in spotlight. Financial Times. November 11, 2015. Jorda,O., B. Richter, M. Schularick, and A. M. Taylor (2017, mar). Bank capital redux: Solvency, liquidity, and crisis. NBER Working Paper Series. La Porta, R., F. L. de Silanes, A. Shleifer, and R. Vishny (2000). Investor protection and corporate governance. Journal of Financial Economics 58(1), 3–27. 25

c Martien Lubberink (2019) Comments on RBNZ’s Capital Review Paper 4. Lubberink, M. J. P. and A. Renders (2018). Are Banks’ Below-Par Own Debt Repurchases a Cause For Prudential Concern? Journal of Accounting, Auditing & Finance 33(1), 1–29. Mariathasan, M. and O. Merrouche (2014, July). The manipulation of basel risk-weights. Journal of Financial Intermediation 23(3), 300–321. Melis, A. and K. Weissenberg (2019). EU Banks’ Journey Towards an Enhanced Capital Framework. London: European Banking Authority. https://eba.europa.eu/documents/10180/2259345/EU+banks+ journey+towards+an+enhanced+capital+framework++-+March+2019.pdf. Micossi, S. (2019). Testing the EU Framework for the Recovery and Resolution of Banks: the Italian Experience. Rome: LUISS School of European Political Economy. https://sep.luiss.it/brief/2019/02/ 15/s-micossi-testing-eu-framework-recovery-and-resolution-banks-italian-experience. Mortlock, G. (2019a, April). Geof Mortlock argues that independent professionals should be brought in to undertake a comprehensive assessment of the Reserve Bank proposals, plus alternative options. interest.co.nz. April 23, 2019. Mortlock, G. (2019b, May). Geof Mortlock argues that New Zealand should follow Australia’s example and have periodic independent capability reviews of our financial regulators. interest.co.nz. May 19, 2019. Mortlock, G. (2019c, February). Geof Mortlock makes the case for a shake-up to RBNZ governance and a fundamental, independent review of the regulatory framework for NZ’s financial system. inter￾est.co.nz. February 12, 2019. Nouy, D. (2017). ECB’s Nouy defends handling of Banco Popular resolution. Financial Times. June 20, 2017. Plosser, M. C. and J. A. C. Santos (2014, December). Banks’ Incentives and the Quality of Internal Risk Models. SSRN eLibrary (ID 2543535), 1–42. RBNZ (2017a, December). Capital Review Paper 2 (Part II): What should qualify as bank capital? Response to submissions. Wellington: Reserve Bank of New Zealand. https://www.rbnz.govt.nz/-/media/ReserveBank/Files/Publications/Policy-development/Banks/ Review-capital-adequacy-framework-for-registered-banks/Capital%20Review%20Response% 20to%20Numerator%20Submissions191217.pdf. RBNZ (2017b, July). Capital Review Paper 2: What should qualify as bank capital? Issues and Options. Wellington: Reserve Bank of New Zealand. https://www.rbnz.govt.nz/-/media/ReserveBank/Files/ Publications/Policy-development/Banks/Review-capital-adequacy-framework-for-registered-banks/ Capital-review-paper-what-should-qualify-as-capital.pdf. RBNZ (2017c, December). Consultation Paper: Review of the Capital Adequacy Framework for locally incorporated banks: calculation of risk weighted assets. Wellington: Reserve Bank of New Zealand. https://www.rbnz.govt.nz/-/media/ReserveBank/Files/Publications/Policy-development/Banks/ Review-capital-adequacy-framework-for-registered-banks/Capital%20Review%20Denominator% 20Consultation%20Paper%20002191217.pdf. RBNZ (2017d, May). Issues Paper: Review of the Capital Adequacy Framework for locally incorporated banks. Wellington: Reserve Bank of New Zealand. https: //www.rbnz.govt.nz/-/media/ReserveBank/Files/Publications/Policy-development/Banks/ Review-capital-adequacy-framework-for-registered-banks/capital-review-issues-paper-may2017. pdf. 26

c Martien Lubberink (2019) Comments on RBNZ’s Capital Review Paper 4. RBNZ (2018, July). The calculation of risk weighted assets - Response to submissions. Wellington: Reserve Bank of New Zealand. https://www.rbnz.govt.nz/-/media/ReserveBank/Files/Publications/ Policy-development/Banks/Review-capital-adequacy-framework-for-registered-banks/Responses/ march/RBNZ-response-to-march-submissions-capital-adequacy-review.pdf. RBNZ (2019, January). Capital Review Paper 4: How much capital is enough? Welling￾ton: Reserve Bank of New Zealand. https://www.rbnz.govt.nz/-/media/ReserveBank/Files/ Publications/Policy-development/Banks/Review-capital-adequacy-framework-for-registered-banks/ Capital-Review-Consultation-How-much-capital-is-enough.pdf. 27

From: Martin Taylor To: Capital Review Subject: Submission: Reserve Bank Review of the capital adequacy framework for registered banks Date: Monday, 22 April 2019 5:52:33 PM Attachments: Submission_ RB capital adequacy review.pdf Hi Please find attached my submission. My contact details, not for publication with the submission, are below. I would be happy to make an oral submission if the opportunity is available. Best regards Martin Taylor OIA s9(2)(a)

Submission To: Reserve Bank Review of the capital adequacy framework for registered banks Submitted by: Martin Taylor Date: 21 April 2019 I am submitting as a member of the public. My submission is IN FAVOUR of significantly strengthening and enlarging bank capital for systemically-important banks. Summary of Key Points For whom are we regulating? The case for greater capital The moral and economic case Is the proposed capital increase enough? Addressing claims of higher interest rate costs A small ‘premium’ is OK Shareholders will not abandon the banks Bring on the competition Addressing claims of credit slowdowns and slower growth Lower housing inflation Encourage business lending An expanded fiscal role The public and bankers: Misaligned risk appetites References Summary of Key Points

1. The Reserve Bank must put the interests of the economy and wider societal well-being at the core of this decision.

2. A substantial boost to capital is justified to reduce damage from bank failure to the economy and ordinary Kiwis. Much tighter regulation would be an alternative. Inaction should not be an option.

3. There is a strong moral, as well as economic case, for banks to contribute a greater capital cushion. New Zealand’s main banking vulnerabilities stem from high land prices and associated debt that has been inflated by banks’ expansionary lending policies.

4. Given the increasing frequency and severity of major financial failures, the 1/200 year risk benchmark should be strengthened rather than watered down.

5. Claimed adverse impacts and costs, such as interest rate hikes and credit squeezes, are certainly overstated and can be mitigated with net public benefits. An (unlikely) small cost to the economy would still be justified due to the ‘insurance’ benefit.

6. If the big banks choose together to raise rates or tighten credit, this will open a window for greater competition and market share by smaller banks with a net gain to the economy and reduction in ‘too-big-to-fail’ risk.

7. Fiscal interventions should also be planned as part of the mix to preempt economic impacts from any reduction in bank credit access. Such interventions themselves can be designed to produce a net public gain.

8. The stability gains from increased capital should be boosted further by better aligning the risk appetites of the public and bankers. This means that more of the direct consequences of failure should fall on banks and senior bankers. For whom are we regulating? I’d like to open this submission with the Reserve Bank’s own words. While the Reserve Bank’s explicit mandate does not make reference to societal well-being, the social impacts of crises are a relevant consideration for society and society’s attitudes towards the risk of harm from crises that we seek to represent in our capital policy. The idea that economic policies (and by association, financial regulations) should be set after considering outcomes in the broadest sense is not novel. (p19 RB Capital Review Background Paper, April 2019) It’s certainly ‘not novel’ to ordinary Kiwis who would wonder in whose interests the financial system was being run, if not the interests of the economy and society. What would surprise ordinary Kiwis is that such a statement should need defending by the Reserve Bank to the banking industry. But it must since “the Reserve Bank’s explicit mandate does not make reference to societal well-being…” Alas, for decades now, prudential policy has focused on protecting banks from themselves rather than protecting the economy and wider society from banks. Let’s indeed put societal well-being at the centre of this discussion — and use it as a step to shift all of the Reserve Bank’s prudential and monetary functions to serve this end. The case for greater capital The security that would come from greater bank capital is unequivocally a public benefit. Simply, the banking system will be safer and the consequences of catastrophic failure will be

greatly reduced. There are few alternatives — short of much tighter regulation — that will serve this purpose better. The moral and economic case There is a moral as well as economic case for banks to provide a substantial capital buffer. The biggest risk to New Zealandʻs banking system stems from our vulnerability to interest rate and asset price changes associated with residential and farm land. As this chart shows, it has been banks’ deliberate choice to dramatically expand credit into land-based sectors — credit could have been more constrained and Kiwis would still have bought and sold land to one another, albeit at lower prices. This sustained, four-fold expansion cannot be explained by a sudden population explosion or loss of land, nor does it reflect building cost increases. Having deliberately inflated the bubble, banks must now inflate the cushion to protect us from it. Is the proposed capital increase enough? The boom/bust cycles that drive most bank failures are increasing globally in frequency and severity. Because of this, setting the proposed 1/200 year benchmark by looking back means it is probably too low and should be increased. This can perhaps be done with a second tranche, once banks have absorbed this proposed increase.

On the costs, any ‘harm’ this might do to the largest banks is minor, given that they already enjoy super-profits and healthy balance sheets. And most, if not all, of the costs banks might choose to impose on the wider economy and society can be mitigated, as shown below. Addressing claims of higher interest rate costs A small ‘premium’ is OK A small increase in interest rates, if it happens, should not deter the RBNZ from imposing greater capital requirements. It will be a small ‘premium’ to insure against disaster. Shareholders will not abandon the banks The profitability of the major NZ banks already exceeds most others, so even with higher capital cost and similar interest spreads, returns will remain attractive relative to alternative investments. Those returns are boosted, and to a large extent protected, by the privileged legislative position these banks enjoy. That won’t change. Bring on the competition The current super-profits could be competed away without undue hardship to the big banks’ balance sheets or shareholders. If the market is competitive, one or two of the big banks should see competitors’ rate hikes as an opportunity to boost market share. And if the big banks, acting in ‘cartel’ fashion, should all decide to raise rates this would give smaller banks like Kiwibank, cooperative banks, etc excellent opportunities to boost their market share by a few points. It should also give the RBNZ a clear signal that competition isn’t working well. Given that this big-bank action would be a conscious preference, not a necessity for survival, competiton from smaller players could be assisted with some regulatory help and/or government support (such as a capital boost to Kiwibank and moving government banking to it.) Overall, this would be a very positive outcome for New Zealand — we need a bigger slice of the market to go to smaller, local banks and we should be doing everything possible to advantage them. We can see this from the overseas success of co-operative models in areas such as housing and local finance, or Germany’s diverse banking industry with 1500 banks providing the financial backbone for its powerhouse small and mid-sized business sector. A more diverse banking sector would make us less reliant on ‘too big to fail’ banks and reduce potential damage from a large-scale bank failure.

Addressing claims of credit slowdowns and slower growth A slow-down of lending by the big four banks doesn’t mean the entire market must fall by this amount. Like jacked-up interest rates, it opens space for smaller competitors to fill the gap which should be welcomed and encouraged. Lower housing inflation Most lending now goes to housing. There’s a strong case to reduce the share of bank lending that goes to housing since years of excessive lending to this sector have helped drive up land prices. So a slow-down here will have positive economic and societal benefits as the finance-driven portion of housing inflation moderates. A recent RBNZ paper calculated that the ‘wealth effect’ from housing inflation adds real consumer spending activity of about 2.2% of the gain. So $100,000 of extra housing ‘wealth’ would produce about $2200 of economic activity. This is inefficient when compared to more direct economic drivers such as wage growth and government spending — and it’s not cost-free. This small ‘gain’ should be reduced by the economic and social costs it imposes on housing affordability. The same level of economic activity could be added by a small injection of direct government spending without the terrible collateral damage from inflating land prices. Houses will still be bought and sold. Encouraging the wealth effect as a way to boost economic growth should draw the same opprobrium as incentivising polluters to pollute more. For the same reason, we should welcome, not fear, a voluntary restraint of housing credit growth by bankers. Encourage business lending If credit expansion slows (it’s run well above inflation and GDP growth for years due to household, not business lending), it doesn’t necessarily have to impact the productive sector. We can plug gaps through active policy steps. Banks could be incentivised to lend more to business by easing the capital requirements and rules for business lending vs housing. With greater capital buffers there is room for such an adjustment. Another tool is a more active industrial policy backed by direct grants or loans from a development bank. This approach is returning to favour around the world, for good reason.

An expanded fiscal role Let’s address the (unlikely) scenario of a significant credit slowdown with a well-targeted and efficient fiscal spending boost and enjoy the dual benefits of increased public infrastructure and lower housing inflation. This would be a good time to add ‘electronic’ money to the RBNZ toolkit, allowing the RBNZ to step in, under prescribed circumstances, with a direct injection of new money via suitable fiscal interventions. A petition on a similar approach is before parliament now from the Positive Money organisation and warrants careful, open-minded consideration by the RBNZ and politicians. Again, the claimed ‘costs’ of bank capital increases can be turned into public gains. The public and bankers: Misaligned risk appetites Let’s be clear: the public has no appetite whatsoever for major bank failure, nor should the RBNZ as its guardian. Unfortunately, the same cannot be said of bankers. The public knows that ordinary people, businesses and public services will be devastated by major bank failure. But bankers have proven that they underrate both the risk, and the consequences, of failure as they repeatedly get caught up inside overheated, runaway systems whose rules bend to accommodate their own greed and their customers’ FOMO (fear of missing out). So in considering risk appetite, it is important that measures are put in place with the systemically-important banks to align bankers’ risk appetites with the public’s. Risk remains conspicuously misaligned as risky behaviour for bank employees has been incentivised, consequences largely neutralised, and shareholders can confidently bet on getting government bailouts on easy terms. To improve alignment:

1. Senior executives should know that the financial consequences for failure will include substantial personal financial cost, including possible bankruptcy. This could be accomplished by requiring full repayment of all recent bonuses and salaries in excess of five times the bank’s average pay.

2. Shareholders must have certainty that a bank failure will cost them their entire stake. Any ambiguity or ‘out clause’ will leave the current moral hazard in place. A government bailout should result in full nationalisation without compensation if the bank’s capital does not cover the costs of its failure. This can be justified on commercial as well as wider societal grounds since, when government must be the lender of last resort, the business has clearly lost all commercial value. Politically, an angry public will be unlikely to support a cost-free bailout of shareholders, especially when depositors under OBR will be paying.

3. Directors must face significant financial burdens in line with senior executives, and should face jail rather than fines if negligence can be established. Combined with the threat of nationalisation, it might also encourage earlier intervention and improve the chances of rescuing a distressed bank. If these proposed measures seem draconian when compared to other industries, they should be. The impact of a major bank failure is exponentially beyond any other enterprise or sector’s failure. This means senior bankers and bank investors have a special position of trust to uphold. They are compensated by the special privileges they receive and should have standards of behaviour and judgement to match this. References Housing Leverage and Consumption Expenditure by Karam Shaar and Fang Yao, Reserve Bank Discussion Paper, April 2018 New Zealand house prices: a historical perspective by Elizabeth Kendall, Reserve Bank Bulletin, Vol. 79. No.1. January 2016 Positive Money NZ, Petition to Require Reserve Bank of New Zealand to Issue All New Zealand Money

Matt Colvin I agree the amount of capital banks need should be increased. Banks should be about security and safety first and not about risky leveraged investments for the benefit of shareholders. It will be the taxpayer who reaps little of the benefit when things are going well but much of the burden when things inevitably go badly. OIA s9(2)(a)

Mel I am writing to express concern around the impact of restricted capital for banks around lending for alternative development models. We are part of an emerging sector that falls between investment based developments and social housing by charitable trusts. These models include co-housing, baugruppen, cooperative style projects. We would like to see exceptions made that allowed these style of developments receive special lending conditions. These sorts of developments are often fully subscribed and have little to no profit margins in them. It would be fantastic to follow the lead of wellbeing budget and promote lending to housing that has superior wellbeing outcomes for people and the planet. Ideally what is considered commercial financing should come in as residential financing as these are resident led developments. Primarily they provide owner occupier accomodation provide a better level of energy efficiency (passive house) and create better social connectivity. Please consider if there is room for RBNZ to encourage these sorts of housing models in New Zealand. Kind Regards Mel Halliday OIA s9(2)(a) OIA s9(2)(a)

Submission to Reserve Bank of New Zealand on Capital Review Paper 4: How much capital is enough? by Michael Reddell 15 May 2019 Introduction and summary The Reserve Bank’s December 2018 consultative document proposed three main changes: • Much higher minimum ratios of capital (CET1) to risk-weighted assets than previously, • Higher minimum capital ratios for systemically-significant banks than for other locally-incorporated banks, and • A significant narrowing in the gap between the calculation of risk-weighted assets as between the big banks using internal models and the remaining locally-incorporated banks using the standardised approach. This submission outlines my response to these proposals, with a focus on the first of them. Most of the points are developed in greater detail in a succession of posts over recent months at my Croaking Cassandra blog (https://croakingcassandra.com/category/bank￾capital-requirements/) which should be read as part of the submission. I also associate myself, to a considerable extent, with the views and analysis contained in two papers on this issue by Ian Harrison (http://www.tailrisk.co.nz/documents/HowMuchCapitalIsEnough.pdf and http://www.tailrisk.co.nz/documents/RBNZ NewAnalysis May.pdf ). In summary: • I support the proposal to increase the minimum risk-weighted assets calculation to around 90 per cent of the amount that would be generated, for the same set of assets, using the standardised approach, • The 90 per cent floor should be regularly highlighted by the Reserve Bank, both on its dashboard, and in speeches or reports, to assist in the cross-country comparability of bank capital ratios. • The case has not been compellingly made for materially higher minimum core capital ratios (in particular, not on top of the additional capital that the larger banks would probably choose to hold as a result of raising the floor on the calculation of risk-weighted assets), • Similarly, no compelling case has been made for imposing a higher minimum capital ratio for systemically significant banks than for small banks, at least if the minimum capital ratios for all banks are to be as high as we proposed by the Reserve Bank in this consultation.

• The failure to:

1. Provide a properly documented cost-benefit analysis, with well-documented sensitivity analysis around the impact of key assumptions,
2. Address in depth the case for much higher capital ratios in light of (a) repeated stress tests results, and (b) the experience of the New Zealand financial system following the credit boom of the 2000s,
3. Provide a proper benchmarking exercise which would have enabled readers and potential submitters to evaluate your proposals relative to APRA’s minimum requirements for Australian banks (and ideally against the requirements of other key overseas regulators),
4. Provide any serious analysis of how the transition might unfold, both as between classes of institutions (those affected by these proposals and those not), between classes of borrowers and/or funders, and of any risks/costs of disintermediation all seriously undermined the credibility of the consultation. Having chosen, apparently by not much more than a stab in the dark, a desired outcome, the attempts to provide substantive insightful supporting analysis seemed half-hearted at best. • Concerns about the potentially disruptive transitional effect are only compounded when one allows for the likelihood that the next significant economic downturn will occur during the next five years, a period during which - on the Bank’s own published projections - there will be quite severely limited conventional monetary policy capacity to respond. THE CASE AGAINST MATERIALLY HIGHER MINIMUM (CET1) CAPITAL RATIOS Relevant context An unbiased observer, looking at the New Zealand economy and financial system, would struggle to find a case for higher minimum capital ratios. Among the factors such an observer might consider would be: • The fact that the New Zealand financial system has not experienced a systemic financial crisis for more than hundred years (and to the extent it approximated one in the late 1980s, that was in the idiosyncratic circumstances of an extensive and fast financial liberalisation which left neither market participants nor regulators particularly well-equipped), • Our major banks - the only ones that might pose any serious economywide risks - come from a country with very much the same historical record as New Zealand, • Despite very rapid credit growth in the years prior to 2008 (increases in the credit to GDP ratios among the larger in the advanced world, spread across housing, farm, and other business/property lending), and a severe recession in 2008/09 and afterwards, the banking system emerged with low loan losses,

• Since then, banks have not only increased their actual capital ratios (and been required to calculate farm risk-weighted assets more stringently) but have also substantially improved their funding and liquidity positions (under some mix of regulatory and market pressure). • Over the decade, bank credit growth (relative to GDP) has been pretty subdued and there has been little or no evidence (in, for example, Reserve Bank FSRs) of any serious degradation of lending standards. • The balance sheets of the large banks remain relatively simple, and there has been no sign (per FSRs) of the sort of financial innovation that might raise significant doubts about the adequacy of existing models. • In terms of the wider policy environment, government fiscal policy remains very strong, we continue to have a freely-floating exchange rate, and there has been neither legislation nor judicial rulings that will have materially impaired the ability of banks to realise collateral. • And the Open Bank Resolution option for bank resolution has been more firmly established in the official toolkit (note that if OBR were fully credible then, in the absence of deposit insurance, there would be little case for regulatory minimum capital requirements at all). • And repeated stress tests - over a period when the regulator had no incentive to skew the tests to show favourable results - suggested that even if exposed to extremely severe adverse macro shocks, and associated large price adjustments for houses, farms, and commercial property, not only would no bank fail, but no bank would even drop below current minimum capital requirements. • Consistent with this experience - also observed in Australia, the home jurisdiction of the parents of our major banks - the major banks operating here continue to have strong credit ratings (consistent with a very low probability of default), and the ratings of the parent banks are even higher. • There has been no change in the ownership structure of our major banks, or in the implied willingness of the Australian authorities to support the (systemically significant) parents of the New Zealand banks were they ever to get into difficulty. Add into the mix indications that New Zealand banks CET1 ratios, if calculated on a properly comparable basis, would already be among the highest in the advanced world - in a macro environment with more scope for stabilisation (floating exchange rate, strong fiscal position, little unhedged foreign currency lending) than in many advanced countries - and there would be a fairly strong prima facie case for leaving things much as they are. But the Reserve Bank’s consultative document - and associated material, including speeches and interviews - engages substantively with almost none of this context. Closing the gap between the internal models and the standardised approach to calculating risk-weighted assets There probably is a good case for the Bank’s proposal to narrow the gap between risk￾weighted asset calculations based on internal models and those calculations if done on a

standardised basis (in fact, there would probably be a reasonable case - at least starting from scratch – to get rid of the internal models provision altogether in calculating regulatory capital requirements). This is so even though what the Bank is proposing goes beyond the standard floor that is part of the Basle framework, and is materially more restrictive than is envisaged by APRA (both points which should have been made explicit in the consultative document.) But on the Bank’s own published numbers, that change alone would have increased the capital required by large banks to support their current business by 15 to 20 per cent (assuming banks would choose to maintain much the same margin over the regulatory minimum ratio as at present). Unfortunately, there is no sign that the Bank has considered this leg separately from the proposal to increase minimum ratios themselves. It is likely - but not certain - that there would be efficiency gains from closing the gap between the two approaches to calculating risk-weighted assets, and in respect of the larger banks it is also likely that there would be some gains in financial stability (reduced probability of failure). It would have been preferable to have analysed the costs and benefits of this proposal first, before moving on to consider whether minimum capital ratios should be raised further. The discipline of doing a proper cost-benefit analysis might have led you to do so. As it is, we are left without any clear differentiation between the benefits and costs of these two, quite separate, strands of what you are proposing (indeed Ian Harrison in his second paper suggests that your belated modelling exercise largely overlooks the impact of closing the gap in the RWA calculation methodology). Little sign that the Bank has thought hard about financial crises The consultation document, and supporting material, shows little sign that the Reserve Bank has thought hard about financial crises in bringing together these proposals. There is plenty of discussion of selected research papers, but nothing that stands back and poses plausibility questions. Thus, there is a strong (implicit) tendency in the document to treat financial crises as exogenous shocks, events arising out of the blue, which a decently-managed bank (or financial system) will face every once in a while, (be it once a century, or two). But a moment’s reflection is all it should take to realise that that is simply the wrong approach to be using (especially when, as in this consultation, you are talking of proposals designed to reduce already-low risks to extremely low levels). You could look at the Irish crisis, the Icelandic one, the US crisis, the Korean crisis of the 1990s, the Nordic crises of the early 1990s (and even the New Zealand and Australian experiences in the late 80s and early 90s) to appreciate that the system-threatening problems didn’t arise from exogenous shocks, but from several years of very degraded lending standards. Exogenous shocks may have played some part in determining the timing and nature of the crystallisation of the problems, but they weren’t what determined that there would be a costly re-adjustment at some point. If the Bank believes differently, the onus should have been on it to make its case. There was no sign of such a case in the consultation document.

Linked to this point, there is very little recognition (none in the main document, and very little in subsequent papers) that many or most of the output losses associated (in time) with financial crises have to do with the misallocation of resources (bad lending, bad borrowing, bad investing) in the preceding boom years. Your documents recognise that one cannot simply measure output losses from a pre-crisis peak (typically a period with a positive output gap) but do not go anywhere near far enough to recognise the significance of this, rather larger, point. In such circumstances, estimates of potential GDP itself may be materially overstated. As far as I can tell, the research papers you quote are open to the same criticism (which is not a defence for the Bank, but - probably – an indication of the predispositions of many of the chosen researchers and their institutional sponsors). When an economy and financial system has gone through several years of badly misdirected lending, borrowing, and investment, not only is there an inevitability about output losses because of the bad prior choices crystallising, but there is a near-inevitability about both lenders and borrowers being hesitant about doing new business in the wake of the realisation of past mistakes. Prior assumptions and business models prove invalid, and it takes time for risk appetite to revive, and to identify like projects that would prove profitable. That is likely to be so whether or not banks emerge from the crystallisation phase with ample levels of capital. At best, it is only the marginal additional output losses from banks falling into “crisis” (however defined) that is likely to be eased by much higher initial capital ratios - and yet you made no attempt to distinguish this effect. The Bank also showed no sign of having done any sort of comparative analysis (of that sort done previously on my blog e.g. here https://croakingcassandra.com/2017/07/06/reserve￾bank-dtis-and-the-cost-of-crises/, or here https://croakingcassandra.com/2019/03/04/banking-crises-are-bolts-from-the-blue/ or by PIIE’s William Cline) comparing the output and/or productivity experiences of countries that underwent financial crises with those that did not. This is particularly important in thinking through the experience around 2008/09, when many countries experienced crises and many others did not, all overlaid on what appears to have been a common global productivity growth slowdown. Reasonable people might differ as to how best to do such an adjustment or assessment, but the Bank shows no sign of having even tried. Any plausible assessment of this sort would, however, conclude that plausible additional output losses saved by reducing the probability of any particular loan book incurring losses large enough to run through capital would be much lower than the estimates the Bank uses. Note also that the Cline methodology still overstates the amount that higher capital ratios alone might save, since his output path comparisons include (for the crisis countries) both kinds of losses

• from the initial misallocation of resources, and the pure crises effects. Only the latter should be relevant in assessing the costs and benefits of higher minimum capital ratios. I don’t propose to spend any material time on the temporary vs permanent distinction, but again it should involve no more than a moment’s introspection to realise the implausibility of supposing that a financial crisis 100 years ago (and, as a reminder, the US had many such crises) is still affecting the level of output in the crisis country today. Pure financial crisis effects on output are typically temporary, and are much smaller than the Bank allows.

The Bank also shows little sign of having given sufficient weight to the fact that New Zealand has a floating exchange rate. There seems little doubt that a fixed exchange rate regime - especially when it involves a small country pegging to a currency the economy of which it is not well-aligned with - can exacerbate booms and complicate the management of busts, and the speed with which output resumes a normal growth path. Ireland and Greece offer two recent examples. All else equal, a fixed exchange rate (or common currency) should probably be accompanied by higher minimum capital ratios than otherwise. But that isn’t the New Zealand system: we have a floating exchange rate, Australia has a floating exchange rate, and neither currency has safe-haven characteristics (thus in economic downturns and risk-off events the exchange rate tends to fall readily and substantially). The importance of this point to thinking about financial stability and economic adjustment in New Zealand was highlighted several years ago in a Bulletin article by Hargreaves and Watson. More specifically, it is not clear that the Bank can cite a single example of a systemic financial crisis in an advanced economy that has had liberalised markets for some time, has a floating exchange rate, and where the government has not played a considerable role in steering lending. As is well known, the state plays a very large role in the US housing finance market, in a way not seen in (for example) New Zealand, Australia, or the UK. Social costs In the Deputy Governor’s speech and the April paper, the Bank makes considerable play of the alleged social costs of financial crises. Here I endorse Ian Harrison’s treatment of these, largely specious (particularly in an advanced economy), arguments. I would also note that none of the papers invoked in support of the social cost arguments make any attempt to distinguish between the costs of crises themselves and the costs of the initial misallocation of credit and investment resources. Again, only the former could possibly be relevant to the minimum capital question, even if there were evidence of significant additional social costs in advanced economies with decent social safety nets. Banks with 50 per cent capital ratios may still accommodate demand to finance bad projects, and that bad lending/borrowing/investment will have real economic costs - perhaps even real social costs

• even if no bank ever comes close to failing. Economic cost of the higher capital ratios themselves In his speech in February, the Deputy Governor indicated that the Bank’s own analysis suggested that the output cost of the proposed higher capital ratios would be “up to 0.3 per cent” of the level of GDP. In other words, the annual insurance premium society would pay
• even on your assumptions - might be 0.25 per cent of GDP. As you note, the standard Treasury discount rate is a bit larger than what is used in many of the papers you cite, and applying such a discount rate to this expected annual cost gives a present value of lost output of perhaps $15 billion. That is a high hurdle to get over when the gain on offer is the reduced (from already low levels) probability of output losses resulting (narrowly) from a financial crisis expected in, on average, 75 or 100 years’ time (your claim is that you want to keep the probability of crisis to no more than once in 200 years). On plausible estimates

of those marginal additional output loss savings, the cost-benefit simply would not stack up. (And as Ian Harrison notes, none of these numbers appear to take account of the income loss to New Zealanders from imposing higher capital requirements on - and thus requiring higher expected equity returns to shareholders of - foreign-owned banks.) Linked to this, citizens risk being forced to pay a premium upfront (lost annual output) with no certainty (probably not even a high likelihood) that the policy will be persisted with for long enough to generate any material expected gains at all. Reserve Bank policy decisions are made by a single decisionmaker, who is unlikely to be round for more than 10 years in total. Current plans to review the governance of the Bank could well result in decision￾making being shifted to people with different risk preferences, different assessments of the costs and benefits of the policy etc. Scarcely any policies remain materially unchanged for 50 or 100 years, and this is one that is easily reversible (not like, say, building a flood stopbank that once built stands for decades). A serious assessment of the Bank’s current proposal would put at least a higher weight (than implied simply by the discount rate) on the certain costs in the early decades than on the possibility of some eventual benefits decades hence if (a) the policy is persisted with, and (b) the current sense of the distribution of crises probabilities and costs is correct. The Bank - let alone the current Governor - simply can’t pre-commit. That inability matters in circumstances like these. No benchmarking It is grossly unsatisfactory that throughout months of consultation the Bank has made no effort to illustrate how its proposals for minimum CET1 ratios and the associated floors around the calculation of risk-weighted assets, compare with those planned by APRA for the Australian banks. Such an exercise should have been relatively straightforward, especially if the Reserve Bank had done what most New Zealanders might reasonably have expected, and worked closely together with APRA in formulating its proposals. Of course, New Zealand is a sovereign nation and the Reserve Bank (regrettably) has final decision-making powers in New Zealand but: • APRA has a considerably deeper pool of expertise, including at the top of the organisation, than the Reserve Bank of New Zealand, • The nature of the risks in the two economies and markets is quite similar (including similar legal institutions, and similar housing markets), • If anything there is a case for thinking that APRA minima would be ceilings below which New Zealand requirements for our large banks should be set (since we have the benefit of strong parent banks, and well-regarded supervisor of those banks, whereas the parents - and parents’ supervisors - themselves are on their own, and we have also chosen to have the OBR as a frontline resolution option), • For the institutions that might pose potential systemic issues in New Zealand, any substantial increase in capital requirements can reasonably be seen as an attempt to grab group capital for New Zealand. Why not work these things out together?

The onus should, surely, be on the Reserve Bank of New Zealand to demonstrate - make the case in detail - why the New Zealand subsidiaries of Australian banks should be subject to more onerous capital requirements than the parents, and banking groups as a whole, are subject to. But not once has the Reserve Bank attempted to make that case. The Bank has also been deficient in not engaging in any analysis to show why it is necessary or appropriate to impose higher CET1 capital ratios on large banks than we observe for (typical) financial intermediaries where there is little or no likelihood of bailout and no deposit insurance. That, surely, is closer to the relevant test than handwaving about bank capital ratios from, say, 100 years ago (when the composition of asset portfolios was very different) or about debt-equity ratios observed in firms in other economic sectors. The latter in particular offers no useful insights at all, without much more in-depth analysis. Finally, in this area, the Bank has made no serious attempt to engage with the probability of failure implied by the current standalone ratings the large banks have. These appear to be less frequent than once in 200 years. Those default probabilities appear consistent with the Bank’s assessment, in successive FSRs, about the soundness of the financial system, but not with these proposals, under which it is claimed that much more capital is required to put banks on a secure footing. Transitional effects and the path to a new steady state One of the most glaring omissions from the consultative document, and subsequent material, was any sustained analysis of how the financial system and the economy would react if the proposals the Bank is consulting on were implemented. The absence of any such analysis means that we can have no confidence that the Bank, with all the resources at its disposal, has thought through the issues and risks in depth themselves. The only estimates we’ve seen have been those for possible changes in lending margins for institutions affected by the proposed higher capital ratios. There has been no serious analysis published of the extent to which banks might become less willing to lend. And there has been no discussion about the extent to which business may migrate from regulated banks to either unregulated (i.e. not locally incorporated) banks here or abroad, or to finance companies, or of the possibility of disintermediation (such that more of society’s demand for credit is met without the direct interposition of a financial institution’s balance sheet). There has been no analysis of which economic sectors might be most severely affected. Large corporates for example will have plenty of alternative providers, probably at a price very similar to what they pay now, and many housing mortgages could be relatively easily securitised if necessary, but SMEs and rural borrowers might be more likely to bear the brunt of any price or capacity adjustment. Similarly, there was no analysis of where the brunt of any adjustment to deposit and wholesale funding interest rates might fall, but it seems reasonable to posit that wholesale creditors will not bear most of the burden. Perhaps more concerningly still, there is no sign of any analysis of whether a financial system in which more business has gravitated to institutions not locally-incorporated or to disintermediated markets would be (a) sounder, and (b) more efficient. There is a risk that

the core banks (already low risk) become somewhat safer, but that those institutions in future have a diminished role in the system. Most of the Bank’s analysis appears to, in effect, treat locally incorporated banks as the sum of the financial system, which is less likely to be the case in future if these proposals proceed. Failure to address these issues does not instill confidence. Finally, in this section, there was no discussion at all of the macroeconomic context in which these proposals would take effect. The proposals involved a transition over five years. Nine years into an economic recovery, with slowing domestic growth and growing global risks there has to be a fairly significant chance that the next significant recession will occur in the next five years (i.e. during the proposed transition period). That means a significant risk that regulatory policy would be exacerbating any downturn (through tighter credit constraints, reduced credit appetite, and potential higher pricing), in a downturn in which monetary policy is likely to be hard up against conventional limits (the Bank’s own analysis has suggested the OCR might be able to be cut only to around -0.75 per cent). Of course, if bank balance sheets were looking shaky it would be prudent to move ahead anyway - better ten years ago, but if not then now - but nothing in the Bank’s published analysis (past FSRs, stress tests, consultation document) nor in the credit ratings of the relevant institutions suggests anything like that sort of vulnerability. Without it, you will - with a reasonable probability - make economic management over the next few years more difficult (additional upfront potential economic costs), in exchange for the modest probability of making any real difference to (already very low) financial system risks over that period. It isn’t a trade￾off that appears to be worth making - at least not without much more supporting analysis than we have had to date. Anti-Australianism There has been a consistent subtext throughout the period of consultation that has the Reserve Bank antagonistic, or at best, indifferent to Australia and the Australian-owned banks. To be positive, it is probably marginally preferable to an alternative in which the Reserve Bank is subject to regulatory capture, identifying the public interest with the interests and views of the Australian banks, with whom they deal all the time. But the Bank simply has not demonstrated that it has got the balance right. We see this is the passive-aggressive approach to APRA, who were not apparently consulted in any depth as the proposals were brought together and were only shown the consultation document on the morning it was released (according to a timeline in one of the documents the Bank released in January). As already discussed, the Bank has made no effort to benchmark its proposals against the requirements APRA will be imposing on many of the same banking groups, or to explain why it believes what APRA is proposing is not nearly demanding enough for New Zealand. We’ve also seen it in rather glib comments that perhaps the Australian banks might sell down their stakes in their New Zealand subsidiaries, in a tone which implies that Reserve Bank senior managers think this might be quite a good thing. Anti-Australianism is a

recurring theme in New Zealand political debate around banks, but it should have no place in the assessments or public comments of officials operating under the Reserve Bank of New Zealand Act. In my view, New Zealand benefits considerably - in terms of financial system soundness and efficiency - from the fact that the major banks are all part of much larger banking groups, each headquartered in a friendly country with good institutions, and strong record of financial stability. The Reserve Bank should not lightly jeopardise that situation with proposals that simply aren’t backed by robust analysis of the risks they are supposed to mitigate or of the costs of adjustment. CONCLUSION The Reserve Bank has simply not made a compelling case for an increase in minimum capital ratios that, taken together with the conservative manner in which these ratios are calculated here, would appear likely to position New Zealand with among the very highest core minimum capital ratios anywhere in the advanced world. Without a compelling case, citizens will pay a significant annual insurance premium upfront, with a relatively low probability of ever realising any benefit from the policy. Moreover, owners of private businesses would be coerced into altering their business models without any robust ex ante cost-benefit analysis to support such a regulatory imposition. The Bank’s documents show little sign that it has thought seriously and deeply about the nature, character, and costs of ‘financial crisis’ - all the more troubling since one of the benefits the Bank sometimes claims for having supervision in the same institution as (macro-focused) monetary policy is that it supposedly allows a richer, more multi￾dimensional perspective on relevant financial stability and macroeconomic issues. None of that has been on display in this consultation. Serious recessions are things to seek to mitigate. That is primarily the role of discretionary monetary policy, made possible by a floating exchange rate. Serious misallocations of resources are likely to be costly, but the misallocations arise in the good times - when credit is growing strongly - not in the subsequent bust. The marginal additional losses arising from financial crises themselves appear to be (typically) small, and these proposals in any case involve only a further modest reduction in an already low risk of serious problems (in a country with little history of serious systemic financial problems). There are limits to what any regulators and officials can do about initial misallocations, but my recommendation to the Bank would be to abandon the push for higher minimum capital ratios (while proceeding to level the playing field between advanced and standardised model banks) and to focus its energies instead on sharpening its ability to recognise, and respond vigorously to, any sharp deteriorations in lending standards promptly when and if they get underway. Complement that with robust championing of (a) the importance of the floating exchange rate regime - especially in a country with neutral interest rates higher than the rest of the world - and (b) of keeping the government out of the business of directing credit and, together with existing demanding capital standards, you are likely to best serve the interests of New Zealanders. Better that approach than the (probably costly)

steep increases in capital requirements proposed in the consultation document without anything like adequate, carefully and independently scrutinised, supporting analysis. New Zealanders deserve better than they have had in the poor process and weak substance that together made up this consultation.

Michael Reynolds I fully support this proposal. As an NZ tax payer who carries a lot of risk should the banks need to be protected from failure after a financial shock I find this proposal entirely reasonable. I believe raising the capital should not be a problem for the highly profitable banking sector and the those owners and shareholders who benefit should shoulder more risk associated with banking instead of NZ taxpayers, depositors and society at large. Even if there are effects on lending I believe this is a worthy trade off for a more stable banking sector. OIA s9(2)(a)

Michael Taylor Increasing the levels of capital required to be held by banks to lower the risk of the NZ banking system may sound good in theory but there are several potential unintended consequences which need to be carefully considered:

1. If the shareholders of the 4 banks owned by Australian Banks decided to sell this could increase the risk of the banking system. e.g. A New Zealand bank owned by a strong parent is a lower risk and has lower funding costs than a NZ owned bank. In a sever financial downturn, a NZ owned bank may struggle to raise the necessary capital to cover lending losses and would be more likely to fail (e.g. BNZ during 1980s which had to be bailed out twice and ended up being owned offshore).
2. Increased capital could lead to Banks rationing capital to low returning sectors e.g. Agri/Rural lending. This could push some farmers out of business and could also create downward pressure on land prices which could result in significant losses for the Banks etc. Higher risk sectors (e.g. property development) could also suffer.
3. Increased capital, regulatory and compliance costs for Banks has lead to increasing levels of second tier finance companies and offshore credit funds (examine what has happened in Australia) which are not restricted to the same degree as the banks. This increases the overall risk to the economy as these lenders could go under in a recession (e.g. Finance companies following the GFC) or could suddenly withdraw from the market if circumstance change creating a credit crunch.
4. Increased capital requirements will force banks to increase margins (through both higher lending margins and lower deposit rates) to achieve adequate returns. This will result in a significant increased cost to the overall economy.

Mike whittaker any increase in capital for the banks will mean more costs for the end user, yes New Zealanders will pay more and we are sick of increased costs by the government Your own report clearly stated a number of times that the NZ banks where in a very safe position so why are we doing this the best country in the world has had enough of more costs, the banks will pass it on to the everyday New Zealander OIA s9(2)(a)

Murray Burdan I agree with the RBNZ to increase the capital of the banks. I also agree with the new requirements. The only thing I think you should review is the amount of time you take to implement it. To lessen the impact on depositors this should be transitioned over a longer period. OIA s9(2)(a)

M E Jackson May 17th, 2019 Attention: Ian Woolford Financial System Policy and Analysis Department Reserve Bank of New Zealand PO Box 2498 Wellington 6140 Email: CaptialReview@rbnz.govt.nz Dear Mr Woolford Re: Bank Capital Review I write on my own behalf, as a depositor in local banks, and seek your consideration of my comments on the Reserve Bank's consultation papers of January 2019 (and other dates). I support the Reserve Bank’s proposal and the Reserve Bank’s innovation to consider the social impacts of its policies in its cost/benefit analyses. I can only hope that this innovation will extend to monetary policy. My primary feeback to the Reserve Bank’s proposal for Bank Capital are:

1. The Reserve Bank’s proposal does not go far enough (See Attachment A)
2. The IRB approach to assessing credit risk is fundamentally flawed, and underestimates financial risk of banks (See Attachment B);
3. Any transgression of the Modigliani–Miller theorem (MM offset less than 1.0) should be considered a reason to increase capital requirements for large banks not decrease them (See Attachment C).
4. The economic efficiency benefits of Tier 2 capital have not been adequately considered. (See Attachment D). The balance sheets of New Zealand banks show a large imbalance in favour of unproductive lending to existing residential property over productive businesses . It is clear that incentives arising from 1 current weak capital regulation of the largest banks (such as low risk weightings for investment in existing housing) cause large banks to prefer lending to the residential sector to the cost of the productive sector, resulting in lower productivity, high levels of debt relative to income and high levels of borrowing from overseas. A significant increase in bank capital, at least to the levels proposed by the Reserve Bank, is necessary to reduce the incentives for this misapplication of the country’s credit by the four largest banks. Thank you for consideration of my submission. Yours faithfully 1 More than 60 percent of bank lending is to the household sector. OIA s9(2)(a)

M E Jackson 2

Attachment A: Utility of Increased Bank Capital Bank equity capital has several important roles. However, in the Reserve Bank’s Capital Review analyses, it seems the bank is focused on only one: the function of bank capital to act as a buffer to absorb financial losses that a bank may suffer. However there are two other critical roles that have not been adequately addressed by the Reserve Bank:

1. Limitations to both bank leverage (minimum capital requirements) and asset leverage (such as LVR limits) are key factors for ensuring financial stability. Without an adequate level of pre-existing savings being required to support lending against existing assets (of at least 38.2% made up from a combination of minimum deposits or bank capital), the financial 2 system is inherently unstable (unless the resultant excess money creation is balanced by Crown surpluses). Without such a minimum level of pre-existing savings being applied to support lending, the risk of financial crisis is not just a probabilistic speculation but an inevitability;
2. A higher level capital acts to regulate the rate of return to bank equity. The rate of return to equity matters because it affects the rate of credit growth that the bank can support. Rapid credit expansion directed towards the purchase of existing assets (investment for no gain in productivity) is a predictor of future financial instability and economic loss. In New Zealand, with bank equity capital at current levels the four major banks have enjoyed rates of return to their equity of more than 13% after tax, significantly greater than New Zealand owned banks. This high rate of return has allowed the major banks to fund credit growth into existing housing assets at a growth rate of more than 8% per annum while still at the same time allowing dividend returns much greater than any deposit they offer to their customers. These high rates of return, that are far above their cost of capital, shows that the large banks hold substantial market power, which power they can use to increase costs to their customers to preserve their excess returns on capital. A perspective on the magnitude of the excess returns is indicated by the reported price-to-book ratios of banks (share price as a ratio of NTA per share) that are listed on the 3 share markets in Australia and New Zealand . The typical NTA ratio of at least 1.5 shows that the rate of return to equity of banks is at least 50% higher than the shareholders true cost of capital. The low levels of capital required of large banks allows them to increase their market share ahead of smaller banks, which is harmful because: ● it has been demonstrated that it is smaller banks who tend to lend to support productive enterprises, such as small businesses; ● It provides the large banks with market power by which to preserve their excess rate of return that they would be otherwise prevented from doing by virtue of have to compete with the smaller banks on a level playing field.. In summary, the proposed cost/benefit analysis must consider the benefit of levelling the playing field and increasing the competitiveness of the small banks which tend to lend for productive purposes rather than property speculation. 2 I would be happy meet and explain the basis for this assertion. 3 NTA =net tangible assets, being the market capitalisation divided by the balance sheet equity including retained earnings. 3

Attachment B: IRB Methodology is Fundamentally Flawed IRB methodology is fundamentally flawed when applied to regulation of bank capital because it presupposes that financial risks can be determined statistically when they cannot. In particular, the assumption that financial risk can be simulated or estimated by normal or lognormal distributions severely underestimates actual financial risk. The methodology assumes that risk can be determined from observation of an external market, and fails to consider that the amount of capital might itself be a factor in system risk. This matter has been discussed in depth by others (see for example “Capital Inadequacies: The Dismal Failure of Bank Capital Regulation” by Dowd, Hutchinson, Hinchliffe 2011). In addition, Vice Chairman of the FDIC Thomas Hoenig observes in relation to risk weights; “If risk weights could be assigned that anticipate and calibrate risks with perfect foresight, adjusted on a daily basis, then perhaps risk-weighted capital standards would be the preferred method for determining how to deploy capital. However, they cannot. To believe they can is a fallacy that puts the entire economic system at risk.”; Further, Andrew Haldane of the Bank of England also considers (in his article Capital Discipline (2011)), that the output of capital calculations under Pillar I are “no longer easily verifiable or transparent. They are as much an article of faith as fact, as much art as science. This weakens both Pillars II and III. For what the market cannot observe, it is unlikely to be able to exercise discipline over. And what the regulator cannot verify, it is unlikely to be able to exercise supervision over.“ He reminds us of Hayek’s caution regarding the “pretence in the minds of risk managers and regulators” as to their understanding the dynamics of complex systems. He warns that policy built on such pretence risks catastrophic error. Even the Reserve Bank has acknowledged that “Critics of the Basel framework maintain that ever-increasing risk granularity has led to a spurious sense of sophistication and that the pendulum should swing back to simpler, more objective measures of risk”. New Zealand can reduce the wasteful expenditure of analytical resources involved in fiddling with the flawed IRB methodology by avoiding its “spurious sophistication” entirely. For example, Admati has advocated that equity requirements be set at 30 percent of total assets (see “The Missed Opportunity and Challenge of Capital Regulation”,) If this flawed and optimistic methodology is going to be used, the criterion for the tolerable estimated frequency of crises should be far greater than 1/200. Recognising that a financial crises will impact heavily on health and well-being in a society, with such impacts giving rise to heavy loss of life through stress and suicide, any other industry would require a tolerance of such an event of 1/1000 or perhaps 1/10000 years. I submit that a tolerance of at least 1/1000 years should be required. 4

Attachment C: MM Offset and Market Power The main proposition of‘ “capital structure irrelevance’ theorem of Modigliani and Mille r (MM) is that 4 “the market value of any firm is independent of its capital structure”. The degree to which banks absorb the potential cost increase due to higher capital is represented by the Modigliani and Miller ‘offset’ (or the ‘MM offset’). The greater the MM offset, the more the MM theorem is assumed to hold so that with an MM offset of 1.0 the banks are assessed to be unsuccessful in imposing high rates of return on the incremental equity that displaces funding by creditors. The Reserve Bank notes that recent empirical studies of the Modigliani-Miller theorem suggest that “changes in banks’ funding structures towards costlier sources (equity) pass through to higher lending rates at a rate of around 50% ”. Therefore it appears that that the Reserve Bank accepts that banks will be quite successful in obtaining a high equity return on any incremental capital that displaces deposits (deposits that would otherwise be paid a pitifully low rate of interest). This seeming passive acceptance by the Reserve Bank that the MM theorem “may not hold in practice” supports the critics of the Reserve Bank’s proposals. However, the Reserve Bank should not treat the MM offset as simply a parameter to be empirically ascertained. Rather that Reserve Bank should consider any finding that MM offset is less than 1 as a deficiency in regulation and evidence of the application of market power by large banks to the detriment of New Zealand. An MM offset of less than 1.0 can only arise because the large banks have market power by virtue of bias in capital requirements regulation, and the means to attenuate that dominating market power is by levelising the playing field in favour of smaller banks by regulating for standardised capital requirements. The Reserve Bank has expressed the view that the MM offset is the degree to which the shareholders accept a reduction in the return on equity. This must be incorrect. Financial theory would suggest that shareholders simply place a value on the degree to which the managers of the bank succeed in maintaining a high rate of return on capital by the exercise of market power. It is concerning that the Reserve Bank might be accepting of the proposition that an increase in capital requirement for a given asset base has a high cost that might be legitimately passed on when any accepted theory as to the cost of capital would determine that the actual marginal cost of incremental capital from the perspective of shareholders is the risk free rate of return. 4 There are two variants of the MM theorem. When considering the welfare of New Zealand, the relevant MM theorem is the one that does not consider the extraction of tax subsidies from the tax payer as a value to the enterprise. 5

Attachment D: Benefits of Tier 2 capital

For the reasons asserted in Attachment C, if the banks’ see high leverage as optimum it is only because creditors (including depositors) are not properly compensated for the risk that the bank imposes on them by high leverage. This effect is profound and leads to systematic underpricing of financial risk in the whole financial system. Unless bank capital requirements are modified, this underpricing of risk in the intermediation between lenders and borrowers is damaging to the welfare of the national economy. To increase intermediation efficiency between depositors and banks, capital requirements must be structured such that at some point in the process of a bank increasing its leverage there is the opportunity for at least some lenders to the bank, at least at the margin, to negotiate fair arms-length commercial terms to reflect the increasing risk of their position. Such a dynamic could be developed by requiring the minimum total capital ratio to be set much higher than the minimum Total Tier 1 capital ratio. A bank could then chose as to whether it wanted to fund the difference by (CET1) equity or by Tier 2 capital . Holders of Tier 2 instruments have powerful incentives to closely scrutinize the risks of their investments, and would seek adequate risk pricing. If an increase capital requirements were to be effected by allowing banks to meet that increase by Tier 2 capital requirement alone, then such increase of total capital requirements from current levels will have zero welfare cost to the economy. It would have zero welfare cost because: (i) it does not require that bank shareholders raise equity, rather it need only require that existing creditors exchange one form of debt (deposits) for another (e.g. subordinated debt): (ii) there is no increase in creditor risk given constant Tier 1 equity; and (iii) any increase in return to creditors would simply be reflective of the Tier 2 holders being able to obtain a fair market-based return for the risk they are exposed to. For example, if a requirement to meet total capital requirements of 19% is put in place, banks will have the choice of meeting this requirement by choosing the relative proportion of Tier 1 to Tier 2 capital, according to which they perceive as having the lower cost. In exercising this choice there will improved efficiency of intermediation between creditors and banks, because creditors will be able to obtain a fair risk premium for their exposure. In short, if the banks were enabled and required to optimise between Tier 1 and Tier 2 “ risk will have been identified by all parties it affects, priced by those able freely to choose to bear it, allocated to those who are best placed to bear it, and managed by those parties in the most efficient manner.” This is the precondition for financial stability, as the Reserve Bank has previously observed. 6